COMPOSITES SCIENCE AND TECHNOLOGY

Composites Science and Technology 67 (2007) 128 www.elsevier.com/locate/compscitech

Review

Vibrations of carbon nanotubes and their composites: A review

a,*

, Emmanuel O. Ayorinde a, Yuan-Feng Wen

Mechanical Engineering Department, Advanced Composites Research Laboratory, Wayne State University, Detroit, MI 48202, USA b Department of Logistics Management, National Kaohsiung Marine University, Kaohsiung, Taiwan, ROC Received 25 January 2006; received in revised form 15 March 2006; accepted 30 March 2006 Available online 24 May 2006

Abstract Spurred by the discovery of carbon nanotubes in 1991 and the subsequent nanotechnology revolution, there has been a marked increase in the volume of scientic and engineering literature dealing with various aspects of the vibratory behavior of carbon nanotubes and their composites in the last few years. This paper is an attempt to present a coherent yet concise review of as many of these publications as possible. The major topics covered in this review are modeling and simulation of vibrating nanotubes, studies of nanomechanical resonators and oscillators, the use of vibration measurements to characterize nanotube mechanical properties, nanotube augmentation of dynamic structural properties of composites, vibrations of nanotube-based sensors and actuators, sonication of nanotube-reinforced polymer resins, Raman scattering from nanotubes, and high frequency wavenanotube interactions. Summaries of both theoretical and experimental studies are included, along with key conclusions and recommendations from these studies. 2006 Elsevier Ltd. All rights reserved.
Keywords: A. Nanostructures; A. Polymer matrix composites; B. Vibration

Contents 1. 2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modeling and simulation of vibrating nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Continuum models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Hybrid atomistic/continuum and atomistic models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Modeling nanotube waviness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nanomechanical resonators, oscillators and eld emission devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Use of vibration measurements to characterize nanotube mechanical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nanotube augmentation of dynamic mechanical properties of composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vibrations in nanotube-based sensors and actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sonication of nanotubes and nanotube-reinforced polymer resins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Raman scattering from nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Absorption of high frequency waves by nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2 2 3 6 6 9 10 14 17 19 21 23 23 23

3. 4. 5. 6. 7. 8. 9. 10.

Corresponding author. Tel.: +1 313 577 3702; fax: +1 313 577 8789. E-mail address: gibson@eng.wayne.edu (R.F. Gibson).

0266-3538/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.compscitech.2006.03.031

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

1. Introduction The number of publications on carbon nanotubes (CNTs) and CNT-reinforced composite materials has grown very quickly since CNTs were discovered by Iijima [1] more than a decade ago, as evidenced by several recent review articles on mechanical behavior [2,3], fabrication [4,5] and applications [610] of such materials. In 2003 a special issue of this journal was dedicated to modeling and characterization of nanostructured materials [11]. Vibrations of CNTs are of considerable importance in a number of nanomechanical devices such as oscillators, charge detectors, clocks, eld emission devices and sensors. In addition, CNT vibrations occur during certain manufacturing processes (e.g., ultrasonication in nanocomposite processing) and as part of some nondestructive evaluation processes (e.g., Raman spectroscopy). Electron microscope observations of vibrating CNTs have also been used to indirectly and nondestructively determine the eective elastic moduli and other aspects of mechanical behavior of the CNTs. Microwave excitation has been found to cause intense heating of CNTs. So there is considerable motivation for studying vibration characteristics of CNTs, and the purpose of this article is to review recent literature which is relevant to all aspects and ramications of CNT vibrations. 2. Modeling and simulation of vibrating nanotubes It is important to have accurate theoretical models for the natural frequencies and mode shapes of carbon nanotubes for several reasons. For example, if the nanotubes are to be used as nanomechanical resonators, the oscillation frequency is a key property of the resonator. In addition, the eective elastic modulus of a nanotube may be indirectly determined from its measured natural frequencies or mode shapes if a suciently accurate theoretical model is used. Several recent publications have reviewed the modeling and simulation of mechanical properties of CNTs and nanocomposites [12,13] but these review articles do not cover vibrating nanotubes. 2.1. Continuum models A continuum approach using the well known frequency equation for a beam has been employed by several investigators to estimate the theoretical fundamental exural resonance frequencies of nanotubes [2] and nanowires [14]. For example, according to the well-known BernoulliEuler beam theory, the equation describing transverse, or exural motion of a continuous, homogeneous, isotropic, linear elastic beam which is without rotary inertia or shear eects and whose properties do not vary along its length may be expressed as EI o4 w o2 w qA 2 0; ox4 ot 1

where E is the eective Youngs modulus of beam material, I the moment of inertia of beam cross-section about its neutral axis, A the cross-sectional area of beam, q the mass density of beam material, x the distance along beam, w(x, t) the transverse displacement of beam and t is the time. The corresponding frequency equation for exural vibration is  1=2 k n L2 EI ; 2 fn 2pL2 qA where L is the length of beam, n the mode number, knL the eigenvalue for nth mode and fn is the nth mode frequency, Hz. The eigenvalues depend on the boundary conditions, and the choice of boundary conditions is obviously very important for an accurate simulation. For example, the rst three eigenvalues for a xed-free cantilever beam are k1L = 1.875, k2L = 4.694, k3L = 7.855, whereas the corresponding values for a free-free beam are k1L = 4.73, k2L = 7.853, k3L = 10.995. Eigenvalues and the corresponding mode shape functions have been tabulated by Blevins [15] for the rst few modes and the most commonly encountered boundary conditions. For a nanotube or any beam having tubular geometry with outside diameter Do and inside diameter Di, the ratio I/A is such that Eq. (2) becomes 1=2 2 k n L ED2 D2 o i : 3 fn q 8pL2 As shown later, Eqs. (2) and (3) have been used in a variety of applications such as estimating the frequencies of nanoresonators and determining the eective exural modulus of CNTs from measured modal frequencies. Such beam models are only useful for calculating exural modes. Noncoaxial intertube resonances of multi-wall carbon nanotubes (MWCNTs or MWNTs) are those modes of vibration which involve relative motions, or intertube displacements between the concentric tubes [16]. Yoon et al. [17] have shown that such resonances can be modeled for an N-wall tube by modifying the BernoulliEuler equation to include van der Waals interaction forces and solving N coupled equations of the form c1 w2 w1 EI 1 o4 w1 o2 w1 qA1 2 ; ox4 ot o4 w 2 o2 w 2 c2 w3 w2 c1 w2 w1 EI 2 4 qA2 2 ; ox ot . . . cN 1 wN wN 1 EI N o4 wn o2 w n qAN 2 ; ox4 ot

where ck is the van der Waals interaction coecient for the kth tube, wk(x, t) the deection of the kth tube, k = 1, 2, . . . (N 1). This simulation demonstrated that for low aspect ratio MWNTs, the lowest predicted noncoaxial intertube

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

resonance frequencies are in the same range as the higher exural modes of the entire MWNT, and are easily in the terahertz range. Zhang et al. [18] used a similar approach to model the van der Waals interactions between transversely vibrating double-walled CNTs under compressive axial load. It was shown that increasing compressive axial loads cause reductions in the frequencies, but that the amplitude ratios of the inner to the outer tubes are independent of axial load. For short beams, higher modes, and/or beam materials having high E/G ratios (where G is the shear modulus of the beam), the eects of through-thickness shear deformation and rotary inertia can become signicant, in which case the Timoshenko beam theory is more accurate than the BernoulliEuler theory. Such eects can be of particular signicance in the case of high frequency intertube resonances in MWNTs. In the Timoshenko theory [19], the transverse deection w(x, t) and the slope u(x, t) due to bending of the beam are governed by the coupled equations   ou o2 w o2 w KGA 2 p qA 2 ; ox ox ot 5   2 ou ow o2 u EI 2 KGA u qI 2 ; ox ox ox where K is the shear factor, which depends on cross-sectional geometry and p is the distributed loading per unit length of beam. Yoon et al. [20] developed a double Timoshenko beam model to characterize the motions of the inner and outer tubes in a double wall CNT and then used the model to study the eects of rotary inertia and shear deformation on terahertz frequency transverse wave propagation in double wall CNTs. It was concluded that, because of the relative motions between the inner and outer tubes at high frequencies, the Timoshenko beam model is more relevant than the BernoulliEuler model for terahertz frequency wave propagation in double wall CNTs. A similar Timoshenko beam model was used to study terahertz frequency vibrations of short CNTs [21]. The fundamental frequency of a typical single SWNT vibrating in the free-free mode is easily in the GHz range, and when the nanotube is embedded in a polymer matrix which provides elastic restraint, the frequencies are even greater. For MWNTs embedded in an elastic medium such as a polymer matrix, the eect of the matrix on the resonance frequencies of the MWNT has been estimated by using a beam-on-elastic foundation model [22]. In this model, Eqs. (4) are modied by adding a distributed elastic reaction force to the last of Eqs. (4) which applies to the outer tube in a MWNT, as p cN 1 wN wN 1 EI N o4 wn o2 w n qAN 2 ; ox4 ot 6

of this simulation show that when the elastic medium surrounding the MWNT is very compliant, the vibrational modes consist of both global exure and noncoaxial intertube deformations, but when the elastic medium is very sti, almost all of the vibrational modes are essentially of the noncoaxial intertube type. Taking into account the van der Waals forces between nanotubes, a self-consistent model for the Youngs modulus and density of SWNTs and their hexagonal arrays [23] as well as the natural frequencies of hexagonal arrays of SWNTs have been estimated [24]. In related work, Wang et al. [25] developed a multiple elastic shell model based on the Flugge elastic shell theory [26] and used it to study vibrations of MWNTs. By using a shell model, radial breathing modes were included along with the exural modes governed by the beam models described in Eqs. (1)(6). This work demonstrated that interlayer van der Waals forces have a particularly signicant eect on the radial breathing modes of large radius MWNTs, but a smaller eect on radial modes of small radius MWNTs, and practically negligible eects on the torsional and longitudinal modes of MWNTs. In a separate paper, Wang et al. [27] developed simplied elastic shell models for buckling and free vibration of CNTs based on the shell equations of Flugge and Donnell. Classical theory of elasticity solutions for the vibrational modes of thin-walled hollow isotropic cylinders have been applied by Mahan [28] to calculate all of the low frequency modes of single wall carbon nanotubes (SWCNTs or SWNTs), including exural, torsional, radial breathing, and longitudinal modes. If the element size is chosen to be suciently small relative to the tube size, nite element approximations to the elasticity solutions can be used to study vibrations of nanotubes. As shown in Fig. 1, nite element simulations based on cylindrical shell approximations of CNTs have been employed to explain observed reductions in eective Youngs modulus of CNTs with increased diameter in exural resonance experiments [29]. As the diameter of the nanotube increases, so does the likelihood of rippling on the compressive side of the bent nanotube, which lowers the resonance frequency and the corresponding eective modulus which is determined from the frequency. 2.2. Hybrid atomistic/continuum and atomistic models The basic premise in the articles referenced above is that a CNT can be modeled as a continuum having continuous distributions of mass and stiness, and the atomistic nature of the CNT is not accounted for. In an attempt to relate the atomistic and continuum approaches, Odegard et al. [30] developed an equivalent continuum model for structure property relationships in nanostructured materials. This approach serves to link computational chemistry of molecular properties with solid mechanics of macroscopic mechanical behavior. The linkage is accomplished by equating the molecular potential energy of the nanostructured material with the mechanical strain energy of a

where p = KwN is a Winkler-type elastic reaction force per unit length and K is a constant which depends on the material in which the MWNT is embedded. The results

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

Fig. 1. Finite element simulations explain observed reductions in eective Youngs modulus of CNT with increased diameter in exural resonance experiments. From Wang and Wang [29].

representative volume element of a continuum model. This method was used to determine the bending rigidity of a graphene sheet and was later integrated into a more comprehensive multi-scale modeling approach [31]. In related work, Bodily and Sun [32] used the potential energy for the interatomic forces of a molecular mechanics model to develop equivalent pin-jointed truss models for SWNTs, then determined the extensional, torsional and bending rigidities of the truss. These stinesses were then assigned to an equivalent thin-walled continuum hollow cylinder in order to calculate the Youngs modulus of the SWNT, but vibrations were not considered. Li and Chou [33,34] developed a similar hybrid atomistic/continuum approach which they refer to as molecular structural mechanics and used it to predict both the static and dynamic mechanical behavior of CNTs. The nanotubes were assumed to be analogous to geometrical frame-like structures, with the primary bonds between neighboring atoms being analogous to massless load-bearing members and individual atoms acting as concentrated masses at the joints in the structure. Fig. 2 shows the dierent types of interatomic interactions that were modeled with this approach. From molecular mechanics, the total steric potential energy for the structure is assumed to be the sum of energies due to valence or bonded interactions and nonbonded interactions X X X X U Ur Uh U u U x U VDW ; 7 where Ur, Uh, Uu, Ux and UVDW are the energy terms due to bond stretching, bond angle bending, dihedral angle torsion, out-of-plane torsion and van der Waals interaction, respectively. From structural mechanics, the deformation of the frame-like structure leads to a change in the total strain energy of the structure X X X X U UA UM UT U V; 8 where UA, UM, UT and axial, bending, torsion The eective stinesses the structure are found UV are the strain energies due to and shear loading, respectively. of the load bearing elements in by evaluating the various energy

Fig. 2. Interatomic interactions in the molecular structural mechanics approach. From Li and Chou [33].

terms and then setting Eq. (7) equal to Eq. (8), then the global stiness matrix for the structure is assembled from these individual element stinesses. The global mass matrix is developed on the assumption that the masses of the carbon nuclei are concentrated at the joints of the structure. Once the stiness and mass matrices for the structure are determined, the methods of structural mechanics can be used to solve static or dynamic problems. For example, for free vibration of the nanotube in its natural modes, the equation of motion is Mfg Kfyg f0g; y 9 where [M] and [K] are the global mass and stiness matrices, respectively, and fg and {y} are the nodal acceleray tion and displacement vectors, respectively. After condensation of the mass and stiness matrices, converting to principal coordinates, and assuming harmonic motion, the natural frequencies and mode shapes are determined by solving the eigenvalue/eigenvector problem

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

Ks x2 Ms fy p g f0g;

10

where [K]s and [M]s are the condensed stiness matrix and mass matrix, respectively, {yp} are the principal coordinates and x is the natural frequency. Li and Chou [35] used this method to predict the frequencies and mode shapes of SWNT and double-walled nanotube (DWNT) nanomechanical resonators, and some typical mode shapes are shown in Fig. 3. Note that the molecular structural mechanics approach yields all possible modes, including bending modes, radial breathing modes, and noncoaxial intertube modes. This approach was also used to determine the eects of tube diameter, length and boundary conditions on the frequencies. The results in Fig. 3 are based on the assumption that the nanotubes are rigidly clamped on both ends, and indeed the specication of boundary conditions is essential to the solution of such eigenvalue/eigenvector problems. Getting the theoretical boundary conditions to exactly simulate the actual boundary conditions for real structures is always a challenge, and for nanostructures it may be even more dicult. For example, is it possible to rigidly clamp a cantilevered nanotube as shown in Fig. 4, or is it necessary to include the elasticity of the support, as also shown in Fig. 4? This is one of the diculties in using measured resonance frequencies to back out the eective modulus of a nanotube, as described later. For vibration testing of composite beams and plates at the macromechanical level [36], it has been found that the free edge boundary condition (i.e., freely suspending the specimen by thin laments or other means) is more easily simulated than the clamped or hinged edge condition, but it is not yet known whether this conclusion is valid or not for nanostructures. For a macromechanical cantilever beam specimen [37], it is possible to indirectly determine the support stiness (i.e., the

Fig. 4. Boundary conditions for a cantilever beam.

torsional stiness K in Fig. 4), but it remains to be seen if this approach will work at the nanomechanical level. Atomistic approaches to the analysis of CNT vibrations include classical molecular dynamics simulation, ab initio and tight binding molecular dynamics simulation methods A detailed review of these methods has been published by Qian et al. [2] and further review is beyond the scope of this paper. These atomistic methods tend to be very computationally intensive, and are typically limited to a relatively small number of atoms and very short times. As shown in Fig. 5, Zhou and Shi [38] used molecular dynamics simulation to show that the application of tensile loading to a SWNT caused a self-excited vibration in the radial direction, and that the frequency of the vibration was independent of the tensile strain rate. This frequency was found to be 4700 GHz, which was conrmed by experiments. However, using a tight binding molecular dynamics simulation of the same problem, Dereli and Ozdogan [39] observed the same induced radial vibration at the same frequency, but found it to be sensitive to tensile strain rate. Ab initio calculations of structural, elastic and vibrational properties of SWNTs have included the eects of chirality and radius [40].

Fig. 3. Mode shapes for SWNT and DWNT from molecular structural mechanics approach. Nanotubes are assumed to be rigidly clamped on both ends. From Li and Chou [34].

Fig. 5. Molecular dynamics simulation of radial vibration in SWNT induced by applied tensile strain. From Zhou and Shi [38].

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

2.3. Modeling nanotube waviness In all of the models and simulations referred to above, the nanotube was assumed to be perfectly straight. However, photomicrographs of nanocomposites such as those in Fig. 6 from Qian et al. [41] indicate that nanotubes may exhibit signicant waviness. The eects of waviness on the elastic moduli of CNT-reinforced polymer composites have been shown to be signicant. For example, Fisher et al. [42,43] have employed 3D nite element models to show that when waviness is taken into account, the predicted composite moduli are in much better agreement with experimental values than they are if waviness is neglected. Gibson and Anumandla [44] combined the ber waviness models of Chan and Wang [45] and Hsaio and Daniel [46] with the 3D averaging procedure of Christensen and Waals [47] to develop a closed form micromechanics model for predicting the elastic modulus of a wavy nanotube-reinforced polymer composite. The nanotube waviness is described by the simple sinusoidal model   2px z A sin 11 LNT as shown in Fig. 7. When the predictions of this model are compared with experimental results for a MWNT-reinforced polystyrene composite [48], the model signicantly overpredicts the composite modulus unless signicant waviness is assumed (Fig. 8). Although it appears that no similar work has been done on the eects of waviness on the vibration characteristics of nanotubes, the corresponding macromechanical models for vibrating curved beams and arches show that the curvature of the beam or arch has a strong eect on modal frequencies [49,50].

Fig. 7. Representative volume elements for closed form micromechanics model of wavy nanotube-reinforced composite. From Gibson and Anumandla [44].

Fig. 8. Comparison of predicted elastic modulus values for a MWNTreinforced polystyrene composite with experimental values. w = waviness = A/LNT where A and LNT are dened in Fig. 8. From Gibson and Anumandla [44].

3. Nanomechanical resonators, oscillators and eld emission devices Nanometer-scale high frequency resonators are critical components of many nanoelectromechanical systems (NEMS) such as oscillators, charge detection devices, parametric ampliers, nanoscale clocks and resonant sensors [5153]. For example, Sazanova et al. [52] have developed a nanoelectromechanical oscillator based on the electrostatic actuation and detection of a double-clamped CNT with a gate electrode, as shown in Fig. 9. Zheng et al. [54,55] have proposed the design of an oscillator which should have an operating frequency of several Gigahertz based on the oscillation of the inner tube of a MWNT with respect to the outer tube of the MWNT. As shown in Fig. 10, the inner tube, or core, is assumed to oscillate inside the outer tube about the position of minimum van der Waals potential energy. Liu et al. [56] used molecular dynamics simulations of MWNTs to show that the oscillatory frequencies saturate as the number of outer tubes exceeds three. Schwab [57] has shown that by fabricating

Fig. 6. Transmission electron microscope image of MWNT-polystyrene composite lm showing waviness of nanotubes. From Qian et al. [41].

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

Fig. 10. Modeling of gigahertz frequency oscillations of MWNT core inside outer shell due to excess van der Waals interaction energy and weak intershell sliding resistance. From Zheng and Jiang [54].

Fig. 9. Nanomechanical oscillator based on doubly clamped CNT and gate electrode. From Sazonova et al. [52].

a single-electron transistor onto a mechanical system in a high magnetic eld, the mechanical spring constant and the damping constant of a CNT nanomechanical resonator can be manipulated by adjusting the potential of a nearby gate electrode. A molecular structural mechanics approach [33] has been used to predict the fundamental frequencies of cantilevered or bridged single wall carbon nanotube (SWNT)

nanomechanical resonators, and the results shown in Fig. 11 indicate that these frequencies could reach the level of 10 GHz1.5 THz [34]. Typical predicted mode shapes were shown previously in Fig. 3. It has been reported that ropes consisting of hundreds of SWNTs have exural frequencies in the MHz range [58]. Individual cantilevered SWNTs having aspect ratios (i.e., ratio of length L-to-outside diameter Dout) of approximately L/Dout = 100 were also found to have fundamental exural resonance frequencies in the MHz range [59]. High Q (quality factor) resonator or oscillator arrays based on CNTs have been shown to be eective nanoelectromechanical signal processing devices for transmission of RF signals [60]. The nanotube array, which is based on vertical CNT arrays integrated with waveguides and bias electrodes, serves to transmit the RF signal only at its

Fig. 11. Use of molecular structural mechanics approach to study SWNTs as nanomechanical resonators. From Li and Chou [34].

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

mechanical resonance frequency and reects RF power at all frequencies except the mechanical resonance frequency. At the CNT array resonance frequency, the incident RF wave couples to nanotubes near the input electrode due to Coulomb forces between the RF electric eld and the electric charge on the CNTs. In related work, it has been proposed that a CNT lm grown on a quartz surface could be used as a miniaturized RF lter [61]. The sensitivity of a resonator depends strongly on its quality factor, Q, which is the inverse of mechanical damping or energy dissipation. Accordingly, several studies have been concerned with investigations of the energy dissipation in CNTs. For example, Jiang et al. [62] used classical molecular dynamics simulations of a cantilevered CNT oscillator (Fig. 12) to study the eects of temperature and interlayer binding strength on the Q factor for the oscillator. The simulation predicted that the Q factor for the SWNT varies with the temperature T according to Q / 1/T0.36, where T is the temperature in K. This is dierent from the classical thermoelastic dissipation in an isotropic crystalline beam, which follows the relationship Q / 1/ T. It was also found that the energy dissipation was much greater (i.e., Q factor was lower) for DWNTs than for SWNTs, apparently because of the interlayer interactions in the DWNTs. An interesting related observation was that the energy dissipation is much greater in DWNTs when the lattice structure for the two tubes is commensurate (e.g., armchair/armchair or zigzag/zigzag) than it is when the lattice structure for the two tubes is incommensurate (e.g., zigzag/armchair). The same conclusion was reported by Guo et al. [63], who also employed molecular dynamics simulations of a bitube oscillator. Sazonova et al. [52] showed that air drag signicantly aects the Q factor and the resonance frequency of CNTs, and that when the pressure is above about 10 Torr, resonance is completely damped out. According to Bhiladvala and Wang [64], beams of nanometer scale are smaller than the mean free path of air molecules under standard conditions, so the Stokes air drag calculation must be replaced by the appro-

priate calculation for raried gas ow. Ono et al. [65] reported that heating a MWNT in high vacuum caused desorption of contaminants from the surface of the nanotube, which in turn decreased the mechanical energy loss during vibration. The use of CNTs as eld electron emission sources seems to have great potential due to their sharp tips with nanometer-scale radii, high mechanical stiness, high aspect ratio, excellent chemical inertness and electrical conductivity [66,67]. The relatively small current capacity of a single CNT can be overcome by using CNT arrays or lms, as in at panel displays [6870]. In related work, the electron emission threshold elds for MWNT/polystyrene nanocomposite lms have been shown to depend on mechanical stress level [71]. The eect of stress on eld emission was measured using the three-point bending setup shown in Fig. 13. Heyd et al. [72] have shown analytically that a uniaxial stress applied parallel to the axis of carbon nanotubes can signicantly alter the band gap structure

Fig. 13. Three-point bending test of MWNT/polystyrene nanocomposite lm for determination of eect of mechanical stress on electron eld emission. From Poa et al. [71].

Fig. 12. Schematic drawing of a SWNT beam oscillator for modeling by classical molecular dynamics simulation. From Jiang et al. [62].

Fig. 14. Schematic of experimental setup for using electric eld emissions of MWNTs to directly observe their resonances. From Purcell et al. [74].

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

and induce a semiconductormetal transition. Minot et al. [73] also demonstrated that the CNT band gaps can be tuned by the application of mechanical strain an AFM tip was used to simultaneously vary the strain and electrostatically gate the CNT specimens. The implication from these reports is that vibratory mechanical strain could be used to generate corresponding oscillatory behavior in the band gap structure of CNTs, although no publications have been found on this subject. Further evidence of electromechanical eects in CNTs has been published by Purcell et al.[74], who demonstrated the use of an applied electrical eld to tune the resonance frequencies of CNTs by up to a factor of 10 in much the same way that a guitar string is tuned by varying the tension on the string. In addition, the eld emission made it possible to directly observe the resonances of CNTs using the experimental setup shown schematically in Fig. 14. The resonances are detected by monitoring changes created in the eld emission patterns as the corresponding applied electric eld changes. 4. Use of vibration measurements to characterize nanotube mechanical properties Both static and dynamic mechanical test methods have been employed to measure the elastic moduli of CNTs. As an example of static testing, Salvetat et al. [75] loaded nanobeams consisting of SWNT ropes at midspan with the tip of an atomic force microscope while the ends were xed at the edges of a pore in a porous ultraltration membrane. The elastic modulus was found to be on the order of 1 TPa, but the shear modulus was estimated to be only about 1 GPa. Modal vibration response measurements have been used for many years to characterize composite materials and structures at the macromechanical level [36], and the same approaches are proving to be eective for characterizing materials and the micromechanical and nanomechanical levels. As mentioned earlier, eld emission has been used to observe resonance frequencies of multiwall carbon nanotubes (MWCNTs) in the MHz range [74]. Variable frequency electric eld excitation has been used by Poncharal et al. [76] to drive cantilevered MWCNT specimens while observing resonance in a TEM, then the driving frequency at resonance was used in the frequency equation for a vibrating beam (i.e, Eqs. (1) and (2)) to back out the eective bending modulus of the nanotube (see Fig. 15). It was reported that the eective modulus decreased sharply from about 1 to 0.1 TPa as the nanotube diameter increased from 8 to 40 nm, and this was believed to be due to the development of a wavelike distortion or rippling on the compression side of the larger diameter nanotubes during large amplitude exure. A subsequent nonlinear large deformation analysis of a vibrating nanotube conrmed the presence of the rippling mode [77]. Gao et al. [59] measured the exural modulus of individual CNTs from aligned arrays by in situ electromechanical resonance in a TEM, as shown in Fig. 16. Similar resonance

Fig. 15. Variation of eective bending modulus of cantilevered MWNTs with diameter, and typical resonance curve from which the modulus is estimated. From Poncharal et al. [76].

Fig. 16. A carbon nanotube (a) in a stationary position, (b) vibrating in the rst resonance mode. The corresponding frequency response curve is shown in (c). From Gao et al. [59].

experiments in a TEM were reported by Wang et al. [78]. In related work, resonance frequencies of polymer nanotubes excited by the tip of an atomic force microscope have been employed, along with analytical models for the system, to indirectly determine the elastic modulus of the nanotubes [79]. As shown in Fig. 17, the method is based on the measurement of the resonance frequency of the AFM cantilever tip in contact with the nanotube. The resonance frequency of the AFM tip changes as it contacts the nanotube at dierent points along the nanotube length. The measured frequencies are combined with a mathematical model for the AFM tip/nanotube system, making it possible to indirectly determine the stiness of the nanotube. Thermally activated nanotube vibrations have been employed to indirectly determine the elastic moduli of the

10

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

Fig. 17. Measurement of nanotube modulus by resonant contact AFM. Resonance frequency of AFM tip changes as it contacts nanotube at dierent points along the nanotube length. From Cuenot et al. [79].

E E0 iE00 E0 1 ig;

12

Fig. 18. Intrinsic thermal vibrations of a doubly clamped SWNT detected by blurring of nanotube TEM image, as shown by arrow. From Babic et al. [83].

nanotubes by several investigators. For example, the eective Youngs moduli of MWCNTs have been measured indirectly by observing the intrinsic thermal vibration amplitudes of the nanotubes as a function of temperature and using the theory of thermal vibrations to back out the modulus value [80]. Similar methods were used later for SWNTs [81,82]. The mean square amplitude of vibration of cantilever nanotubes was found to be linearly proportional to the absolute temperature in Kelvin, and the vibration appeared in TEM images as a blurring of the TEM image that increased towards the free end of the MWCNTs. Similarly, instrinsic thermal vibrations of suspended doubly clamped SWCNTs have been observed as a blurring of the nanotube images in a SEM (see Fig. 18) that increased with increasing temperature [83]. 5. Nanotube augmentation of dynamic mechanical properties of composites There are numerous reports in the literature on the use of CNTs to augment mechanical properties of composite materials, and these reports have focused primarily on static strength, stiness and fracture toughness [2,3,12,13,84]. By contrast, the number of publications dealing with dynamic mechanical properties of CNT-reinforced composites is much smaller. Dynamic mechanical properties are usually described by the use of complex modulus notation

where E* is the complex modulus, E 0 the elastic storage modulus, E00 the loss modulus, a measure dissipation p of or damping, i the imaginary operator = 1, g the loss factor, another measure of dissipation or damping = E00 / E 0 = tan d, where tan d is the loss tangent and d is the phase lag between sinusoidally varying stress and strain. The complex modulus relates sinusoidally varying stress to sinusoidally varying strain in a linear viscoelastic material, and is typically measured by exciting small specimens in sinusoidally varying forced vibration with a dynamic mechanical analyzer (DMA) instrument or similar apparatus while simultaneously measuring the input and response, or by observing the free vibration decay after some initial excitation. The complex modulus notation, dynamic mechanical test methods and data analysis for composite materials have been described in detail by Gibson [85,86]. Such tests are often conducted over a wide range of frequencies and temperatures, and are frequently used to determine the glass transition temperature, Tg, since E 0 drops sharply and E00 peaks in the vicinity of Tg. For example, Velasco-Santos et al. [87] reported on the results of DMA testing of MWNT-reinforced polymers. It was found that the addition of only 1% by weight MWNTs in a methylethyl methacrylate copolymer led to an increase in the storage modulus by more than 200% and an increase in the glass transition temperature of more than 10 C. Finegan et al. [88] conducted DMA testing of vapor-grown carbon nanober/polypropylene (VGCF/PP) composites and compared measurements with predictions from an analytical model for predicting the complex modulus of such materials. Since the specimens were injection molded, the nanober orientations were characterized by X-ray diffraction. The resulting probability distribution for the nanober orientations was included in a modied HalpinTsai approach for the elastic modulus, then the elasticviscoelastic correspondence principle was employed to develop the model for the complex modulus. the model predicts an optimum nanober aspect ratio L/D of about 500 for maximizing the loss modulus, where L is the nanober length and D is the nanober diameter. Fig. 19 shows the

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

11

Fig. 20. Comparison of frequency response curve of baseline beam with that of CNT lm-reinforced sandwich beam. From Koratkar et al. [91].

Fig. 19. Predicted and measured storage modulus vs. ber volume fraction for VGCF/PP composites having nanober aspect ratio L/ D = 19 at a temperature of 21 C. From Finegan et al. [88].

good agreement between predicted and measured storage moduli for several nanober volume fractions and a xed aspect ratio. The prediction of an optimum aspect ratio for maximizing the loss modulus had been previously reported for discontinuous-aligned ber composites by Gibson et al. [89] and experimental data suggesting the existence of such an optimum L/D was presented by Suarez et al. [90]. Several papers have reported signicant increases in internal damping of polymer nanocomposites when CNTs are used as reinforcement. Koratkar et al. [91,92] reported that densely packed MWNT nano-lms have been embedded as interlayers within laminated piezo-silica sandwich beams to enhance both stiness and damping of the laminates. Fig. 20 shows the marked dierences in the frequency response curves of the baseline beam and the CNT lm-reinforced beam it was reported that damping increased by up to 200% and exural stiness increased by 30%. As shown in Fig. 21, increases in the loss modulus of more than 1000% were reported by Koratkar et al. [93] for SWNT/polycarbonate nanocomposites containing up to 2% by weight SWNTs. It was shown that the increase in damping was primarily due to frictional sliding at the nanotubepolymer interfaces as the strain level increases. In this case, the damping tests were done in the uniaxial tensiontension mode. This points out an advantage of nanollers in that they present a very large interfacial contact area with the matrix material by comparison with conventional reinforcements. Zhou et al. [94] fabricated and tested SWNT/epoxy beams with varying amounts of SWNT reinforcement and developed a stick-slip micromechanical model that captured the observed trends in the damping. The

Fig. 21. Variation of loss modulus with increasing strain for SWNT/ polycarbonate nanocomposite. From Koratkar et al. [93].

increased damping eect (i.e., more rapid free vibration decay) with increased CNT content is clearly shown in Fig. 22. Another key conclusion from the analytical model developed here was that the size and specic surface area of ller materials are extremely important factors which aect damping in composites, and that the very large interfacial contact area of nanotubes is a denite advantage when it comes to damping. For example, in Fig. 23, the predicted loss factors of SWNT reinforced composites are compared with those of composites reinforced with llers having smaller interfacial contact areas (Filler A is similar to carbon black and Filler B is similar to carbon whiskers). It is seen that the SWNT composite has greater damping than the other composites or the unreinforced epoxy over the entire strain range. Rajoria and Jalili [95] generated a large amount of experimental data on damping of CNT/epoxy/ steel sandwich beams having various weight percentages

12

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

of CNTs. As shown in Fig. 24, beams containing MWNT reinforcement had greater damping than those containing either SWNT reinforcement or conventional carbon ber (CF) reinforcement. In addition, the frequency dependence of CNT-reinforced beams was more pronounced than that of the other beams, and it appears that once the weight percentage of either SWNTs or MWNTs reaches 5%, there is no further improvement in damping. Up to 700% improvement in damping for 5% MWNT-reinforced beams was reported. On the other hand, corresponding changes in the storage modulus were found to be much less pronounced. Wang et al. [96] conducted DMA measurements on SWNT buckypaper/epoxy composites with SWNT loading up to 39% by weight and found that, while the storage modulus increased up to 492% with SWNT loading,

Fig. 22. Free vibration decay curves for SWNT/epoxy nanocomposite beams having dierent amounts of SWNT reinforcement. From Zhou et al. [94].

Fig. 23. Predicted composite loss factors associated with dierent llers (Filler A is similar to carbon black, Filler B is similar to carbon whiskers). From Zhou et al. [94].

Fig. 24. Measured frequency dependence of damping ratio for various composite beams. From Rajoria and Jalili [95].

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

13

the familiar peak in the tan d vs. frequency curve for the epoxy matrix material at Tg disappeared with the introduction of SWNT buckypaper reinforcement. It appears that the eect of the SWNT reinforcement on damping for these materials was not as pronounced as has been reported in other publications. During processing of polymers and polymer composites (including CNT-reinforced polymers), the rheological behavior of the melt is of great interest. Polymer melt rheology is typically characterized with a parallel plate rheometer where the polymer melt sample is placed between two parallel plates, with one plate held stationary while the other plate oscillates sinusoidally at a variable frequency [97]. From measurements of input excitation and response, the complex shear modulus of the melt, G* = G 0 + iG00 = G 0 (1 + igG) is determined, where G 0 is the shear storage modulus, G00 is the shear loss modulus, and gG is the shear loss factor, which is the same as tan d. The complex viscosity is also determined from rheometer measurements. A number of interesting observations have been reported by those using rheometers to study the rheology of various CNT-reinforced polymer melts. Kinloch et al. [98] reported that MWNTs dispersed in water exhibited non-Newtonian shear thinning (i.e., viscosity decreases with increased shear rate), which is typical of polymers but not water. AbdelGoad and Potschke [99] conducted rheological studies of MWNT/polycarbonate composites and observed a transition from liquid-like behavior to solid-like behavior at a MWNT concentration of 0.5 wt%. The signicant changes in the appearance of both the storage modulus-frequency and loss factor-frequency curves at 0.5 wt% MWNTs in polycarbonate resin are shown in Figs. 25 and 26, respectively. Some studies of the rheological behavior and the electrical conductivity of CNT-lled polycarbonate melts [100] and CNT-lled polyamide-6 melts [101] having various weight fractions of CNTs have shown that the percolation threshold for transition from electrical insulator to

Fig. 26. Variation of loss factor with frequency for various concentrations of MWNTs in polycarbonate resin. From Abdel-Goad and Potschke [99].

Fig. 25. Variation of storage modulus with frequency for various concentrations of MWNTs in polycarbonate resin. From Abdel-Goad and Potschke [99].

electrical conductor and the rheological threshold for transition from liquid-like behavior to solid-like behavior occurred at approximately the same weight percentage of CNTs (about 2%). The percolation threshold for CNTs in polymer matrix materials is so low because the high aspect ratios of CNTs make it relatively easy to form a contiguous conducting path along the CNTs in the insulating polymer matrix. McNally et al. [102] found a similar threshold of about 7.5% for both rheological and electrical measurements on a MWNT/polyethylene composite, but concluded that the similarity may be coincidental. Potschke et al. [103] conducted rheological measurements on MWNT/polycarbonate composites over the temperature range 170280 C and found that the rheological threshold changed from about 5 to 0.5 wt% by increasing the temperature from 170 to 280 C, whereas the percolation threshold for electrical conductivity was found to be about 0.5 wt%. Du et al. [104] investigated rheological and electrical properties of SWNT/polymethylmethacrylate composites and found the rheological threshold to be about 0.12 wt%, whereas the percolation threshold for electrical conductivity was about 0.39 wt%. The lower rheological threshold was explained by using the diagrams in Fig. 27. As the CNT concentration increases, the rheological threshold, or liquidsolid transition, occurs when the size of the polymer chain is somewhat larger than the separation between the CNT bundles, whereas the percolation threshold for electrical conductivity occurs at a larger concentration of CNTs when the CNT bundles are close enough to one another to form a contiguous conducting path through the insulating polymer. In the above publications, the CNT has been the reinforcing constituent in a matrix material to form a composite. Though it may be dicult to imagine, other materials can be embedded or contained inside the CNT to form a different type of composite, and the dynamic behavior of such composites has been the subject of several publications.

14

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

and that the critical ow velocity at which structural instability of the CNTs occurs could fall within the range of practical signicance, but that these eects would be mitigated when the CNT is embedded in a matrix material in a composite. 6. Vibrations in nanotube-based sensors and actuators The use of carbon nanotubes as sensors has received considerable attention [112], and in some CNT-based sensors, the CNTs are subjected to vibrations. Raman scattering from CNTs will be discussed in more detail in a later section of this paper, but sensors based on Raman spectroscopy will be discussed here. For example, the Raman frequencies of CNTs depend on pressure and strain [113,114], so Raman shifts can be the basis of nanoscale pressure or strain transducers. In addition, the Raman spectroscopy appears to be particularly sensitive to polymer transitions [115], and stress elds around defects and bers in a polymer [116]. The use of CNT probes to increase sensitivity and resolution in the atomic force microscope has been reported in several publications. The sensing element, or stylus, of a conventional atomic force microscope (AFM) is a silicon microcantilever with a sharp tip attached to its free end [117]. Dai et al. [118] rst attached CNTs to the silicon microcantilever sensing tips of an AFM and demonstrated that the CNTs were resistant to damage from tip crashes, and that their large aspect ratio led to sharper, higher resolution imaging of surface topography. Fig. 28 from Snow et al. [119] shows a CNT attached to the tip of an AFM. In this paper, an equation of motion was developed to simulate the response of the AFM tip during noncontact-mode imaging. It was found that the elastic response of long CNTs causes the CNT tip to jump into contact with the surface, and this made the sensor response unsuitable. On the other hand, for short CNTs, stable noncontact mode

Fig. 27. Schematic of SWNT/polymer networks showing how increased SWNT concentration is related to the rheological and electrical percolation thresholds. From Du et al. [104].

For example, Danailov et al. [105] developed a model based on empirical potentials and atomistic simulations for bending of a SWNT encapsulating a metal nanowire. The presence of the nanowire inside the nanotube was found to lower the vibration frequency and increase the vibration damping by comparison with the unlled nanotube. Storage of uids and gases inside CNTs has been the subject of much investigation and controversy [106,107], and transport of gases and uids inside nanotubes has been the subject of several studies [108,109]. In an attempt to understand the uidstructure interactions in CNTs, Tuzun et al. [110] developed molecular dynamics simulations of uids owing through CNTs. It was found that in a uid/CNT system, the motion of the CNT plays a critical role in the uid ow. For example, a uid owing through the CNT tends to straighten out the CNT as it exes, and simultaneously excites longitudinal vibration modes of the CNT. In related work, Yoon et al. [111] developed simple beam models for vibrating CNTs containing uids. It was concluded that the resonance frequencies depend on the uid ow velocity,

Fig. 28. Schematic of a CNT attached to AFM tip. From Snow et al. [119].

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

15

imaging was achieved by imposing a small vibration amplitude on the tip. In the tapping mode AFM, the microcantilever sensing tip is vibrated as it drags over the surface that is being imaged, so that the tip contacts the surface intermittently. CNTs attached to the silicon microcantilevers have been successfully used to improve the tapping mode AFM as well. Moloni et al. [120] found that by using CNT tips and covering the sample in a shallow layer of water, the undesirable eects of hydrodynamic damping and acoustic excitation in the conventional tapping mode AFM operation in water were greatly reduced. The nonlinear tapping mode response of an AFM microcantilever with an attached MWNT probe has been investigated both experimentally and theoretically by Lee et al. [121]. It was found that the nonlinear frequency response of such probes is signicantly dierent from that of conventional AFM microcantilever probes. Nanobalances based on CNTs have been developed for detection of the masses of extremely tiny particles. Wang et al. [78] reported on the measurement of the mass of a particle as light as 22 fg (22 1015 g). The measurement was accomplished by observing the resonance frequency shift of a CNT nanocantilever under a transmission electron microscope (TEM) after the mass is attached to the tip of the CNT, as shown in Fig. 29. Li and Chou [122] applied their molecular structural mechanics approach to study the frequency shifts in cantilevered and bridged SWNTs with attached particles. The two cases are shown schematically in Fig. 30, and results for the cantilevered case are shown in Fig. 31. It was predicted that mass sensitivity of such CNT nanobalances can reach 1021 g, and that a logarithmically linear relationship exists between

Fig. 30. (a) Cantilevered and (b) bridged nanotube resonator with an attached mass. From Li and Chou [122].

Fig. 31. Variation of fundamental frequency shifts of cantilevered CNTs having dierent lengths and attached masses. From Li and Chou [122].

Fig. 29. TEM photographs of a small particle attached at the end of a CNT nanocantilever (a) in stationary position, (b) vibrating at rst harmonic resonance frequency. From Wang et al. [78].

the resonant frequency and the attached mass when the mass is larger than about 1020 g. Absorption of gases or uids by CNTs can also cause the CNT resonance frequencies to shift, and this phenomenon has opened up a new area of research on CNT-based sensors [123126]. CNT-based actuators have attracted considerable interest in the emerging eld of nanoelectromechanical systems (NEMS). One way to manufacture NEMS is to scale down conventional silicon-based MEMS, but CNT-based NEMS have several advantages over such devices, many of which are related to dynamic mechanical behavior. For example, nanometer-sized silicon-based systems do not have the desired high-Q mechanical resonances due to dominant surface eects and thermoelastic damping, and CNT-based systems have much higher strength, stiness, thermal and electrical conductivity [127]. A rotational NEMS actuator employing a MWNT as a shaft is shown in Fig. 32. CNT-based actuators appear to be capable of withstanding much greater strains than even natural muscle [128,129]. In related work, SWNT sheets subjected to electrochemically generated gas bubbles were observed to expand pneumatically by up to 300% in the thickness direction and contract by up to 3% in the in-plane direction [130]. Extremely fast

16

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

Fig. 32. Rotational actuator with MWNT shaft (a) conceptual drawing of device, (b) SEM image of actual device. From Fennimore et al. [127].

actuation rates are possible with CNT actuators. For example, Kang and Hwang [131] used molecular dynamics simulations to demonstrate a gigahertz actuator based on MWNTs encapsulating potassium ions. An external electric eld was used to accelerate the potassium ions inside the MWNTs. Bearings are essential components of any mechanical actuator system, and Cumings and Zetti [132] have demonstrated that MWNT linear bearings are capable of controlled and reversible telescopic extension with very low friction. The use of carbon nanotubes as molecular chemical sensors has been widely researched. Kong et al. [133] demonstrated chemical sensors based on individual single-walled carbon nanotubes (SWNTs). The functionality is based upon observed dramatic change in the electrical resistance of a semiconducting SWNT upon exposure to gaseous molecules such as NO2 or NH3. At room temperatures, the nanotube sensors were found to respond faster and at much higher sensitivities than existing solid-state sensors. Sensor reversibility was achieved by slow recovery under ambient conditions or by heating to high temperatures. Bradley et al. [134] designed a carbon nanotube eld-eect transistor that could determine the nanotube depletion length. It featured metal contacts and adjacent nanotubes coated with impermeable silicon oxide, with an exposed central region of nanotubes, and sensitivity to NH3 was measured. Polyethylene-imine (PEI) and NH3 were found to produce similar responses in passivated devices and in normal, nonpassivated devices. This device design thus appeared to passivate the metalnanotube contacts while retaining chemical sensor characteristics. PEI yielded nega-

tive threshold shifts of tens of volts, even though its contact was only with the device central region. From the characterization test results, it was concluded that the length scale of the covered nanotubes in this device was comparable to the decay length of the depletion charge. He and Dai [135] chemically grafted single-strand DNA chains onto aligned carbon nanotube electrodes, to make novel aligned carbon nanotube-DNA sensors with high sensitivity and selectivity for probing complementary DNA and target DNA chains of specic sequences. The use of nanotubes as membrane chemical and pH sensors was described by Steinle et al. [136] Synthetic micropore and nanotube membranes were described, that manifest on-o (little/no versus high ion current) switch states like ligand-gated ion channels, before and after introducing a chemical stimulus. Microporous alumina and Au nanotube membranes were investigated. Chemisorption of appropriate materials was utilized to induce hydrophobic behavior in the membranes for the o state, while the on state was precipitated by partitioning a hydrophobic ionic species into the membranes.The membrane switched to the on state because at a suciently high concentration of this ionic analyte species, the pores/nanotubes ooded with water and electrolyte. The pH-sensitive membrane was prepared by attaching a hydrophobic alkyl carboxylic acid silane to the alumina membrane. Research on CNT ow sensors has continued to be of interest. Ghosh et al. [137] reported that the ow of a liquid on single-walled carbon nanotube bundles induces a voltage in the sample along the ow direction. The voltage was found to depend logarithmically on velocity, and to be sensitive to the ionic conductivity and polar nature of the liquid. It was concluded that a direct forcing of the free charge carriers in the nanotubes by the uctuating Coulombic eld of the liquid owing past the nanotubes played a dominant role in the response, and it was suggested that the phenomenon of pulsating asymmetric ratchets could probably explain the observed behavior. This research suggested the potential use of nanotubes as sensitive ow sensors and energy converters. Carbon nanotubes have also been under development as eld-eect transistor sensors [11,12]. Kim et al. [138] give a basic description of such a device as nanotubes lying on SiO2 surfaces exposed to the ambient environment. The nanotube transistors manifest hysteresis of electrical characteristics due to charge trapping by water molecules around the nanotubes, including SiO2 surface-bound water near the nanotubes. Hysteresis still occurs in vacuum because the SiO2-bound water does not completely desorb in vacuum at room temperature due to SiO2 surface chemistry. The water can be signicantly removed and the hysteresis reduced by heating in dry conditions. Dierent treatments could yield various sensor types, for instance, appropriate water-sensitive coatings would yield nanotube humidity sensors, while hysteresis-free transistors would result from passivating the devices with polymers that hydrogen bond with silanol groups on SiO2 (e.g., with

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

17

poly(methyl methacrylate) (PMMA)). Na et al. [139] examined the eect of humidity on the electrical transport properties of single-walled carbon nanotube eld eect transistors (SWNT-FETs). Water molecules were found to behave as electron donors to the nanotube and current through the p-type carbon nanotube device was found to rst decrease under modest humidity, and then start to increase as the humidity increases over 65%, probably because of the opening of electron channels. Fundamental calculations were utilized to deduce that water molecules can donate electrons to the carbon nanotube, and that a hydrogen-bonded water monolayer would be formed around the nanotube at fully covered conditions. The authors suggest that their result could provide a systematic understanding of the humidity eect on SWNT-FETs. Grimes et al. [140] described a wireless sensor network for in situ monitoring of atmospheric hydrogen concentration. A thick lm of titanium was anodized to produce titanium nanotubes. The electrical resistance of the sensor increases immensely as hydrogen concentration increases. The authors developed a network system of these sensors integrated with automated data acquisition and digitization. An apparent low level resolution of approximately 0.05 ppm, was achieved. Investigations by Chen et al. [141] claried the action of single-walled carbon nanotube eld eect transistors (FETs) on protein adsorption. The sensors utilize gold and palladium electrodes. The authors used extensive characterization to show that the electronic eects occurring at the metalnanotube contacts due to protein adsorption constitute a more signicant contribution to the electronic biosensing signal than adsorption solely along the exposed lengths of the nanotubes. Novel glucometers have been under development from CNTs. Loh et al. [142] demonstrated the superiority of CNT electrodes as glucose sensors, in a systematic comparison with boron-doped diamond (BDD) electrodes. Modication with 3,3 0 -diaminobenzidine (DAB) allowed the CNT to show selective response to glucose at a lower electrode voltage of 0.3 V, and with suppressed interference signals from ascorbic acid (LAA) and uric acid (UA). No selective response to glucose was obtained from DABmodied BDD electrode. CNTs have been applied to thermal sensing. Fung et al. [143] introduced a feasible technology for batch production of carbon nanotube (CNT) devices using ac electrophoretic technique to organize multiwalled bundles on an Si/SiO2 substrate between microfabricated electrodes, in a parallel assembly process, The fabricated devices functioned well as novel thermal sensors with low microwatts. CNTs have also been utilized as strain and pressure sensors. The basis is that the electrical properties of single-walled carbon nanotubes depend on mechanical deformation at the nanoscale thus making a carbon nanotube lm utilizable as a strain sensor at the macroscale [144]. Li and Chou [145] proposed single-walled carbon nanotube-based sensors for measuring strain and pressure at

the nanoscale, based on the principle of resonant frequency shift of a carbon nanotube resonator when it is strained by an external load. The resonant frequency shifts were shown to be linearly dependent on the applied axial strain and pressure. Reduction of tube length was found to enhance sensitivity of axial strain sensing as did reduction of tube diameter for pressure sensing. Barber et al. [146] utilized CNTs as stress sensors to assess the interfacial strength between E-glass bers and an isotactic polypropylene matrix. These sensors even proved excellent in detecting variations in the composite interfacial strength due to the glass bers being coated with dierent polymer coatings prior to composite processing. These interfacial measurements correlated well with results from conventional micro-mechanical fragmentation testing. 7. Sonication of nanotubes and nanotube-reinforced polymer resins One of the major diculties encountered during processing of carbon nanotube-reinforced or nanoparticlereinforced polymer composites is the inability to achieve a uniform dispersion of the nanotubes or nanoparticles in the liquid polymer. Nanotubes tend to cluster or agglomerate due to physical entanglements of the tubes, van der Waals forces between the carbon surfaces, and the fact that the surface energy of the nanotube clusters is thought to be less than that of the corresponding collection of individual nanotubes. Various methods of achieving uniform dispersion of nanotubes in liquids have been reviewed by Hilding et al. [147], and one of the most widely used methods is sonication. Sonication (or ultrasonication, since ultrasound at frequencies P20 kHz is typically used) of the nanotube liquid resin or nanoparticleliquid resin mixture involves the use of ultrasonic excitation of the ller/resin mixture to break up the nanotube or nanoparticle clusters through cavitation (i.e., nucleation and collapse of bubbles) in the liquid and/or exciting resonant vibrations of the clusters. Sonication of nanotubes immersed in solvents such as acetone or alcohol is also used as a preliminary step in the preparation of nanotubes for subsequent processing or for electron microscopy studies. Sonication is typically conducted either by immersing a vessel containing the nanotubeliquid mixture in a ultrasonic bath, or by submerging or partially submerging an ultrasonic wand (or tip, or horn) in the nanotubeliquid mixture. Long nanotubes can be broken up into shorter segments or otherwise damaged by sonication [148], and the amount of damage depends on the sonication time. For example, Fig. 33 shows the eect of ultrasonication time on the mean average length of MWNTs most of the length reduction occurs during the rst few minutes, then the rate of length reduction levels out. The use of sonication to separate nanotubes by length (and for MWNTs, by diameter) has been reported in several publications [149,150]. Purication of nanotube mixtures by removing

18

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

Fig. 33. Mean average MWNT length as a function of time in ultrasonic bath. From Hilding et al. [147].

carbon impurities and metal particles has been accomplished by using ultrasonically-assisted ltration [151]. The combination of tip sonication and solvent dilution with acetone led to more than 50% increase in the modulus of SWNT/epoxy nanocomposites [152]. The eect of ultrasonication on CNT/solvent or CNT/surfactant interactions has also been the subject of several investigations [153 155]. The eects of process parameters including CNT weight percent, bath and tip sonication, sonication time, surfactant type and solvent type on the dispersion of SWNTs in epoxy resin have been explored [156]. Ultrasonication without any chemical treatment has been used to disperse CNTs in polystyrene [157] and polyimide [158] resins to produce nanocomposites, and the measured mechanical properties of both materials compared well with micromechanics predictions. In related work on polymers and composites, Koda et al. [159] probed the eect of ultrasonic irradiation on copolymerization between sodium styrene sulphonate and vinylpyrrolidone at three frequencies (20, 40, and 540 kHz). Acceleration of the rate of polymerization under sonication was observed at three frequencies. Copolymer yield was found to depend on the frequency and the maximum ultrasonic eect was observed in the range between the extreme test frequencies. The rate of copolymerization increased with increasing acoustic intensity at frequencies 20 and 40 kHz while the molecular weight of the product increased with reaction time but was almost independent of the excitation frequencies. Price [160] described some examples of sonication-assisted polymerization reactions, and showed that the three-category Luche typication still applies for low molar mass systems. He concluded that the use of ultrasound allows much control on the polymer structure and resulting properties, especially molar weight distributions. Gronroos et al. [161] studied the factors involved in ultrasonication eects. Basically, industrially-acceptable, polymer degradation procedures must be capable of prescribing particular sonication conditions like irradiation power, temperature, concentration and irradiation time required for a target relative molar mass distribution. Park

et al. [162] reported a 10 orders of magnitude conductivity enhancement in addition to improvements in mechanical properties and thermal stability were also improved with the incorporation of the 0.1 vol % SWNT as reinforcement into polyimide in in situ polymerization of monomers of interest in the presence of sonication. Brown et al. [163] studied the sonication dispersion of single-walled carbon nanotubes (SWNTs) into diamine curing agents in a bid to enhance the dispersion of SWNTs in cured epoxy. Many NDE methods like light microscopy, electron microscopy, light scattering (LS), ultra small-angle X-ray scattering (USAXS), electrical conductivity, high resolution transmission electron microscopy (TEM) and Raman spectroscopy were utilized to characterize specimens. Small ropes of SWNTs (mostly under 15 nm) networked into fractal clusters were observed. In Kims work [164], high intensity ultrasound was utilized to induce mechano-chemical degradation during melt processing of polycarbonate (PC) and styrene-co-acrylonitrile (SAN). Macroradicals of PC and SAN were generated during this ultrasound assisted melt processing, thus providing a useful route to achieve in situ compatibilization for the blends of PC and SAN by their mutual coupling. Domain size was reduced and the stability of morphology was well maintained even after annealing treatment. Mechanical properties such as elongation at break and tensile strength were enhanced, revealing the utility of sonication of melt-blends in enhancing intermolecular interaction by promoting chemical bonds between dissimilar polymers without use of any compatibilizers. Liu et al. [165] observed that nanophase titania/polylactic-co-glycolic acid (PLGA) composites combine the advantages of a degradable polymer with nano-size ceramic grains to optimize physical and biological properties for bone regeneration Especially, nanophase titania being a nanostructured composite of nanometer-sized hydroxyapatite well dispersed in a mostly collagen matrix, mimics very well the size scale of constituent components of bone. They therefore studied osteoblast (bone-forming cell) adhesion and long-term functions on nanophase titania/PLGA composites, employing dierent levels of sonication. Dispersion of titania in PLGA was found to be enhanced by increasing sonication intensity and that greater osteoblast adhesion was found to correlate with improved nanophase titania dispersion in PLGA. Their study showed that PLGA composites with well dispersed nanophase titania can enhance osteoblast functions needed for better bone tissue engineering applications. In a bid to prepare 18 A Al-pillared vermiculites, de Haro et al. [166] sonicated high-charge Spanish vermiculites from Santa Olalla, and its low-charge version from Ojen reduce particle size, morphology and structural ordering. It was found that sonication caused a rise of the S-BET (specic surface area of solids by nitrogen/krypton gas adsorption according to the Brunauer, Emmett and Teller method) from the initial value 1 m2/g to 36 and 54 m2/g for the high- and lowcharge samples, respectively. X-ray diraction analysis

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

19

showed the crystal structure of both materials to be practically unaected by sonication, but the 001 diraction yielded particle thickness values of 185 and 260 A for the low- and high-charged specimens, respectively. The sonicated Santa Olalla sample (high-charge) treated with Al13 solution shows a d(001) powder diraction peak at 14 angstrom, whereas the sonicated Ojen sample (low-charge) shows peaks at 14 and 18 A. The 18 A diraction peak is preserved at high temperatures (550 C) and it is responsible of an increase of surface area and microporosity. The ultrasound treatment of low-charge vermiculite from Ojen is a new method for obtaining 18 A Al-pillared vermiculites. Sonication using H2O2 in the medium broadly produced signicant delamination and a decrease of both particle size and layer charge thus yielding Al-polymer intercalation. Baxter et al. [167] explored the inuence of high-intensity ultrasonication on the molecular weight and degree of acetylation of chitosan (processed, more water-soluble form of chitin, a polysaccharide starch component of the exoskeletons of arthropods (ants, shellsh, spiders, etc.) and the walls of many molds, yeast and fungi). A 1% (w/v) chitosan in 1% (v/v) aqueous acetic acid was sonicated for varying time durations at varying power levels. Results showed that degree of deacetylation of chitosan molecules was insensitive to both power level and sonication time. However, increased sonication time exponentially decreased the viscosity of the samples. The Schmid polymer degradation model was used to analyze the molecular weight decay as a function of sonication time. Their results suggested a rst order dependence of reaction rate ultrasonic intensity, and indicated that high-intensity ultrasonication is eective in reducing chitosan molar weight even as degree of acetylation remains same. The typical dimensions of CNTs turn out to be in the same range as those of DNA molecules, and in recent years there has been an explosion in the literature dealing with the possible interactions between DNA molecules and CNTs [168171]. In related work, it has also been found that various proteins easily absorb on the sidewalls of acid-oxidized SWNTs, making it possible to use CNTs as intracellular protein transporters [172]. With regard to a connection between this nano-bio research and vibrations of CNTs, Zheng et al. [173] found that bundled

SWNTs can be eectively dispersed in water when they are sonicated in the presence of single-stranded DNA (ssDNA). Their molecular modeling suggests that the sonication-assisted dispersion and separation of CNTs is enhanced when the ssDNA binds to the CNT by helically wrapping the CNT surface as shown in Fig. 34. Such nanohybrids of man-made materials with biological materials open up the possibility of unlimited future developments in biotechnology. 8. Raman scattering from nanotubes Raman spectroscopy is a widely used technique in materials science which involves the excitation of a sample with intense monochromatic light (typically laser light) followed by observations of the scattering of the incident radiation [174]. The scattered radiation occurs at frequencies corresponding to dierent vibrational modes of the molecules in the sample (the Raman spectrum), and which are dierent from the frequency of the incident radiation. The vibrational spectra of molecules are compared with other well known frequency ranges in the electromagnetic spectrum in Fig. 35 [175]. Since the dimensions of CNTs fall in the same range as those of molecules, Raman spectroscopy is used extensively to study vibrational modes of the nanotubes. The basic concepts and characteristics of Raman spectra from carbon nanotubes have been summarized by Ecklund et al. [176], Jorio et al. [177] and Dresselhaus et al. [178]. A review article on Raman spectroscopy of individual SWNTs has been published by Dresselhaus et al. [179]. The radial breathing mode (RBM) and the G-band mode, which are two of many possible modes of vibration of the atoms in a nanotube, are shown schematically in Fig. 36. Raman intensity of each MWNT vibration mode has been found to vary with polarization direction, and the relationship between Raman intensity and polarization direction can be represented by analytical functions [180]. The use of analytical functions can avoid burdensome numerical calculation and also gives clear relations between Raman intensity and polarization direction. Applying these analytical functions, the orientation of the SWNT through

Fig. 34. Simulation of helical wrapping of carbon nanotube by single strand DNA molecule. From Zheng et al. [173].

Fig. 35. The electromagnetic spectrum. From Fadini and Schnepel [175].

20

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

Fig. 36. Schematic representations of the radial breathing mode (RBM) and the G-band mode for a zigzag CNT. From Jorio et al. [177].

polarized Raman spectra can be easily veried [181]. This also provides a useful way of evaluating the application of SWNTs. Besides, the Raman-active mode of SWNTs can be distinguished through the polarized Raman spectrum. This gives analytical functions based on the symmetry of SWNT vibration modes. Since nanotubes are of one-dimensional symmetry, the use of polarized light parallel or perpendicular to the tube axis can identify the symmetry for the specic Raman signal and give information about the low dimensionality of the nanotubes [182]. Yet, no measurement on polarized Raman scattering has been reported for carbon nanotubes because of the diculty in getting aligned SWNTs with mono-dispersed tube size. Sun et al. [183] presented the polarization dependence of Raman intensity for mono-dispersed SWNTs in one-dimensional channels of porous crystallites. Vibration symmetries of the vibrational modes are directly obtained from the polarized Raman spectra measured using dierent polarization congurations. The polarization eects of the Raman scattering are commonly observed in low-dimensional materials such as SWNTs. The applied light that is polarized parallel or perpendicular to the tube axis will show the low dimensionality of the nanotubes. Raman spectra analysis has proven to be a unique characterization tool for SWNTs [184186]. Rao et al. [195] showed that Raman signals from (m, n) SWNTs not only consist of the graphite-oriented E2g mode, which occurs around 15501600 cm1, but also contain a strong low frequency A1g-active mode, known as a radial breathing mode. Saito et al. [196] presented a detailed calculation on the Raman intensity of a SWNT as a function of the polarization of light of the carbon nanotube using non-resonant bond-polarization theory [187]. The Raman intensity of each vibration mode as a function of the SWNT sample orientation was not in analytical form. An analytical function was developed for identifying the Raman modes of SWNTs based on the symmetry of the vibration modes [187]. Polarized Raman experiments have also been performed on aligned SWNT samples [188]. Another technique which is enhanced by

polarization is surface-enhanced spectroscopy (SERS), which has been used to investigate MWNT properties [189]. The standard SERS technique may induce depolarizing eects which lead to ineciencies on performing polarized Raman spectroscopy [190]. Some papers have demonstrated that SERS allows one to select a single class of SWNTs [191193]. Local SERS Raman spectroscopy has been demonstrated to be a powerful tool for performing polarized Raman experiments at a nanooptical scale [182]. The results obtained from a single class of nanotubes shows that polarized Raman spectroscopy is possible when a single hot site of a SERS substrate is selected and unambiguous assignment of the modes is possible. It is clear that the polarized Raman signals have not only been used in the G-band domain of SWNTs but also can be widely used for other samples. Raman scattering experiments have been used to study the dependence of SWNT breathing mode resonances on the nanotube diameter [185]. Similar results have shown that the diameter of a SWCNT or the inner diameter of a MWNT can be calculated from the RBM frequencies [194]. In particular, the Raman scattering dependent on the excitation energy leads to a diameter-selective signal [195], so that the electronic structure of a nanotube having a certain diameter can be directly evaluated [196]. The diameter of nanotubes can be easily estimated from the peak position of the signal associated with the specic vibration of tube diameter [196] (radial breathing mode, RBM). Synthesis of nanotube samples with dierent diameter distributions is possible by using dierent types and concentrations of catalysts [198]. The electron and phonon nanotube properties are signicantly dependent on their symmetry and diameter [197]. Resonant Raman scattering is one of the most useful techniques for the characterization of SWNTs, since it can probe both their vibrational spectra and electronic structures via the resonant Raman eect [196]. This eect has been used to study samples with dierent diameter distributions, showing the diameter dependence of the electronic properties of these samples [199]. Therefore, the study of the nanotubes diameters by Raman scattering can provide complementary information for the direct diameter measurements by TEM, STM and other techniques. Most of the Raman studies about the diameter distribution in SWNT samples have been focused on the behavior of the radial breathing mode (RBM) [200 202]. Because the frequency of a given isolated tube is inversely proportional to the tube diameter, it is possible to estimate the diameter distribution from the spectral shape and position of the RBM Raman band. It is well known that the resonant Raman eect is very strong in SWNTs, due to the sharp Van Hove singularities in the 1D electronic density of states (DOS) [184], and the nanotubes that contribute most strongly to the Raman scattering for a given excitation wavelength are those which are in resonance with the incident or scattered light. There is another procedure for determining the diameter distribution from the Raman spectra of carbon nanotubes that is

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

21

associated with the resonant behavior of the tangential modes of metallic nanotubes. The Gaussian distribution of diameters can be obtained from the laser energy region in which the metallic modes are enhanced in the Raman spectra [203,204]. Corio et al. [205] investigated dierent samples of carbon nanotubes synthesized with dierent combinations of catalysts to verify the dependence of the resonant Raman behavior of the metallic modes on the diameter distribution, and to estimate the diameter distribution within each sample. The radial breathing mode (RBM) of Raman scattering can be used to distinguish SWNTs from DWNTs [206,207]. Similar to that of SWNTs, the Raman spectra of DWNTs have been experimentally shown to be a resonant process which displays the radial breathing mode (RBM) and the splitting tangential mode (TM). It has been shown that the peak positions of the Raman bands of DWNTs are strongly dependent on the laser energies due to resonant Raman vibration [208]. CNTs with all kinds of chiralities can be synthesized, and a DWNT can have any combination of chiralities of the inner and outer tubes. Dierent RBM peaks of dierent tube diameters have dierent Raman frequency shifts [209]. Raman spectroscopy has proven to be an ecient tool for probing the interfacial properties of SWNT/epoxy nanocomposites. It has been shown that Raman spectroscopy can monitor the load transfer from the epoxy matrix to the nanotubes with applied mechanical [210] or thermal [211] loads. Loads applied on the composites result in strains of the nanotubes that can be detected as changes of the carbon vibration frequencies. The carbon vibrations that involve CC bond-stretching motions with eigenvectors being tangential to the nanotube surface are particularly sensitive to the nanotube strain [212]. The highest frequency tangential carbon vibrations in SWNTs give rise to a graphite-like G-band around 15001590 cm1 in the Raman spectra. The G-band usually exhibits a doublet structure with higher frequency component G+ corresponding to the longitudinal carbon vibrations [213,214]. The second band G involves transversal carbon vibrations. The G+-band vibrations have been studied and the correspondence between the change of the vibration frequencies and strain on the SWNT has been established [210212]. SWNTs in the nanocomposite act as unidirectional strain sensors that can be read when the incident light polarization has a component along the nanotube axis [210]. Siochi et al. [215] investigated SWNT/polyimide composites and characterized the SWNT orientations by comparison of Raman spectra parallel and perpendicular to the ber axis and high resolution scanning electron microscopy (HRSEM) on cross sections. Haggenmueller et al. applied polarized Raman spectroscopy on PMMA/SWNT [216] and PE/SWNT [217] bers and used the Raman intensity of the radial breathing mode as a function of the angle with respect to the incident polarization axis [216] or the intensity ratio parallel/perpendicular [217] to determine the full

width at half maximum (FWHM) of the distribution function. Bhattacharyya et al. [218] determined the SWNT Hermans orientation factor of PP/SWNT composites from polarized Raman spectroscopy by using the peak intensity of the tangential mode at dierent angles between ber axis and polarization direction. Sreekumar et al. [219] also conducted the same procedure for polyacrylonitrile/SWNT composite bers. Changes in the intensities of RBMs have been shown to be related to induced uniaxial strain in SWNT-reinforced epoxy composites, and it has also been shown nanotube chirality can be identied from RBM data [220]. Raman spectroscopy experiments have been used to measure the degree of alignment of SWNTs in a polymer matrix [221], and to study the stress transfer between SWNTs and silicone-based elastomers and reorientation of nanotubes under strain [222]. Deformation micromechanics of polymer matrix nanocomposites reinforced with SWNTs and MWNTs have been studied using Raman spectroscopy [223]. Micro-Raman spectroscopy has been used to measure the compressive deformation of carbon nanotubes embedded in a polymer matrix in order to indirectly determine the Youngs modulus of the nanotubes [211]. Based on experiments and molecular dynamics simulations, reductions in radial and tangential band Raman frequencies of SWNTs with increasing temperature have been attributed to thermal expansion in the radial direction, softening of the CC intratubular bonds and softening of the van der Waals interactions in SWNT bundles [224]. Glass transition temperatures and secondary transitions in polymers have been measured by observing Raman spectral shifts of SWNTs embedded in the polymers [225]. 9. Absorption of high frequency waves by nanotubes Experiments with microwave excitation of carbon nanotubes have yielded several very interesting conclusions related to heating of the nanotubes. This has important applications for processing of nanotube-reinforced polymer composites, since microwave radiation is often used for curing of thermosetting polymer matrix materials in composites [226228] and because carbon nanotubes are often grown by using a microwave-enhanced chemical vapor deposition process [229232]. Wadhawan et al. [233] studied the eects of microwave irradiation on both unpuried and puried iron-catalyzed high-pressure disproportionation (HiPco)-grown singlewalled carbon nanotubes (SWNTs) in ultrahigh vacuum. They found that with microwave irradiation, unpuried HiPco SWNTs rapidly attained high temperatures of approximately 1850 C, yielding hydrogen, steam, carbon monoxide, carbon dioxide and methane gases. The iron (Fe) catalyst nanoparticles were observed to melt and coalesce into larger crystallites approximately four times their original diameter. In contrast, carbon black and puried HiPco SWNTs only attained temperatures of 500650 C. The authors concluded that the iron catalysts were

22

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

responsible for the much-enhanced heating of unpuried HiPco SWNTs. It had been known that CNT/crystalline Fe nanocomposites have excellent microwave-absorption characteristics. Che et al. [234] showed that this absorption property actually comes from the connement of crystalline Fe in the carbon nanoshells, resulting mainly from magnetic rather than electric eects. The complex permittivity and permeability were found to depend both on the shape and phase of the CNT/Fe nanocapsulates. Roberts et al. [235] used a resonant microwave cavity operating in the TM010 mode to investigate the microwave susceptibility of single walled carbon nanotubes (SWNT) blended in polymer matricies, using probe signal frequencies of 9.8, 11.4, and 35.93 GHz. Blends of 911% SWNT in polycarbonate were found to be quite eective in shielding the electric vector of electromagnetic waves, thus protecting devices from microwaves at these frequencies Both the electric vector and the magnetic vectors were probed by the process to determine the nature of coupling between the SWNTs and the applied elds. Relative microwave absorption for dierent blends of the SWNTs with polycarbonates was quantitatively documented. Microwave response properties of the ZnO nanowirepolyester composites were examined by Chen et al. [236]. Enhancement of the maximum absorption was noted as the concentration of the nanowires increased in the composites, and strong microwave absorption was observed in the X band. The low values of complex permittivity and dissipation of the pure nanowires observed identied pure nanowires as low-loss materials for microwave absorption in the X band. It was concluded that the strong microwave absorption was related to interfacial multipoles at the interface between the polyester and the ZnO nanowires, and to a high surface-to-volume ratio and a similar shape of the nanowires to antenna. Liu et al. [237] theoretically and experimentally investigated the eective electromagnetic properties of planar composites at microwave frequencies, with randomly and periodically embedded conductive bers of varying volume concentrations. Transmission coecient and eective permittivity obtained by analysis using the free space method, and numerically by the FEM agreed well. Lower microwave permittivity with broader peak response was observed for composites with randomly distributed bers as opposed to that with periodically distributed bers. Strong absorption of 2.45 GHz microwaves has been observed to cause ignition, burning, outgassing, intense mechanical motions and reconstruction of unpuried SWNTs [238]. In this case, the SWNTs were produced by the HiPco process [239]. For short cycles of microwave exposure, the SWNTs appeared to expand to twice their original volume, but contracted back to near their original size when the microwave eld was turned o. Repeated cycling of the microwave eld resulted in reduced motion, and it is believed that the reduced motion was caused by cross-linking or welding of the nanotubes. In related work, it was found that unpuried HiPco SWNTs heated up to

signicantly higher temperatures than puried SWNTs or carbon black particles, and the eect was attributed to greater microwave absorption by a higher concentration of Fe catalyst nanoparticles which were present in the unpuried nanotubes [240]. Experiments to date seem to have been focused on standard microwave generators operating at 2.45 GHz, which is the resonance frequency of a water molecule, and it appears that there have been no systematic attempts to nd the optimum excitation frequency or frequency range for exciting CNT vibrations. Bright uorescence of SWNTs has been observed under laser excitation in the 8001600 nm range [241]. In a related medical application which appears to have great promise for cancer therapy, near-infrared (NIR) light has been used to excite SWNTs within living biological cells [242]. Although biological systems are transparent to NIR light having wave lengths of 7001200 nm, SWNTs exhibit strong optical absorbance of such radiation, which causes intense local heating of the SWNTs, which in turn kills the cancerous cells. Thus, it appears that by injecting SWNTs into cancerous cells and then subjecting the cells to NIR, it is possible to selectively target and destroy the cancerous cells without harming the normal cells. A similar approach to cancer therapy which involves the use of NIR to excite gold-coated dielectric silica nanoshells has been reported [243]. In addition, the nanoshells can be engineered to both scatter and absorb light, so that the technique oers integrated imaging and therapy. The references cited previously in this section and in the section on Raman scattering have reported mainly on experimental observations of the interaction of high frequency waves with CNTs and their composites, and there seems to be a need for more work in modeling and simulation of CNT/high frequency wave interactions such as dispersion. This is particularly important, since continuum models typically break down for high frequencies where the wave lengths are small enough to be on the order of the characteristic structural dimension. However, the usefulness of continuum theories can often be extended by adding features that reect the nature of the system being modeled. For example, Wang and Hu [244] compared molecular dynamics simulations and continuum models in the study of dispersion of exural waves propagating in SWNTs, and showed that the Euler beam theory agrees well with molecular dynamics simulation in describing dispersion of waves having small wave numbers (i.e., long wave lengths), but the Timoshenko beam model is better for large wave numbers. Sun and Zhang [245] and Zhang and Sun [246] have reported on the development of a semi-continuum nanoplate model which was used to study dispersion of exural waves in nanostructured materials. The semi-continuum model takes into account the discrete nature of the material in the thickness direction while using the continuum description of the in-plane properties. More innovations of this type may allow further extensions of the range of usefulness of continuum theories in modeling CNT/wave interactions.

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128

23

10. Concluding remarks The knowledge base associated with the vibrations of CNTs and their composites has expanded greatly in recent years, as the number of applications involving dynamic behavior of nanometer-scale systems continues to grow. The explosive growth in the literature on this subject is such that the vast majority of the papers reviewed here were published within the last ve years, with most of the publications appearing in physics journals. Developments in the simulation of vibrating CNTs span the spectrum from the continuum mechanics-based models by engineers to the molecular dynamics-based models of physicists and chemists. The interdisciplinary nature of the subject is brought out in hybrid atomistic/continuum approaches such as molecular structural mechanics. The use of new nanoscale measurement techniques such as atomic force microscopy, transmission electron microscopy and Raman spectroscopy has led to a better understanding of the dynamic properties of CNTs and their composites. On the other hand, the unique properties of vibrating CNTs are being employed to develop new nanomechanical devices and instruments such as sensors, actuators, high frequency resonators and oscillators, and eld emission devices. CNT vibrations induced by the use of ultrasonics, microwaves or other high frequency wave propagation are widely used for improving the dispersion of nanotubes in resin matrix materials of nanocomposites, yet much work remains if the process is to be optimized. Vibrating CNTs are even beginning to have an impact in some of the emerging areas of nanobio technologies, such hybrid biological/man-made materials and cancer treatments. In conclusion, the seemingly narrow subject of CNT vibrations has had an amazingly rapid, widespread, and interdisciplinary impact in science and engineering in recent years, but much analytical and experimental work remains if we are to take full advantage of the unique dynamic behavior of these materials. Acknowledgment The support of National Science Foundation Grant No. CMS-03-19767 for Development of Instrumentation for Measurement of Microscopic Dynamic Motions in Physical Systems is gratefully acknowledged. References
[1] Iijima S. Helical microtubules of graphitic carbon. Nature 1991;354: 568. [2] Qian D, Wagner JG, Liu WK, Yu MF, Ruo RS. Mechanics of carbon nanotubes. Appl Mech Rev 2002;55(6):495533. [3] Thostenson ET, Ren Z, Chou TW. Advances in the science and technology of carbon nanotubes and their composites: a review. Compos Sci Technol 2001;61:1899912. [4] Dai HJ. Carbon nanotubes:opportunities and challenges. Surface Sci 2002;500:21841. [5] Liu J, Fan SS, Dai HJ. Recent advances in methods of forming carbon nanotubes. MRS Bull 2004;29(4):24450.

[6] Collins PG, Avouris P. Nanotubes for electronics. Sci Am 2000;283:629. [7] Avouris P, Appenzeller J, Martel R, Wind SJ. Carbon nanotube electronics. Proc IEEE 2003;91(11):177284. [8] Tsukagoshi K, Yoneya N, Uryu S, Aoyagi Y, Kanda A, Ootuka Y, et al. Carbon nanotube devices for electronics. Physica B 2002;323(14):10714. [9] Baughman RH, Zakhidov AA, de Heer WA. Carbon nanotubes the route toward applications. Science 2002;297:78792. [10] Choi WB, Bae E, Kang D, Chae S, Cheong B, Ko J, et al. Aligned carbon nanotubes for nanoelectronics. Nanotechnology 2004;15: S5126. [11] Gates TS. Modeling and characterization of nanostructured materials. Compos Sci Technol 2003;63(11):1497724 [special issue]. [12] Lau KT, Chipara M, Ling HY, Hui D. On the eective elastic moduli of carbon nanotubes for nanocomposite structures. Composites B 2004;35:95101. [13] Valavala PK, Odegard GM. Modeling techniques for determination of mechanical properties of polymer nanocomposites. Rev Adv Mater Sci 2005;9:3444. [14] Dikin DA, Chen X, Ding W, Wagner G, Ruo RS. Resonance vibration of amorphous SiO2 nanowires driven by mechanical or electrical eld excitation. J Appl Phys 2003;93(1):22630. [15] Blevins RD. Formulas for natural frequency and mode shape. New York: Van Nostrand Reinhold; 1979. [16] Ru CQ. Intrinsic vibration of multiwalled carbon nanotubes. Int J Nonlinear Sci Numer Simul 2002;3(34):735. [17] Yoon J, Ru CQ, Mioduchowski A. Noncoaxial resonance of an isolated multiwall carbon nanotube. Phys Rev B 2002;66: 233402. [18] Zhang Y, Liu G, Han X. Transverse vibrations of double-walled carbon nanotubes under compressive axial load. Phys Lett A 2005;340:25866. [19] Timoshenko S. Vibration problems in engineering. New York: Wiley; 1974. [20] Yoon J, Ru CQ, Mioduchowski A. Timoshenko-beam eects on transverse wave propagation in carbon nanotubes. Composites B 2004;35:8793. [21] Yoon J, Ru CQ, Mioduchowski A. Terahertz vibration of short carbon nanotubes modeled as Timoshenko beams. J Appl Mech 2005;72(1):1017. [22] Yoon J, Ru CQ, Mioduchowski A. Vibration of an embedded multiwall carbon nanotube. Compos Sci Technol 2003;63: 153342. [23] Pipes RB, Frankland S-J, Hubert P, Saether E. Self-consistent properties of the SWCN and hexagonal arrays as composite reinforcements. Compos Sci Technol 2003;63(10):134958. [24] Pipes RB, Hubert P, Kamrazzaman M. Harmonic properties of carbon nanotubes and their hexagonal arrays. Presented at symposium honoring Prof. C.T. Sun, 18th annual technical conference of the American Society for Composites, Gainesville, FL, October 19 22; 2003. [25] Wang CY, Ru CQ, Mioduchowski A. Free vibration of multiwall nanotubes. J Appl Phys 2005;97:114323. [26] Flugge W. Stresses in shells. Berlin: Springer; 1960. [27] Wang CY, Ru C, Mioduchowski A. Applicability and limitations of simplied elastic shell equations for carbon nanotubes. J Appl Mech 2004;71:62231. [28] Mahan GD. Oscillations of a thin hollow cylinder: carbon nanotubes. Phys Rev B 2002;65:235402. [29] Wang XY, Wang X. Numerical simulations for bending modulus of carbon nanotubes and some explanations for experiment. Composites B 2004;35:7986. [30] Odegard GM, Gates TS, Nicholson LM, Wise KE. Equivalent continuum modeling of nano-structured materials. Compos Sci Technol 2002;62:186980.

24

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128 [54] Zheng Q, Jiang Q. Multiwalled carbon nanotubes as gigahertz oscillators. Phys Rev Lett 2002;88(4):045503-13-3. [55] Zheng Q, Liu JZ, Jiang Q. Excess van der Waals interaction energy of a multiwalled carbon nanotube with an extruded core and the induced core oscillation. Phys Rev B 2002;65(24):245409-19-6. [56] Liu P, Zhang YW, Lu C. Oscillatory behavior of gigahertz oscillators based on multiwalled carbon nanotubes. J Appl Phys 2005;98:014301. [57] Schwab K. Spring constant and damping constant tuning of nanomechanical resonators using a single-electron transistor. Appl Phys Lett 2002;80(7):12768. [58] Reulet B, Kasumov AY, Kociak M, Deblock R, Khodos II, Gorbatov YB, et al. Acoustoelectric eects in carbon nanotubes. Phys Rev Lett 2000;85(13):282932. [59] Gao R, Wang ZL, Bai Z, de Heer W, Dai L, Gao M. Nanomechanics of individual carbon nanotubes from pyrolytically grown arrays. Phys Rev Lett 2000;85(3):6225. [60] Davis JF, Bronikowski M, Choi D, Epp L, Hoenk M, Hoppe D, et al. High-Q mechanical resonator arrays based on carbon nanotubes. In: Third IEEE conference on nanotechnology, vol. 2; 2003. p. 6358. [61] Prokudina NA, Shishchenko ER, Joo O-S, Hyung K-H, Han S-H. A carbon nanotube lm as a radio frequency lter. Carbon 2005;43:18159. [62] Jiang H, Yu M-F, Huang Y. Intrinsic energy loss mechanisms in a cantilevered carbon nanotube beam oscillator. Phys Rev Lett 2004;93(18):185501. [63] Guo W, Guo Y, Gao H, Zheng Q, Zhong W. Energy dissipation in gigahertz oscillators from multiwalled carbon nanotubes. Phys Rev Lett 2003;91(12):125501. [64] Bhiladvala RB, Wang ZJ. Eect of uids on the Q factor and resonance frequency of oscillating micrometer and nanometer beams. Phys Rev E 2004;69:036307. [65] Ono T, Sugimoto S, Miyashita H, Esashi M. Mechanical energy dissipation of multiwalled carbon nanotube in ultrahigh vacuum. Japanese J Appl Phys 2 2003;42(6B):L6834. [66] de Heer WA, Chatelain A, Ugarte D. A carbon nanotube eld emission electron source. Science 1995;270:117980. [67] Saito Y, Hamaguchi K, Hata K, Uchida K, Tasaka Y, Ikazaki F, et al. Conical beams from open nanotubes. Nature 1997;389:5545. [68] Choi WB, Chung DS, Kang JH, Kim HY, Jin YW, Han IT, et al. Fully sealed, high brightness carbon nanotube eld emission display. Appl Phys Lett 1999;75(20):312931. [69] Hsu DSY, Shaw JL. Robust and regenerable integrally gated carbon nanotube eld emitter arrays. J Appl Phys 2005;98:014314. [70] Bonard J-M, Salvetat J-P, Stockli T, de Heer WA. Field emission from single-wall nanotube lms. Appl Phys Lett 1998;73(7): 91820. [71] Poa CHP, Smith RC, Silva SRP, Sun CQ. Inuence of mechanical stress on electron eld emission of multiwalled carbon nanotubepolymer composites. J Vac Sci Technol B 2005;23(2):698701. [72] Heyd R, Charlier A, McRae E. Uniaxial stress eects on the electronic properties of carbon nanotubes. Phys Rev B 1997;55(11): 68204. [73] Minot ED, Yaish Y, Sazanova V, Park J-Y, Brink M, McEuen P. Tuning carbon nanotube band gaps with strain. Phys Rev Lett 2003;90(15):156401-11-4. [74] Purcell ST, Vincent P, Journet C, Binh VT. Tuning of nanotube mechanical resonances by electric eld pulling. Phys Rev Lett 2002;89(27):276103-13-4. [75] Salvetat J-P, Briggs GAD, Bonard J-M, Bacsa RR, Kulik AJ, Stockli T, et al. Elastic and shear moduli of single-walled carbon nanotube ropes. Phys Rev Lett 1999;82(5):9447. [76] Poncharal P, Wang ZL, Ugarte D, de Heer WA. Electrostatic deections and electromechanical resonances of carbon nanotubes. Science 1999;283(5407):15136. [77] Liu JZ, Zheng Q, Jiang Q. Eect of a rippling mode on resonances of carbon nanotubes. Phys Rev Lett 2001;86(21):48436.

[31] Gates TS, Odegard GM, Frankland SJV, Clancy TC. Computational materials: multi-scale modeling and simulation of nanostructured materials. Compos Sci Technol 2005;65(1516):241634. [32] Bodily BH, Sun CT. Structural and equivalent continuum properties of single-walled carbon nanotubes. Int J Mater Product Technol 2003;18(4/5/6):38197. [33] Li C, Chou T-W. A structural mechanics approach for the analysis of carbon nanotubes. Int J Solids Struct 2003;40:248799. [34] Li C, Chou T-W. Single-walled nanotubes as ultrahigh frequency nanomechanical resonators. Phys Rev B 2003;68:073405. [35] Li C, Chou T-W. Vibrational behaviors of multiwalled-carbonnanotube-based nanomechanical resonators. Appl Phys Lett 2004;84(1):1213. [36] Gibson RF. Modal vibration response measurements for characterization of composite materials and structures. Compos Sci Technol 2000;60:276980. [37] Kulkarni A, Gibson RF. Nondestructive characterization of eects of temperature and moisture on elastic moduli of vinyl ester resin and E-glass/vinylester composite. In: Proceedings of the American Society for Composites 18th annual technical conference, October 2022, Gainesville, FL; 2003 [on CD-ROM]. [38] Zhou LG, Shi SQ. Molecular dynamics simulations on tensile mechanical properties of single-walled carbon nanotubes with and without hydrogen storage. Comput Mater Sci 2002;23:16674. [39] Dereli G, Ozdogan C. Structural stability and energetics of singlewalled carbon nanotubes under uniaxial strain. Phys Rev B 2003;67(3):035416. [40] Sanchez-Portal D, Artacho E, JSoler JM. Ab initio structural, elastic and vibrational properties of carbon nanotubes. Phys Rev B 1999;59(19):1267888. [41] Qian D, Dickey EC, Andrews R, Rantell T. Load transfer and deformation mechanisms in carbon nanotube-polystyrene composites. Appl Phys Lett 2000;76(20):286870. [42] Fisher FT, Bradshaw RD, Brinson LC. Fiber waviness in nanotubereinforced polymer composites I: modulus predictions using eective nanotube properties. Compos Sci Technol 2003;63: 1689703. [43] Bradshaw RD, Fisher FT, Brinson LC. Fiber waviness in nanotubereinforced polymer composites II: modeling via numerical approximation of the dilute strain concentration tensor. Compos Sci Technol 2003;63:170522. [44] Gibson RF, Anumandla V. A comprehensive closed form micromechanics model for estimating the elastic modulus of nanotubereinforced composites. Composites A [in press]. Available online 2 November 2005. [45] Chan WS, Wang JS. Inuence of ber waviness on the structural response of composite laminates. J Thermoplastic Compos Mater 1994;7:24360. [46] Hsaio HM, Daniel IM. Elastic properties of composites with ber waviness. Composites A 1996;27(10):93141. [47] Christensen RM, Waals FM. Eective stiness of randomly oriented ber composites. J Compos Mater 1972;6:51832. [48] Andrews R, Jacques D, Minot M, Rantell T. Fabrication of carbon multiwall nanotube/polymer composites by shear mixing. Macromol Mater Eng 2002;287:395403. [49] Lee BK, Wilson JF. Free vibrations of arches with variable curvature. J Sound Vibr 1990;136(1):7589. [50] Kim N-I, Seo K-J, Kim M-Y. Free vibration and special stability of non-symmetric thin-walled curved beams with variable curvatures. Int J Solids Struct 2003;40:310728. [51] Pelesko JA, Bernstein DH. Modeling MEMS and NEMS. Boca Raton (FL): Chapman & Hall/CRC; 2003. [52] Sazonova V, Yaish Y, Ustenel H, Roundy D, Arias TA, McEuen P. A tunable carbon nanotube electromechanical oscillator. Nature 2004;431:2847. [53] Kwon YW, Manthena C, Oh JJ, Srivasta D. Vibrational characteristics of carbon nanotubes as nanomechanical resonators. J Nanosci Nanotechnol 2005;5(5):70312.

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128 [78] Wang ZL, Poncharal WA, de Heer WA. Measuring physical and mechanical properties of individual carbon nanotubes by in situ TEM. J Phys Chem Solids 2000;61:102530. [79] Cuenot S, Fretigny C, Demoustier-Champagne S, Nysten B. Measurement of elastic modulus of nanotubes by resonant contact atomic force microscopy. J Appl Phys 2003;93(9):56505. [80] Treacy MMJ, Ebbesen TW, Gibson JM. Exceptionally high Youngs modulus observed for individual carbon nanotubes. Nature 1996;381(6584):67880. [81] Krishnan A, Dujardin E, Ebbesen TW, Yianilos PN, Treacy MMJ. Youngs modulus of single-walled nanotubes. Phys Rev B 1998;58(20):140139. [82] Treacy MMJ, Krishnan A, Yianilos PN. Inferring physical parameters from images of vibrating carbon nanotubes. Microsc Microanal 2000;6(4):31723. [83] Babic B, Furer J, Sahoo JS, Farhangfar S, Schoenberger C. Intrinsic thermal vibrations of suspended doubly clamped single-wall carbon nanotubes. Nanoletters 2003;3(11):157780. [84] Gojny FH, Wichmann MHG, Kopke U, Fiedler B, Schulte K. Carbon nanotube reinforced epoxy composites: enhanced stiness and fracture toughness at low nanotube content. Compos Sci Technol 2004;64:236371. [85] Gibson RF. Principles of composite material mechanics. New York: McGraw-Hill; 1994. [86] Gibson RF. Vibration test methods for dynamic mechanical property characterization. In: Jenkins CH, editor. Chapter 14 in Manual on experimental methods of mechanical testing of composites. Society for Experimental Mechanics; 1998. [87] Velasco-Santos C, Martinez-Hernandez AL, Fisher F, Ruo R, Castano VM. Dynamic mechanical and thermal analysis of carbon nanotube-methyl-ethyl-methacrylate nanocomposites. J Phys D 2003;36:14238. [88] Finegan IC, Tibbetts GG, Gibson RF. Modeling and characterization of damping in carbon nanober/polypropylene composites. Compos Sci Technol 2003;63:162935. [89] Gibson RF, Chaturvedi SK, Sun CT. Complex moduli of aligned discontinuous bre reinforced polymer composites. J Mater Sci 1982;17:3499509. [90] Suarez SA, Gibson RF, Sun CT, Chaturvedi SK. The inuence of ber length and ber orientation on damping and stiness of polymer composite materials. Exp Mech 1986;26(2):17584. [91] Koratkar NA, Wei BQ, Ajayan PM. Multifunctional structural reinforcement featuring carbon nanotube lms. Compos Sci Technol 2003;63(11):152531. [92] Koratkar NA, Wei BQ, Ajayan PM. Carbon nanotube lms for damping applications. Adv Mater 2002;14(1314):9971000. [93] Koratkar NA, Suhr J, Joshi A, Kane RS, Schadler LS, Ajayan PM, et al. Characterizing energy dissipation in single-walled carbon nanotube polycarbonate composites. Appl Phys Lett 2005;87: 063102. [94] Zhou X, Shin E, Wang KW, Bakis CE. Interfacial damping characteristics of carbon nanotube composites. Compos Sci Technol 2004;64:242537. [95] Rajoria H, Jalili N. Passive vibration damping enhancement using carbon nanotube-epoxy reinforced composites. Compos Sci Technol 2005;65:207993. [96] Wang Z, Liang Z, Wang B, Zhang C, Kramer L. Processing and property investigation of single-walled carbon nanotube (SWNT) buckypaper/epoxy resin matrix nanocomposites. Composites A 2004;35:122532. [97] Macosko CW. Rheology: principles, measurements and applications. New York: VCH Pub. Co.; 1994. [98] Kinloch IA, Roberts SA, Windle AH. A rheological study of concentrated acqueous nanotube dispersions. Polymer 2002;43: 748391. [99] Abdel-Goad M, Potschke P. Rheological characterization of melt processed polycarbonate-multiwalled carbon nanotube composites. Journal non-Newtonian Fluid Mech 2005;128:26.

25

[100] Potschke P, Fornes TD, Paul DB. Rheological behavior of multiwalled carbon nanotube/polycarbonate composites. Polymer 2002;43:324755. [101] Meincke O, Kaempfer D, Weickmann H, Friedrich C, Vathauer M, Warth H. Mechanical properties and electrical conductivity of carbon nanotube lled polyamide 6 and its blends with acrylonitrile/ butadiene/styrene. Polymer 2004;45:73948. [102] McNally T, Potschke P, Halley P, Murphy M, Martin D, Bell SEJ, et al. Polyethylene multiwalled carbon nanotube composites. Polymer 2005;46:822232. [103] Potschke P, Abdel-Goad M, Alig I, Dudkin S, Lellinger D. Rhelogical and dielectric characterization of melt mixed polycarbonate-multiwalled carbon nanotube composites. Polymer 2004;45:886370. [104] Du F, Scogna RC, Zhou W, Brand S, Fischer JE, Winey K. Nanotube networks in polymer nanocomposites: rheology and electrical conductivity. Macromolecules 2004;37:904855. [105] Danailov D, Keblinski P, Nayak S, Ajayan PM. Bending properties of carbon nanotubes encapsulating solid nanowires. J Nanosci Nanotechnol 2002;2(5):5037. [106] Gadd GE, Blackford M, Moricca S, Webb N, Evans PJ, Smith AM, et al. The worlds smallest gas cylinders? Science 1997;277:9336. [107] Li J, Furuta T, Goto H, Ohashi T, Fujiwara Y, Yip S. Theoretical evaluation of hydrogen storage capacity in pure carbon nanostructures. J Chem Phys 2003;119(4):237685. [108] Skoulidas A, Ackerman DM, Johnson KJ, Sholl DS. Rapid transport of gases in carbon nanotubes. Phys Rev Lett 2002;89:185901. [109] Gogotsi Y, Libera JA, Guvenc-Yazicioglu A, Megardis M. In situ multiphase uid experiments in hydrothermal carbon nanotubes. Appl Phys Lett 2001;79(7):10213. [110] Tuzun RE, Noid DW, Sumpter BG, Merkle RC. Dynamics of uid ow inside carbon nanotubes. Nanotechnology 1996;7:2416. [111] Yoon J, Ru CQ, Mioduchowski A. Vibration and instability of carbon nanotubes conveying uid. Compos Sci Technol 2005;65:132636. [112] Wood JR, Zhao Q, Frogley MD, Meurs ER, Prins AD, Peijs T, et al. Carbon nanotubes: from molecular to macroscopic sensors. Phys Rev B 2000;62(11):75715. [113] Wood JR, Frogley MD, Meurs ER, Prins AD, Peijs T, Dunstan DJ, et al. Mechanical response of carbon nanotubes under molecular and macroscopic pressures. J Phys Chem B 1999;103(47):1038892. [114] Wood JR, Wagner HD. Single-wall carbon nanotubes as molecular pressure sensors. Appl Phys Lett 2000;76(20):28835. [115] Zhao Q, Wood JR, Wagner HD. Using carbon nanotubes to detect polymer transitions. J Polym Sci B 2001;39(13):14925. [116] Zhao Q, Wood JR, Wagner HD. Stress elds around defects and bers in a polymer using carbon nanotubes as sensors. Appl Phys Lett 2001;78(12):174850. [117] Binnig G, Quate CF, Gerber C. Atomic force microscope. Phys Rev Lett 1986;56(9):9303. [118] Dai H, Hafner JH, Rinzler AG, Colbert DT, Smalley R. Nanotubes as nanoprobes in scanning probe microscopy. Nature 1996;384(6605):14750. [119] Snow ES, Campbell PM, Novak JP. Single wall carbon nanotube atomic force microscope probes. Appl Phys Lett 2002;80(11):20024. [120] Moloni K, Buss MR, Andres RP. Tapping mode scanning force microscopy in water using a carbon nanotube probe. Ultramicroscopy 1999;80(4):23746. [121] Lee SI, Howell SW, Raman A, Reifenberger R, Nguyen CV, Meyyappan M. Nonlinear tapping dynamics of multi-walled carbon nanotube tipped atomic force microcantilevers. Nanotechnology 2004;15:41621. [122] Li C, Chou TW. Mass detection using carbon nanotube-based nanomechanical resonators. Appl Phys Lett 2004;84(25):52468. [123] Wei BY, Lin CS, Lin HM. Examining the gas-sensing behaviors of carbon nanotubes using a piezoelectric quartz microbalance. Sensors Mater 2003;15(4):17790.

26

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128 [148] Lu KL, Lago RM, Chen YK, Green MLH, Harris PJF, Tsang SC. Mechanical damage of carbon nanotubes by ultrasound. Carbon 1996;34(6):8146. [149] Heller DA, Mayrhofer RM, Baik S, Grinkova YV, Usrey ML, Strano MS. Concomitant length and diameter separation of single walled carbon nanotubes. J Am Chem Soc 2004;126(44): 1456773. [150] Tan YQ, Resasco DE. Dispersion of single walled carbon nanotubes of narrow diameter distribution. J Phys Chem B 2005;109(30): 1445460. [151] Shelimov KB, Esenaliev RO, Rinzler AG, Human CB, Smalley RE. Purication of single-walled carbon nanotubes by ultrasonically assisted ltration. Chem Phys Lett 1998;282:42934. [152] Liao YH, Marietta-Tomblin O, Liang Z, Zhang C, Wang B. Investigation of the dispersion process of SWNTs/SC-15 epoxy resin nanocomposites. Mater Sci Eng A 2004;385:17581. [153] Niyogi S, Hamon MA, Perea DE, Kang CB, Zhao B, Pal SK, et al. Ultrasonic dispersions of single walled carbon nanotubes. J Phys Chem B 2003;107:8799804. [154] Lau KT, Lu M, Lam CK, Cheung HY, Sheng FL, Li HL. Thermal and mechanical properties of single-walled carbon nanotube bundlereinforced epoxy nanocomposite: the role of solvent for nanotube dispersion. Compos Sci Technol 2005;65(5):71525. [155] Strano MS, Moore VC, Miller MK, Allen MJ, Haroz EH, Kittrell C, et al. The role of surfactant adsorption during ultrasonication in the dispersion of single-walled carbon nanotubes. J Nanosci Nanotechnol 2003;3(12):816. [156] Ramasetty A, Haque A, Shamsuzzoha AM. Study of ultrasonic dispersion method in CNT reinforced polymer matrix composites. Proceedings of the American Society for Composites 20th technical conference. Lancaster (PA): DesTech Publications; 2005. Paper No. 154 [CD-ROM]. [157] Safadi B, Andrews R, Grulke EA. Multiwalled carbon nanotube polymer composites: synthesis and characterization of thin lms. J Appl Polym Sci 2002;84(14):26609. [158] Park C, Ounaies Z, Watson KA, Crooks RE, Smith J, Lowther SE, et al. Dispersion of single wall carbon nanotubes by in situ polymerization under sonication. Chem Phys Lett 2002;364:3038. [159] Koda S, Amano T, Nomura H. Ultrasonics Sonochem 1996;3(2):S915. [160] Price GJ. Ultrasonically enhanced polymer synthesis. Ultrasonics Sonochem 1996;3(3):S22938. [161] Gronroos A, Pirkonen P, Heikkinen J, Ihalainen J, Mursunen H, Sekki H. Ultrasonic depolymerization of aqueous polyvinyl. Ultrasonics Sonochem 2001;8(3):25964. [162] Park C, Ounaies Z, Watson KA, Crooks RE, Smith J, Lowther SE, et al. Dispersion of single wall carbon nanotubes by in situ polymerization under sonication. Chem Phys Lett 2002;364(34): 3038. [163] Brown JM, Anderson DP, Justice RS, Lafdi K, Belfor M, Strong KL, et al. Hierarchical morphology of carbon single-walled nanotubes during sonication in an aliphatic diamine. Polymer 2005;46(24):1085465. [164] Kim H. Compatibilization of PC/SAN blends via ultrasoundirradiated melt mixing. Polymer-Korea 2004;28(3):22531. [165] Liu HN, Slamovich EB, Webster TJ. Increased osteoblast functions on nanophase titania dispersed in poly-lactic-co-glycolic acid. Nanotechnology 2005;16(7):S6018 [special issue]. [166] de Haro MCJ, Perez-Rodriguez JL, Poyato J, Perez-Maqueda LA, Ramirez-Valle V, Justo A, et al. Eect of ultrasound on preparation of porous materials from vermiculite. Appl Clay Sci 2005;30(1): 1120. [167] Baxter S, Zivanovic S, Weiss J. Molecular weight and degree of acetylation of high-intensity ultrasonicated chitosan. Food Hydrocolloids 2005;19(5):82130. [168] Matyshevska OP, Karlash AY, Shtogun YV, Benilov A, Kirgizov Y, Gorshinskyy KO, et al. Self-organizing DNA/carbon nanotube lms. Mater Sci Eng C 2001;15:24952.

[124] Chopra S, McGuire K, Gothard N, Rao AM, Pham A. Selective gas detection using a carbon nanotube sensor. Appl Phys Lett 2003;83(11):22802. [125] Chopra S, Pham A. Carbon nanotube-based resonant circuit sensor for ammonia. Appl Phys Lett 2002;80(24):46324. [126] Grujicic M, Cao G, Roy WN. A computational analysis of the carbon nanotube-based resonant circuit sensors. Appl Surf Sci 2004;229(14):31623. [127] Fennimore AM, Yuzvinsky TD, Han WQ, Fuhrer MS, Cumings J, Zetti A. Rotational actuators based on carbon nanotubes. Nature 2003;424:40810. [128] Baughman RH, Cui CX, Zakhidov AA, Iqbal Z, Barisci JN, Spinks GM, et al. Carbon nanotube actuators. Science 1999;284(5418): 13404. [129] Frysse J, Minett AI, Jaschinski O, Deusberg GS, Roth S. Carbon nanotubes acting like actuators. Carbon 2002;40(10):17359. [130] Spinks GM, Wallace GG, Field LS, Dalton LR, Mazzoldi A, De Rossi D, et al. Pneumatic carbon nanotube actuators. Adv Mater 2002;14(23):172831. [131] Kang JW, Hwang HJ. Gigahertz actuator of multiwall carbon nanotube encapsulating metallic ions: molecular dynamics simulations. J Appl Phys 2004;96(7):39005. [132] Cumings J, Zetti A. Low friction nanoscale linear bearing realized from multiwall carbon nanotubes. Science 2000;289:6024. [133] Kong J, Franklin NR, Zhou CW, Chapline MG, Peng S, Cho KJ, et al. Nanotube molecular wires as chemical sensors. Science 2000;287(5453):6225. [134] Bradley K, Gabriel J-CP, Star A, Grunerb G. Short-channel eects in contact-passivated nanotube chemical sensors. Appl Phys Lett 2003;83(18):38213. [135] He PG, Dai LM. Aligned carbon nanotube-DNA electrochemical sensors. Chem Commun 2004(3):3489. [136] Steinle ED, Mitchell DT, Wirtz M, Lee SB, Young VY, Martin CR. Ion channel mimetic micropore and nanotube membrane sensors. Anal Chem 2002;74(10):241622. [137] Ghosh S, Sood AK, Kumar N. Carbon nanotube ow sensors. Science 2003;299(5609):10424. [138] Kim W, Javey A, Vermesh O, Wang O, Li YM, Dai HJ. Hysteresis caused by water molecules in carbon nanotube eld-eect transistors humidity sensors. Nanoletters 2003;3(2): 1938. [139] Na PS, Kim HJ, So HM, Kong KJ, Chang HJ, Ryu BH, et al. Investigation of the humidity eect on the electrical properties of single-walled carbon nanotube transistors. Appl Phys Lett 2005;87(9):093101. [140] Grimes CA, Ong KG, Varghese OK, Yang XP, Mor G, Paulose M, et al. A sentinel sensor network for hydrogen sensing. Sensors 2003;3(3):6982. [141] Chen RJ, Choi HC, Bangsaruntip S, Yenilmez E, Tang XW, Wang Q, et al. An investigation of the mechanisms of electronic sensing of protein adsorption on carbon nanotube devices. J Am Chem Soc 2004;126(5):15638. [142] Loh KP, Zhao SL, De Zhang W. Diamond and carbon nanotube glucose sensors based on electropolymerization. Diamond Relat Mater 2004;13(48):10759. [143] Fung CMKM, Wong VTS, Chan RHM, Li WJ. Dielectrophoretic batch fabrication of bundled carbon nanotube thermal sensors. IEEE Trans Nanotechnol 2004;3(3):395403. [144] Li ZL, Dharap P, Nagarajaiah S, Barrera EV, Kim JD. Carbon nanotube lm sensors. Adv Mater 2004;16(7):692536. [145] Li CY, Chou TW. Strain and pressure sensing using single-walled carbon nanotubes. Nanotechnology 2004;15(11):14936. [146] Barber AH, Zhao Q, Wagner HD, Baillie CA. Characterization of E-glass-polypropylene interfaces using carbon nanotubes as strain sensors. Compos Sci Technol 2004;64(1314):19159. [147] Hilding J, Grulke EA, Zhang ZG, Lockwood F. Dispersion of carbon nanotubes in liquids. J Dispersion Sci Technol 2003;24(1): 141.

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128 [169] Lau EY, Lightstone FC, Colvin ME. Dynamics of DNA encapsulated in a hydrophobic nanotube. Chem Phys Lett 2005;412:827. [170] Chou SG, Ribeiro HB, Barros EB, Santos AP, Nezich D, Samsonidze GG, et al. Optical characterization of DNA-wrapped carbon nanotube hybrids. Chem Phys Lett 2004;397:296301. [171] Lu G, Maragakis P, Kaxiras E. Carbon nanotube interaction with DNA. Nanoletters 2005;5(5):897900. [172] Kam NWS, Dai H. Carbon nanotubes as intracellular protein transporters: generality and biological functionality. J Am Chem Soc 2005;127(16):60216. [173] Zheng M, Jagota A, Semke ED, Diner BA, McLean RS, Lustig SR, et al. DNA-assisted dispersion and separation of carbon nanotubes. Nat Mater 2003;2:33842. [174] Weber WH, Merlin R. Raman scattering in materials science. New York: Springer; 2000. [175] Fadini A, Schnepel FM. Vibrational spectroscopy methods and applications. Chichester, England: Ellis Horwood Limited; 1989. [176] Ecklund PC, Holden JM, Jishi RA. Vibrational modes of carbon nanotubes; spectroscopy and theory. Carbon 1995;33(7):95972. [177] Jorio A, Pimenta MA, Souza Filho AG, Saito R, Dresselhaus G, Dresselhaus MS. Characterizing carbon nanotube samples with resonance Raman scattering. New J Phys 2003;5:139.1139.17. [178] Dresselhaus MS, Jorio A, Souza Filho AG, Dresselhaus G, Saito R. Raman spectroscopy of one isolated carbon nanotube. Physica B 2002;323:1520. [179] Dresselhaus MS, Dresselhaus G, Jorio A, Souza AG, Samsonidze GG, Saito R. Science and applications of single nanotube Raman spectroscopy. J Nanosci Nanotechnol 2003;3(12):1937. [180] Benoit JM, Buisson JP, Godon C, Chauvet O, Lefrant S. Lowfrequency Raman studies of multiwalled carbon nanotubes: experiments and theory. Phys Rev B 2002;66:073417. [181] Guo Y, Wu J, Zhang Y. Manipulation of single-wall carbon nanotubes into aligned molecular layers. Chem Phys Lett 2002;362:3148. [182] Saito R, Takeyda T, Kimura T, Dresselhaus G, Dresselhaus MS. Raman intensity of single-wall carbon nanotubes. Phys Rev B 1998;57:414553. [183] Sun HD, Tang ZK, Chen J, Li G. Polarized Raman spectra of single-wall carbon nanotubes mono-dispersed in channels of AlPO45 single crystals. Solid State Commun 1999;109:3659. [184] Rao AM, Richter E, Bandow S, et al. Diameter-selective Raman scattering from vibrational modes in carbon nanotubes. Science 1997;275(5297):18791. [185] Richter E, Subbaswamy KR. Theory of size-dependent resonance Raman scattering from carbon nanotubes. Phys Rev Lett 1997;79(14):273841. [186] Popov VN, Van Doren VE, Balkanski M. Elastic properties of single-walled carbon nanotubes. Phys Rev B 1999;59:8355. [187] Pijun L, Liyue L, Yafei Z. Alignment characterization of single-wall carbon nanotubes by Raman scattering. Phys Lett A 2003;313:3026. [188] Rao AM, Jario A, Pimenta MA, Dantas MS, Saito R, Dresselhaus G, et al. Polarized Raman study of aligned multiwalled carbon nanotubes. Phys Rev Lett 2000;84:18203. [189] Kneipp K, Kneipp H, Corio P, Brown SDM, Shafer K, Motz J, et al. Surface-enhanced and normal Stokes and anti-Stokes Raman spectroscopy of single-walled carbon nanotubes. Phys Rev Lett 2000;84:34703. [190] Lyon LA, Keating CD, Fox AP, Baker BE, He Lin, Nicewarner SR, et al. Raman spectroscopy. Anal Chem 1998;70(12):341R61R. [191] Duesberg GS, Blau WJ, Byrne HJ, Muster J, Burghard M, Roth S. Experimental observation of individual single-wall nanotube species by Raman microscopy. Chem Phys Lett 1999;310:814. [192] Azoulay J, Debarre A, Richard A, Tchenio P. Confocal observation of individual classes of single wall nanotubes by surface-enhanced Raman spectroscopy. J Phys IV 2000;10(P8):2236. [193] Azoulay J, Debarre A, Richard A, Tchenio P. Quenching and enhancement of single-molecule uorescence under metallic and dielectric tips. Europhys Lett 2000;51(4):37480.

27

[194] Zhao X, Ando Y, Qin LC, Kataura H, Maniwa Y, Saito R. Characteristic Raman spectra of multiwalled nanotubes. Physica B 2002;323:2656. [195] Rao AM, Richter E, Bandow S, et al. Diameter-selective Raman scattering from vibrational modes in carbon nanotubes. Science 1997;275(5297):18791. [196] Saito R, Dresselhaus G, Dresselhaus MS. Physical properties of carbon nanotubes. London: Imperial College Press; 1998. [197] Bandow S, Asaka S, Saito Y, Rao AM, Grigorian L, Richter E, et al. Eect of the growth temperature on the diameter distribution and chirality of single-wall carbon nanotubes. Phys Rev Lett 1998;80:377982. [198] Thess A, Lee R, Nikolaev P, Dai H, Petit P, Robert J, et al. Crystalline ropes of metallic carbon nanotubes. Science 1996;273(5274):4839. [199] Sugano M, Kasuya A, Tohji K, Saito Y, Nishima Y. Resonance Raman scattering and diameter-dependent electronic states in single wall carbon nanotubes. Chem Phys Lett 1998;292(46):5759. [200] Alvarez L, Righi A, Guillard T, Rols S, Anglaret E, Laplaze D, et al. Resonant Raman study of the structure and electronic properties of single-wall carbon nanotubes. Chem Phys Lett 2000;316:18690. [201] Takizawa M, Bandow S, Yudasaka M, Ando Y, Shimoyama H, Iijima S. Change of tube diameter distribution of single-wall carbon nanotubes induced by changing the bimetallic ratio of Ni and Y catalysts. Chem Phys Lett 2000;326(3):3517. [202] Kurti J, Kuzmany H, Burger B, Hulman M, Winter J, Kresse G. Resonance Raman investigation of single wall carbon nanotubes. Synth Met 1999;103(13):25089. [203] Pimenta MA, Marucci A, Brown SDM, Matthews MJ, Rao AM, Eklund PC, et al. Resonant Raman eect in single wall carbon nanotubes. J Mater Res 1998;13(9):2396404. [204] Pimenta MA, Marucci A, Empedocles S, Bawendi M, Hanlon EB, Rao AM, et al. Spin-dependent phenomena in digital-magnetic heterostructures: clustering and phase-space lling eects. Phys Rev B 1998;58(24):R160125. [205] Corio P, Temperini MLA, Santos PS, Romero JV, Huber JG, Luengo CA, et al. Resonant Raman scattering characterization of carbon nanotubes grown with dierent catalysts. Chem Phys Lett 2001;350(56):37380. [206] Ci LJ, Xie SS, Tang DS, Yan XQ, Li YY, Liu ZQ, et al. Controllable growth of single wall carbon nanotubes by pyrolizing acetylene on the oating iron catalysts. Chem Phys Lett 2001;349:1915. [207] Ci LJ, Zhou ZP, Yan XQ, Liu DF, Yuan HJ, Song L, et al. Raman characterization and tunable growth of double-wall carbon nanotubes. J Phys Chem B 2003;107:87604. [208] Ci LJ, Zhou ZP, Yan XQ, Liu DF, Yuan HJ, Song L, et al. Resonant Raman scattering of double wall carbon nanotubes prepared by chemical vapor deposition method. J Appl Phys 2003;94:57159. [209] Xie SS, Song LI, Ci LJ, Zhou ZP, Dou XU, Weiya Zhou, et al. Controllable preparation and properties of single-/double-walled carbon nanotubes. Sci Technol Adv Mater 2005;6(7):72535. [210] Hadjiev VG, Iliev MN, Arepalli S, Nikolaev P, Files BS. Raman scattering test of single-wall carbon nanotube composites. Appl Phys Lett 2001;78(21):31935. [211] Lourie O, Wagner HD. Evaluation of Youngs modulus of carbon nanotubes by micro-Raman spectroscopy. J Mater Res 1998;13(9): 241822. [212] Reich S, Jantoljak H, Thomsen C. Shear strain in carbon nanotubes under hydrostatic pressure. Phys Rev B 2000;61(20):R1338992. [213] Dubay O, Kresse G, Kuzmany H. Phonon softening in metallic nanotubes by a Peierls-like mechanism. Phys Rev Lett 2002;88(23): 2355069. [214] Jorio A, Souza AG, Dresselhaus G, Dresselhaus MS, Swan AK, Unlu MS, et al. G-band resonant Raman study of 62 isolated singlewall carbon nanotubes. Phys Rev B 2002;65(15):15541220.

28

R.F. Gibson et al. / Composites Science and Technology 67 (2007) 128 [231] Wang SG, Wang JH, Zhang BH, Ma ZB, Wang CX, Man WD. Aligned carbon nanotubes grown on the inner surfaces of iron ring through microwave plasma chemical vapor deposition. J Mater Sci Lett 2003;22(14):10178. [232] Cheng HF, Tsau YM, Chang TY, Lai TS, Kuo TF, Lin IN. Electron eld emission properties of carbon nanotubes grown on iron needles. Diamond Relat Mater 2003;12(37):4869. [233] Wadhawan A, Garrett D, Perez JM. Nanoparticle-assisted microwave absorption by single-wall carbon nanotubes. Appl Phys Lett 2003;83(13):26835. [234] Che RC, Peng LM, Duan XF, Chen Q, Liang XL. Microwave absorption enhancement and complex permittivity and permeability of Fe encapsulated within carbon nanotubes. Adv Mater 2004;16(5):4015. [235] Roberts JA, Imholt T, Ye Z, Dyke CA, Price DW, Tour JM. Electromagnetic wave properties of polymer blends of single wall carbon nanotubes using a resonant microwave cavity as a probe. J Appl Phys 2004;95(8):43526. [236] Chen YJ, Cao MS, Wanga TH, Wan Q. Microwave absorption properties of the ZnO nanowire-polyester composites. Appl Phys Lett 2004;84(17):33679. [237] Liu L, Matitsine SM, Gan YB, Rozanov KN. Eective permittivity of planar composites with randomly or periodically distributed conducting bers. J Appl Phys 2005;98(6):063512. [238] Imholt TJ, Dyke CA, Hasslacher B, Perez JM, Price DW, Roberts JA, et al. Nanotubes in microwave elds: light emission, intense heat, outgassing and reconstruction. Chem Mater 2003;15:396970. [239] Nikolaev P, Bronikowski MJ, Bradley RK, Rohmund F, Colbert DT, Smith KA, et al. Gas-phase catalytic growth of single-walled carbon nanotubes from carbon monoxide. Chem Phys Lett 1999;313:917. [240] Wadhawan A, Garrett D, Perez JM. Nanoparticle-assisted microwave absorption by single-wall carbon nanotubes. Appl Phys Lett 2003;83(13):26835. [241] OConnell MJ, Bachilo SM, Human CB, Moore VC, Strano MS, Haroz EH, et al. Band gap uorescence from individual singlewalled carbon nanotubes. Science 2002;297:5936. [242] Kam NWS, OConnell M, Wisdom JA, Dai H. Carbon nanotubes as multifunctional biological transporters and near-infrared agents for selective cancer cell destruction. Proc Natl Acad Sci USA 2005;102(33):116005. [243] Loo C, Lowery A, Halas N, West J, Drezek R. Immunotargeted nanoshells for integrated cancer imaging and therapy. Nanoletters 2005;5(4):70911. [244] Wang L, Hu H. Flexural wave propagation in single-walled carbon nanotubes. Phys Rev B 2005;71:195412. [245] Sun CT, Zhang H. Size-dependent elastic moduli of platelike nanomaterials. J Appl Phys 2003;93(2):12128. [246] Zhang H, Sun CT. Nanoplate model for platelike nanomaterials. AIAA J 2004;42(10):20029.

[215] Siochi EJ, Working DC, Park C, Lillehei PT, Rouse JH, Topping CC, et al. Melt processing of SWCNT-polyimide nanocomposite bers. Composites B 2004;35:43946. [216] Haggenmueller R, Gommans HH, Rinzler AG, Fischer JE, Winey KI. Aligned single-wall carbon nanotubes in composites by melt processing methods. Chem Phys Lett 2000;330(34):21925. [217] Haggenmueller R, Zhou W, Fischer JE, Winey KI. Production and characterization of polymer nanocomposites with highly aligned single-walled carbon nanotubes. J Nanosci Nanotechnol 2003; 3(12):10510. [218] Bhattacharyya AR, Sreekumar TV, Liu T, Kumar S, Ericson LM, Hauge RH, et al. Crystallization and orientation studies in polypropylene/single wall carbon nanotube composite. Polymer 2003;44:23737. [219] Sreekumar VT, Liu T, Min BG, Guo H, Kumar S, Hauge RH, et al. Polyacrylonitrile single-walled carbon nanotube composite bers. Adv Mater 2004;16(1):5860. [220] Lucas M, Young RJ. Eect of uniaxial strain deformation upon the Raman radial breathing modes of single-wall carbon nanotubes in composites. Phys Rev B 2004;69:085405. [221] Wood JR, Zhao Q, Wagner HD. Orientation of carbon nanotubes in polymers and its detection by Raman spectroscopy. Composites A 2001;32:3919. [222] Frogley MD, Ravich D, Wagner HD. Mechanical properties of carbon nanoparticle-reinforced elastomers. Compos Sci Technol 2003;63(11):164754. [223] Cooper CA, Young RJ, Halsall M. Investigation into the deformation of carbon nanotubes and their composites using Raman spectroscopy. Composites A 2001;32:40111. [224] Raravikar NR, Keblinski P, Rao AM, Dresselhaus MS, Schadler LS, Ajayan PM. Temperature dependence of radial breathing mode Raman frequency of single-walled carbon nanotubes. Phys Rev B 2002;66(23):235424. [225] Zhao Q, Wood JR, Wagner HD. Using carbon nanotubes to detect polymer transitions. Polym Sci B 2001;39:14925. [226] Thostenson ET, Chou TW. Microwave processing: fundamentals and applications. Composites A 1999;30:105571. [227] National Research Council National Materials Advisory Board. Microwave processing of materials. The National Academies Press; 1994. Available from: http://www.nap.edu/. [228] John MK, Hosur MV, Rangari VK, Jeelani S. On the processability of nanophased composites using microwave. In: Proceedings 22nd southeastern conference on theoretical and applied mechanics (SECTAM XXII), Tuskegee University, Alabama; 2004. p. 5969. [229] Yao BD, Wang N. Carbon nanotube arrays prepared by MWCVD. J Phys Chem B 2001;105(46):113958. [230] Liu RM, Ting JM. Growth of carbon nanotubes using microwave plasma-enhanced chemical vapor deposition process. Mater Chem Phys 2003;82(3):5714.