Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P....

13
Carbon nanotube composites P. J. F. Harris* Carbon nanotubes are molecular-scale tubes of graphitic carbon with outstanding properties. They are among the stiffest and strongest fibres known, with Young’s moduli as high as 1 TPa and tensile strengths of up to 63 GPa. They also have remarkable electronic properties and can be metallic or semiconducting depending on their structure and diameter. There is currently great interest in exploiting these properties by incorporating carbon nanotubes into some form of matrix. A wide range of polymer matrices have been employed, and there is growing interest in nanotube/ceramic and nanotube/metal composites. This review outlines the properties of carbon nanotubes and describes the preparation and properties of carbon nanotube composites. The prospects for commercial exploitation of these materials are discussed. Keywords: Nanotubes, Composite materials, Nanotube/polymer composites IMR/406 Introduction Carbon nanotubes are among the most exciting new materials to have been discovered in the past 30 years. 1 Their remarkable properties have attracted intense inter- est from the scientific community, as well as from industry, and nanotubes are currently the subject of around 7 papers per day. The explosion of interest in carbon nanotubes can be traced back to a 1991 Nature paper by Sumio Iijima. 2 Iijima was using high resolution transmission electron microscopy (HRTEM) to examine carbon produced by the arc evaporation of graphite in an atmosphere of helium. About a year earlier, it had been shown that this method could be used to produce C 60 and other fullerenes in high yield, 3 so it seemed likely that other novel forms of carbon might be formed at the same time. It was while studying the hard deposit which formed on the graphite cathode following arc evaporation that Iijima made his discovery: the central part of the deposit contained large numbers of tiny tubules of graphitic carbon, consisting of concentric graphene cylinders, typically about 10 nm in diameter, with fullerene-like end-caps. Although tubes of carbon, produced catalytically, had been known for decades, the structures discovered by Iijima were far more perfect than any that had been seen before, and promised to have exceptional properties. In subsequent work, two groups showed that single-walled nanotubes could be produced in a similar apparatus in which one or both electrodes contained cobalt, nickel, or some other metal. 4,5 These single-walled tubes generally have smaller dia- meters than the multiwalled nanotubes (typically of the order of 1–2 nm dia.). It was shown in 1996 that single- walled nanotubes, like multiwalled nanotubes, can be produced catalytically. 6 There is now abundant experimental evidence of the outstanding properties of carbon nanotubes. Their stiffness and strength are phenomenal. Measurements using in situ transmission electron microscopy (TEM), and atomic force microscopy (AFM) have produced estimates for the Young’s modulus of the order of 1 TPa. 7,8 For comparison, the stiffest conventional carbon fibres have Young’s moduli of approximately 800 GPa, while glass fibres typically have moduli of about 70 GPa. Direct measurements on individual nanotubes using atomic force microscopy have shown that they can accommo- date extreme deformations without fracturing; 9 they also have the extraordinary capability of returning to their original, straight, structure following deformation. In addition, carbon nanotubes are excellent electrical con- ductors, 10 with current densities of up to 10 11 Am 22 , and have very high thermal conductivities. 11 Many of these outstanding properties can be best exploited by incorporating the nanotubes into some form of matrix, and the preparation of nanotube con- taining composite materials is now a rapidly growing subject. 12,13 In many cases, these composites have employed polymer matrices, but there is also interest in other matrix materials, such as ceramics and metals. The aim of this paper is to review recent work on carbon nanotube composites, and to assess how successful this work has been in exploiting the full potential of nano- tubes. To begin, a brief introduction is given to the science of carbon nanotubes. Structure and properties of carbon nanotubes The nanotubes discovered by Iijima have structures closely related to those of fullerenes. This can be most clearly illustrated by considering the two ‘archetypal’ carbon nanotubes which can be formed by cutting a C 60 molecule in half and placing a graphene cylinder between the two halves. Dividing C 60 along one of the five-fold axes produces the ‘armchair’ nanotube shown in Fig. 1a, while bisecting C 60 parallel to one of the three-fold axes results in the ‘zigzag’ nanotube shown in *Centre for Advanced Microscopy, University of Reading, Whiteknights, Reading, RG6 6AF, UK, email [email protected]. ß 2004 IoM Communications Ltd and ASM International Published by Maney for the Institute of Materials, Minerals and Mining and ASM International DOI 10.1179/095066004225010505 International Materials Reviews 2004 VOL 49 NO 1 31

Transcript of Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P....

Page 1: Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P. J. F. Harris * Carbon nanotubes are molecular-scale tubes of graphitic carbon with

Carbon nanotube composites

P J F Harris

Carbon nanotubes are molecular-scale tubes of graphitic carbon with outstanding properties

They are among the stiffest and strongest fibres known with Youngrsquos moduli as high as 1 TPa and

tensile strengths of up to 63 GPa They also have remarkable electronic properties and can be

metallic or semiconducting depending on their structure and diameter There is currently great

interest in exploiting these properties by incorporating carbon nanotubes into some form of

matrix A wide range of polymer matrices have been employed and there is growing interest in

nanotubeceramic and nanotubemetal composites This review outlines the properties of carbon

nanotubes and describes the preparation and properties of carbon nanotube composites The

prospects for commercial exploitation of these materials are discussed

Keywords Nanotubes Composite materials Nanotubepolymer composites IMR406

IntroductionCarbon nanotubes are among the most exciting newmaterials to have been discovered in the past 30 years1

Their remarkable properties have attracted intense inter-est from the scientific community as well as fromindustry and nanotubes are currently the subject ofaround 7 papers per day The explosion of interest incarbon nanotubes can be traced back to a 1991 Naturepaper by Sumio Iijima2 Iijima was using high resolutiontransmission electron microscopy (HRTEM) to examinecarbon produced by the arc evaporation of graphite inan atmosphere of helium About a year earlier it hadbeen shown that this method could be used to produceC60 and other fullerenes in high yield3 so it seemedlikely that other novel forms of carbon might be formedat the same time It was while studying the hard depositwhich formed on the graphite cathode following arcevaporation that Iijima made his discovery the centralpart of the deposit contained large numbers of tinytubules of graphitic carbon consisting of concentricgraphene cylinders typically about 10 nm in diameterwith fullerene-like end-caps Although tubes of carbonproduced catalytically had been known for decades thestructures discovered by Iijima were far more perfectthan any that had been seen before and promised tohave exceptional properties In subsequent work twogroups showed that single-walled nanotubes could beproduced in a similar apparatus in which one or bothelectrodes contained cobalt nickel or some othermetal45

These single-walled tubes generally have smaller dia-meters than the multiwalled nanotubes (typically of theorder of 1ndash2 nm dia) It was shown in 1996 that single-walled nanotubes like multiwalled nanotubes can beproduced catalytically6

There is now abundant experimental evidence ofthe outstanding properties of carbon nanotubes Their

stiffness and strength are phenomenal Measurementsusing in situ transmission electron microscopy (TEM)and atomic force microscopy (AFM) have producedestimates for the Youngrsquos modulus of the order of1 TPa78 For comparison the stiffest conventional carbonfibres have Youngrsquos moduli of approximately 800 GPawhile glass fibres typically have moduli of about 70 GPaDirect measurements on individual nanotubes using atomicforce microscopy have shown that they can accommo-date extreme deformations without fracturing9 they alsohave the extraordinary capability of returning to theiroriginal straight structure following deformation Inaddition carbon nanotubes are excellent electrical con-ductors10 with current densities of up to 1011 A m22and have very high thermal conductivities11

Many of these outstanding properties can be bestexploited by incorporating the nanotubes into someform of matrix and the preparation of nanotube con-taining composite materials is now a rapidly growingsubject1213 In many cases these composites have employedpolymer matrices but there is also interest in othermatrix materials such as ceramics and metals The aimof this paper is to review recent work on carbonnanotube composites and to assess how successful thiswork has been in exploiting the full potential of nano-tubes To begin a brief introduction is given to thescience of carbon nanotubes

Structure and properties of carbonnanotubesThe nanotubes discovered by Iijima have structuresclosely related to those of fullerenes This can be mostclearly illustrated by considering the two lsquoarchetypalrsquocarbon nanotubes which can be formed by cutting aC60 molecule in half and placing a graphene cylinderbetween the two halves Dividing C60 along one of thefive-fold axes produces the lsquoarmchairrsquo nanotube shownin Fig 1a while bisecting C60 parallel to one of thethree-fold axes results in the lsquozigzagrsquo nanotube shown in

Centre for Advanced Microscopy University of Reading WhiteknightsReading RG6 6AF UK email pjfharrisrdgacuk

2004 IoM Communications Ltd and ASM InternationalPublished by Maney for the Institute of Materials Minerals and Miningand ASM InternationalDOI 101179095066004225010505 International Materials Reviews 2004 VOL 49 NO 1 31

Fig 1b The terms lsquozigzagrsquo and lsquoarmchairrsquo refer to thearrangement of hexagons around the circumferenceThere is a third class of structure lsquochiralrsquo in which thehexagons are arranged helically around the tube axis asshown in Fig 1c Experimentally the tubes are generallyless perfect than the idealised versions shown in Fig 1and as already noted may be either multiwalled orsingle-walled The abbreviations MWNT and SWNTare commonly used for multiwalled and single-wallednanotubes respectively Figure 2a shows a typicalsample of MWNT-containing cathodic soot at moderatemagnification As can be seen the nanotubes are accom-panied by other material including nanoparticles (hollowfullerene-related structures) and some disordered car-bon At high resolution the individual layers making upthe concentric tubes can be clearly imaged as in Fig 2bThe multiwalled tubes range in length from a few tensof nanometres to several micrometres and in outerdiameter from about 25 to 30 nm The end-caps of thetubes are sometimes symmetrical in shape but moreoften asymmetric Conical structures of the type shownin Fig 3 are commonly observed This type of structureis believed to result from the presence of a singlepentagon (at the position indicated by the arrow) withfive further pentagons at the apex of the cone

Multiwalled tubes produced by catalytic methods aregenerally much more defective than those made by arcevaporation as can be seen in Fig 414 However theirproperties can be improved by high temperature heattreatment15 Some typical micrographs of SWNTs canbe seen in Fig 516 As a result of their very smalldiameters the tubes are much more flexible thanMWNTs and are therefore often observed to be curledand looped rather than straight Because SWNTs fre-quently have a narrow range of diameters they tend tocluster together to form lsquoropesrsquo

There have been a large number of studies boththeoretical and experimental of the mechanical proper-ties of carbon nanotubes The first measurements ofstiffness were made in 1996 by Treacy et al7 whocarried out in situ measurements of the intrinsic thermalvibrations of multiwalled carbon nanotubes in a TEMThese indicated that the tubes had extremely high elasticmoduli but the error on the measurements was largeestimates for the Youngrsquos modulus ranged from

a armchair (n m)5(5 5) b zigzag (n m)5(9 0) c chiral

(n m)5(10 5)

1 Three classes of carbon nanotube structure

2 a TEM image of nanotube-containing soot and b

higher magnification image of individual tubes

3 Image of typical multiwalled nanotube (MWNT) cap with arrow indicating approximate position of pentagonal ring

Harris Carbon nanotube composites

32 International Materials Reviews 2004 VOL 49 NO 1

410 GPa to 415 TPa with an average of 18 TPaRather more accurate measurements were made a usinga lsquonanostressing stagersquo inside a scanning electron micro-scope (SEM)17 These experiments indicated that theYoungrsquos modulus of the outermost layers of multiwalledcarbon nanotubes varied from 270 to 950 GPa Thetensile strengths of individual nanotubes were also deter-mined in this work and values ranging from 11 to63 GPa were found Measurements of Youngrsquos modulushave also been carried out using atomic force micro-scopy Wong et al8 described a method of fixing nano-tubes at one end and then determining the bendingforce F of the fibres as a function of displacement d Forsmall deflections linear Fndashd curves were observed andthe results implied a value of y128 TPa for theYoungrsquos modulus However for deflections larger than10u an abrupt change in the slope of the Fndashd curve wasseen This was attributed to elastic buckling of the tubesIn TEM studies of nanotubes buckled tubes are quiteoften seen as a result of stresses imposed by the way thetubes are supported on the grid In some cases thebuckles are regularly spaced as shown in Fig 618 It hasbeen demonstrated that when the stress constraininga tube is released the tube can return to its originalstraight form This behaviour sets nanotubes apart fromconventional carbon fibres and all other currently knownfibres which are much more susceptible to fracture whensubject to stress beyond the elastic limit It also raisesthe fascinating possibility that composites could beproduced which would snap back into place followingdeformation

As already noted catalytically produced MWNTs aremuch less perfect in structure than arc grown tubes andare therefore considerably less stiff Salvetat et al usedatomic force microscopy to record forcendashdisplacementcurves for nanotubes prepared by the catalytic decom-position of acetylene19 This produced an averagevalue for the Youngrsquos modulus of 27 GPa The tensile

strengths of catalytically produced MWNTs have notyet been measured However it is possible that thesemight be relatively high the defects in these imperfectstructures might be expected to assist stress transferbetween layers

Determining the mechanical properties of SWNTspresents an even greater challenge than for MWNTsHowever some measurements have been made using ananostressing stage20 These produced values for theYoungrsquos modulus in the range 320ndash1470 GPa (mean1002 GPa) and an average breaking strength of 30 GPa

There have also been a substantial number of studiesof the electronic properties of carbon nanotubes121

Early theoretical work showed that nanotubes could bemetallic or semiconducting depending on their structureand diameter22ndash24 Ebbesen et al confirmed experimen-tally that the electronic properties of nanotubes couldvary widely from tube to tube25 both semiconductingand metallic behaviour has been observed Measuredresistivities fall into the approximate range 005 mV mndash10 mV m This compares with a value of approximately04 mV m for high quality single crystal graphite(parallel to the c axis) and 0017 mV m for copper Theresistivities of catalytically produced nanotubes arelikely to be similar to those of disordered carbons ieof the order of 10ndash100 mV m but experimental measure-ments on this type of tube have not yet been carried outThe thermal and optical properties of carbon nanotubes

a low magnification TEM image b HRTEM image of

typical nanotube

4 TEM images of multiwalled carbon nanotubes synthe-

sised by decomposition of C2H2 at 850uC over cobalt

catalyst14

a low magnification SEM image b high magnification

TEM image

5 Images of single-walled carbon nanotubes prepared

by decomposition of C2H2 at 950uC over FeMoAl2O3

catalyst16

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 33

have been studied less intensively than the electronicproperties but some interesting results have been obtainedThe high thermal conductivities of nanotubes have alreadybeen mentioned11 It has also been reported that carbonnanotube suspensions have non-linear optical proper-ties suggesting possible applications as optical limiters26

Another area which has been the subject of muchresearch is the chemical modification of carbon nano-tubes Chemical treatments of nanotubes were first carriedout in an attempt to selectively remove the end-caps anduse them as molecular containers It was demonstratedin 1994 that nanotube caps could be removed by treat-ment with hot nitric acid apparently as a result ofselective oxidation at the pentagonal rings27 In sub-sequent work it was demonstrated by acidndashbase titra-tion and by temperature programmed decompositionthat this acid treatment resulted in the formation ofsurface carboxylic acid and other groups on the tubesurfaces28 These surface acid groups can then be furthermodified in a variety of ways29 Functionalisation hasbeen widely used to facilitate the dispersion of nano-tubes in solvents as a prelude to the preparation ofcomposite materials and can also help to improvebonding between tubes and matrix30

Preparation of carbon nanotubepolymercompositesA commonly used method for preparing nanotubepolymer composites has involved mixing nanotube dis-persions with solutions of the polymer and then eva-porating the solvents in a controlled way As noted inthe previous section the nanotubes are often pretreatedchemically to facilitate solubilisation Shaffer et al forexample have demonstrated that acid treatments enablestable aqueous solutions of catalytically producedMWNTsto be prepared31 These workers then showed that ananotubePVA (polyvinyl alcohol) composite could be

prepared simply by mixing one of these aqueous nano-tube dispersions with an aqueous solution of the poly-mer and then casting the mixtures as films andevaporating the water32 Solution based methods havealso been used to produce nanotubepolystyrene com-posites Thus Hill et al have solubilised both SWNTsand MWNTs by functionalising with a polystyrenecopolymer33 This was achieved by first acid treating thetubes and then carrying out esterification of the surface-bound carboxylic acids The polymer-modified carbonnanotubes were shown to be soluble in common organicsolvents To prepare composites polystyrene was dissolvedin the nanotube solution and nanotubepolystyrene thinfilms prepared using wet casting Functionalisation ofthe nanotubes is not always necessary in order to pre-pare a polymer composite Qian et al used a high energyultrasonic probe to disperse MWNTs in toluene andthen mixed the dispersed suspension with a dilute solu-tion of polystyrene in toluene again with ultrasonicagitation34 The low viscosity of the polymer solutionallowed the nanotubes to move freely through thematrix The mixture was cast on glass and the solventremoved to yield MWNT-doped films Specimens of thecomposite were prepared for TEM by placing drops ofthe mixed solution on copper TEM grids On evapora-tion of the toluene a thin film of the composite was leftsuspended across the grid bars A micrograph of one ofthese films is shown in Fig 7 and this illustrates theexcellent dispersion that was achieved Note that in situpolymerisation has also been used to prepare nanotubepolystyrene composites as is discussed below

The solution mixing approach is limited to polymersthat freely dissolve in common solvents An alternativeis to use thermoplastic polymers (ie polymers thatsoften and melt when heated) and then apply meltprocessing techniques Thus shear mixing can be used toproduce a homogeneous dispersion of nanotubes andextrusion to produce nanotube alignment or to fabricateartefacts in the required form by injection moulding

7 TEM image of MWNTpolystyrene thin film (arrows

indicate regions of polymer shrinkage) inset shows

distribution of nanotube lengths346 High resolution TEM image of deformed nanotube

showing periodic buckling18

Harris Carbon nanotube composites

34 International Materials Reviews 2004 VOL 49 NO 1

Andrews et al used shear mixing to disperse catalytically-produced nanotubes in a range of polymers includinghigh impact polystyrene acrylonitrile butadiene styreneand polypropylene35 The dispersion of nanotubes wasdetermined as a function of mixing energy and tem-perature and the composites were formed into fibres andthin films Haggenmueller et al used a combination ofsolvent casting and melt mixing methods to disperseSWNTs in polymethylmethacrylate (PMMA)36 Compositefibres with a high degree of nanotube orientation wereproduced by melt spinning

Polycarbonate is another thermoplastic polymer thathas been used as a matrix for nanotube composites3738

Sennet et al have used melt processing techniques todisperse and align carbon nanotubes in polycarbonate38

Dispersion was achieved by mixing the catalyticallyproduced MWNTs and SWNTs with polycarbonateresin in a conical twin-screw extruder and alignment wascarried out using a fibre spinning apparatus By opti-mising mixing time and fibre draw rates excellent dis-persion and alignment were accomplished

Nanotubeepoxy composites have also been preparedby a number of groups39ndash43 The earliest work in thisarea was carried out by Ajayan et al39 In 1994 thisgroup described a study in which purified tubes wereembedded into an epoxy resin which was then cut intothin slices with a diamond knife Their initial aim indoing this was not to produce a composite material butto obtain thin TEM specimens which it was hoped wouldcontain cross-sectional images of nanotubes However nocross-sections were observed following sectioning insteadthe nanotubes were found to have become aligned in thedirection of the knife movement as shown in Fig 8According to Ajayan et al the alignment is primarily aconsequence of extensional or shear flow of the matrixproduced by the cutting Of course alignment can bemuch more readily achieved when the epoxy is in a liquidstate Xu et al showed that some alignment of nanotubesin nanotubeepoxy composites can been achieved by usingspin coating methods42 The cure reaction of a SWNT

epoxy resin composite has been analysed using thermalanalysis and Raman spectroscopy43 The SWNTs werefound to have a catalysing effect on the cure reaction

An alternative method for preparing nanotubepoly-mer composites is to use the monomer rather than thepolymer as a starting material and then carry out in situpolymerisation Cochet et al were among the first to usethis method to prepare a MWNTpolyaniline compo-site44 Nanotubes prepared by arc evaporation wereused and were sonicated in a solution of HCl to achievea dispersion The aniline monomer again in HCl wasadded to the suspension and a solution of an oxidantwas slowly added with constant sonication and coolingSonication was continued in an ice bath for 2 h and thecomposite was then obtained by filtering rinsing anddrying In this way it was possible to prepare com-posites with high MWNT loadings (up to 50 wt-)Transport measurements on the composite revealedmajor changes in the electronic behaviour confirmingstrong interaction between nanotubes and polymer asdiscussed below A number of other nanotubepolymercomposites have been prepared using in situ poly-merisation including MWNTpolystyrene45 and SWNTpolyimide46 Several groups have used electrochemicalpolymerisation to grow porous composite films ofMWNTand polypyrrole for use as supercapacitors4748

In the techniques described so far the aim has been toproduce composites in which the nanotubes are dis-tributed evenly throughout the polymer (either ran-domly oriented or aligned) For some applicationshowever a layered arrangement is advantageous Photo-voltaic devices containing layers of nanotubes or nanotubepolymer composites have been prepared as discussed inthe section lsquoElectrical and optical properties of carbonnanotubepolymer compositesrsquo below The preparationof multilayer SWNTpolymer composites with excellentmechanical properties has also been described49 Themethod involved the layer-by-layer deposition of SWNTsand polymer onto a substrate followed by crosslinkingIn this way composites with SWNT loadings as high as

8 Alignment of nanotubes in polymer matrix following cutting with microtome arrows indicate buckled nanotubes39

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 35

50 wt- could be obtained Coating was carried out byalternate dipping of a glass slide or silicon wafer intodispersions of SWNTs (stabilised by acid treatment) andpolymer solutions The layers were held together by vander Waals forces and by electrostatic attraction betweenthe negatively charged SWNTs and a positively chargedpolyelectrolyte such as branched polyethyleneimine (PEI)Coatings containing up to 50 SWNTPEI bilayers couldbe built up in this way When the procedure was com-plete the multilayer films were heated to 120uC topromote crosslinking The films could then be lifted offthe substrate to obtain uniform free-standing mem-branes A rather similar method was used to synthesisemultilayer composites films of the polyelectrolyte poly-diallyldimethylammonium chloride and SWNTs50

Before proceeding to discuss the properties of carbonnanotubepolymer composites in detail it is worth brieflyconsidering the effect that the inclusion of nanotubescan have on polymer microstructure Although thistopic has not yet been widely studied a number ofgroups have demonstrated that the crystallisation andmorphology of the polymers can be strongly affected bysmall additions of nanotubes For example Lozano andBarrera51 showed that the addition of catalytically growncarbon nanofibres can promote nucleation of polypro-pylene Similar observations were made by Valentiniet al52 The crystallinity of a polymers has also beenshown to increase with the addition of carbon nano-fibres5354 Changes in the crystal structure of polymerscan also be induced by nanotubes55 More work on theseeffects would be welcome

Mechanical properties of carbonnanotubepolymer compositesMuch of the work on the preparation of nanotubepolymer composites has been driven by a desire toexploit the tubesrsquo stiffness and strength Even where theinterest has been focused on other properties the abilityof nanotubes to improve the mechanical characteristicsof a polymer has often been a valuable added benefit Inearly work the measured properties were often dis-appointing but such studies were of value in under-standing the reasons for composite failure and inidentifying the critical issues which need to be addressed

Shaffer and Windle were among the first to carry outa systematic study of the mechanical properties ofnanotubepolymer composites Their technique for pre-paring MWNTPVA composites was described in theprevious section32 The tensile elastic moduli of thecomposite films were assessed in a dynamic mechanicalthermal analyser as a function of nanotube loading andtemperature The stiffness of the composites at roomtemperature was relatively low From the theory deve-loped for short-fibre composites a nanotube elasticmodulus of 150 MPa was obtained from the roomtemperature experimental data This value is well belowthe values reported for isolated nanotubes even defec-tive catalytically grown tubes have measured stiffness ofthe order of 30 GPa19 The low value may have more todo with poor stress transfer than with the weakness ofthe nanotubes themselves Above the glass transitiontemperature of the polymer (y85uC) the nanotubeshad a more significant effect on the properties of thecomposite

Cadek et al achieved rather better results in 2002with composites made from PVA and arc producedMWNTs56 The composites were prepared by solutionmixing and films were formed on glass substrates bydrop casting The presence of 1 wt- nanotubes increasedthe Youngrsquos modulus and hardness of the PVA byfactors of 18 and 16 respectively This improvementover the early results of Shaffer and Windle reflects boththe superior quality of the tubes and stronger interfacialbonding Evidence for the latter was provided by TEMstudies

The mechanical properties of nanotubepolystyrenecomposites were studied by Qian and co-workers3435

Two samples of catalytically produced MWNTs wereused one with an average length of y15 mm the otherwith an average length of y50 mm With the addition of1 wt- nanotubes they achieved a 36 and 42 increasein the elastic stiffness of the polymer for short and longtubes respectively (the neat polymer modulus wasy12 GPa) In both cases a 25 increase in the tensilestrength was observed Theoretical calculations carriedout assuming a nanotube modulus of 450 GPa pro-duced increases of a 48 and 62 in the stiffness of thepolymer for short and long tubes respectively The closeagreement within y10 between the experimental andtheoretically predicted composite moduli indicated thatthe external tensile loads were successfully transmittedto the nanotubes across the tubepolymer interface Athigher nanotube concentrations the changes in themechanical properties of the composites were morepronounced35 For MWNTpolystyrene composites con-taining from 25 to 25 vol- nanotubes Youngrsquosmodulus increased progressively from 19 to 45 GPawith the major increases occurring when the MWNTcontent was at or above about 10 vol- However thedependence of tensile strength on nanotube concentra-tion was more complex At the lower concentrations((10 vol-) the tensile strength decreased from theneat polymer value ofy40 MPa only exceeding it whenthe MWNT content was above 15 vol-

In an attempt to understand the tensile fracturemechanisms of the nanotubepolystyrene compositesQian et al performed deformation studies inside aTEM34 Their elegant method of preparing TEM speci-mens has been described above Focusing the electronbeam onto the thin film resulted in local thermal stresseswhich initiated cracks in the composite The propaga-tion speed of the crack could be controlled by varyingthe beam flux onto the sample These in situ TEMobservations showed that the cracks tended to nucleateat low nanotube density areas then propagate alongweak nanotubepolymer interfaces or relatively low nano-tube density regions The nanotubes became alignedperpendicular to the crack direction and bridged thecrack faces in the wake as shown in Fig 9 When thecrack opening displacement exceeded y800 nm thenanotubes began to break andor pull out of the matrixThe fact that some of the pulled-out tubes do not appearto be coated with polymer suggests that there is room forimprovement in nanotubendashpolymer bonding

Watts and Hsu have also used TEM to study thefracture behaviour of a nanotubepolymer composite57

Their composite consisted of arc grown MWNTsembedded in a diblock copolymer known as MPCndashDEA To prepare specimens of the composite for TEM

Harris Carbon nanotube composites

36 International Materials Reviews 2004 VOL 49 NO 1

the tubes were mixed with the polymer in an acidifiedaqueous solution and drops of this were applied to TEMgrids which were then dried in vacuum As in the workof Qian et al cracks were produced in the material byirradiation with an electron beam but a low acceleratingvoltage (75 kV) was used to avoid massive disruption ofthe thin composite film Individual tubes and bundles oftubes were observed to become stretched as the cracksgrew wider and pull-out was often seen In contrast tothe work of Qian et al however the tubes did notbreak This was probably because high quality arcgrown nanotubes were used rather than catalyticallyproduced tubes

The early work on nanotubeepoxy composites byAjayan et al has been described in the previoussection39 In 1998 Ajayan and co-workers described astudy of the mechanical behaviour of MWNTepoxycomposites in both tension and compression58 It wasfound that the compression modulus was higher thanthe tensile modulus indicating that load transfer to thenanotubes in the composite was much higher in com-pression The authors suggested that during load trans-fer to MWNTs only the outer layers are stressed intension whereas all the layers respond in compression

Nanotubeepoxy thin films prepared by Xu et alshowed excellent mechanical properties42 Compared toneat resin thin films for which the elastic modulus was42 GPa a 20 increase in modulus was seen when01 wt- MWNTs were added This was attributed topartial alignment of the MWNTs induced by spincoating Fracture behaviour of the films was investigatedby SEM and showed that pulled-out tubes were oftencovered with polymer suggesting strong interfacial adhe-sion Similar results were reported by Gojny et al59 Thisgroup showed that functionalising arc produced MWNTswith amine groups greatly enhanced bonding with theepoxy matrix As a result the outer shells of the tubesoften remained embedded in the matrix following pull-out as shown in Fig 10

All of the work described so far in this section hasinvolved multiwalled carbon nanotubes Ajayan et al

have described a study of SWNT-containing compo-sites60 Two types of composite were prepared first anSWNTepoxy composite and second a pressed pelletcontaining SWNTs and carbonaceous soot materialformed during nanotube synthesis The SWNTs wereused as-prepared ie generally grouped into bundlesrather than separated into individual tubes The com-posites were loaded axially in tension until failureoccurred and the fracture surfaces were then examinedin detail using SEM In some cases the tube bundleswere found to have been pulled out of the matrix duringthe deformation and fracture of the composites while inother cases the nanotubes were not entirely pulled outbut were stretched between two fracture surfaces Theauthors believed that during pull-out the nanotubeswere sliding axially within the ropes and that the failureobserved at large crack distances was not failure ofindividual tubes but the bundles pulling apart Studiesby micro-Raman spectroscopy supported this hypoth-esis It was found that little shift in the second order A1g

band occurred as a result of the application of axialtension showing that the individual nanotubes were notbeing significantly stretched To take full advantage ofthe high Youngrsquos modulus of SWNTs in polymer com-posites it is clear that load must be transferred effec-tively from the matrix to the nanotubes Ajayan et alsuggested that load transfer could be improved by break-ing bundles down into individual tube fragments beforedispersing in the matrix Alternatively the bundles them-selves could be reinforced by crosslinking the tubes withinbundles (eg by chemical treatments or irradiation)

Single-walled nanotubeepoxy composites have alsobeen prepared by Biercuk et al61 The Vickers hardnessof the polymer was found to increase monotonicallywith addition of SWNTs up to a factor of 35 at 2 wt-nanotube loading Greatly enhanced thermal conducti-vities were also observed Thus samples loaded with1 wt- unpurified SWNT material showed a 70increase in thermal conductivity at 40 K rising to125 at room temperature

The preparation of SWNTpolyelectrolyte compositesby the layer-by-layer method has been described in theprevious section49 The films produced in this way wereshown to have exceptional mechanical properties Thusthe average ultimate tensile strength of the films wasfound to be 220 MPa with some measurements as highas 325 MPa This is several orders of magnitude greater

10 Image of epoxy coated MWNT subject to tensile

stress as result of electron irradiation telescopic

pull-out of outer tube layer has occurred59

9 TEM image of crack in MWNTpolystyrene thin film

induced by thermal stresses34

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 37

than the tensile strength of strong industrial plastics andindicates that the layer-by-layer method has great pro-mise for the production of strong nanotube-containingcomposites

Electrical and optical properties ofcarbon nanotubepolymer compositesThe effect of adding nanotubes on the electrical pro-perties of polymers has been studied by many groupsThese studies fall into a number of distinct categories Insome cases the nanotubes have been used to increase theconductivity of relatively low cost polymers as analternative to currently used fillers such as carbon blackOther studies have involved the incorporation of nano-tubes into conducting polymers such as polyaniline Stillfurther work has been aimed at specific applicationssuch as photovoltaic devices or supercapacitors In manycases improved mechanical properties have been avaluable by-product of the inclusion of nanotubes Theaddition of nanotubes to low cost bulk polymers isconsidered first

Electrostatic dissipationImproving the electrical conductivity of bulk polymers isimportant in a number of applications For example insome aircraft components enhanced conductivity isrequired to provide electrostatic discharge and electro-magnetic radio frequency interference protection Staticelectrical dissipation is also needed in other applicationsincluding computer housings and exterior automotiveparts With these types of application in mind Sandleret al investigated the electrical properties of nanotubeepoxy composites40 Matrix resistivities of about 100 V mwith filler volume fractions as low as 01 vol- wereachieved These figures represented an advance on thebest conductivity values previously obtained with carbonblack in the same epoxy matrix

Lozano and co-workers prepared composites ofcatalytically grown carbon nanofibres in a polypropy-lene matrix5162 The nanofibres were purified and func-tionalised to facilitate bonding with the matrix and highshear processing was used to disperse the fibres in thepolymer matrix Rheological and microscopic analysisshowed that a homogeneous dispersion of nanofibreswas formed and a percolation threshold for electricalconduction of 9ndash18 wt- was observed

These and other published studies suggest that carbonnanotubes have great promise in reducing electrostaticcharging of bulk polymers Nanotubes also have otheradvantages over conventional fillers such as carbonblack and carbon fibres in that they are more amenableto processing and can be more easily dispersed through-out the matrix In addition to the studies available in theopen literature there is undoubtedly a great deal ofcommercial work being carried out on the use of nano-tubes to reduce the electrostatic charging of plastics Infact nanotube containing plastics are already being usedin commercial products63 These include fuel lines inautomobiles where the nanotubes help to dissipate anydangerous charge which may build up Thermoplasticpolymers containing nanotubes are also used in someexterior automobile parts so that they can be earthedduring electrostatic painting

Conducting polymer compositesPolyaniline is a conducting polymer with many attrac-tive characteristics such as processability and environ-mental stability There has been interest in addingnanotubes to this polymer to provide both enhancedconductivity and improved mechanical properties Asmentioned above in the section lsquoPreparation of carbonnanotubepolymer compositesrsquo Cochet et al have pre-pared a MWNTpolyaniline composite by in situ poly-merisation44 The nanotubes were found to have astrong influence on the transport properties of thepolymer Thus the room temperature resistivity of thecomposite was found to be an order of magnitude lowerthan that of pure polyaniline while the low temperatureresistivity was much smaller than that of either poly-aniline or of MWNTs The temperature dependence ofresistivity was also weaker than that of polyanilinealone These observations were explained by assumingthat in situ polymerisation favours charge transfer betweenpolyaniline and MWNTs resulting in an overall materialwhich is more conducting than the starting componentsRaman studies indicated that these charge transferprocesses involved a site-selective interaction betweenthe quinoid ring of the polymer and the MWNTs Othergroups have also demonstrated excellent conductivity inMWNTpolyaniline composites6465 while a group fromDuPont have shown that SWNTpolyaniline compositescan be used as printable conductors for organic elec-tronics devices66

Optoelectronic propertiesSeveral groups have prepared nanotube compositesusing the conjugated polymer poly(p-phenylene viny-lene) (PPV) and its derivatives Polymers of this classhave electroluminescent properties and are widely usedin light-emitting diodes They are also used in photo-voltaic devices In an early study workers from Irelandand the USA described the synthesis of a compositematerial containing arc produced MWNTs in a PPVderivative PmPV67 The composite was prepared usinga solution mixing method and good wetting of the tubesby the polymer was achieved One of the aims of addingnanotubes to the polymer was to increase its electricalconductivity Previous attempts to achieve this by dopinghad resulted in degradation of the optical properties ofthe polymer In the case of nanotubes an increase in theelectrical conductivity of the polymer by up to eightorders of magnitude was seen with no concomitantdegradation of optical properties This was apparentlybecause the tubes acted as nanometric heat sinks pre-venting the build up of large thermal effects whichdamage these conjugated systems It was also shown inthis work that the composite could be used as theemissive layer in an organic light-emitting diode (LED)

As mentioned above several groups have been inter-ested in using nanotubepolymer composites in photo-voltaic devices The photovoltaic effect involves thegeneration of electron and hole pairs and their sub-sequent collection at opposite electrodes In organicmaterials the photon absorption excites an electron intothe conduction band which leads to a generation ofbound electronndashhole pairs or excitons These excitonshave to dissociate into free charges in order to be trans-ported to the electrodes Exciton dissociation can occurat a junction between a polymer and an electron acceptor

Harris Carbon nanotube composites

38 International Materials Reviews 2004 VOL 49 NO 1

molecule This allows the preferential transfer of the elec-trons into the electron acceptor molecule while leavingthe holes to be preferentially transported through thepolymer a process known as photoinduced chargetransfer In 1992 photoinduced charge transfer from aconducting polymer to C60 was discovered68 and thishas led to the fabrication of several efficient photo-voltaic systems using fullerenepolymer combinations69

Photovoltaic devices based on nanotubepolymer com-binations have now been fabricated by several groupsand this work will now be summarised

The earliest attempt to fabricate a nanotubepolymerphotovoltaic device appears to be the 1996 work ofRomero et al70 In this work a layer of MWNTs wasdeposited on Teflon and a PPV derivative was cast ontothis with a thin aluminium layer on top The devicedisplayed clear diode behaviour but only a small photo-current was observed when the junction was irradiatedwith light from a HendashNe laser This may have been dueto the relatively thick PPV active layer Ago et alfabricated a nanotube containing photovoltaic deviceby depositing a layer of PPV onto a glass-supportedMWNT film71 The interaction between MWNTs andPPV was investigated by means of photoluminescence(PL) measurements The photoluminescence efficiencyof the PPV was found to be markedly reduced by thepresence of the MWNT layer Since photoluminescenceis a result of radiative recombination a reduction mayimply improved exciton separation which is what onewishes to achieve in a photovoltaic device The devicefabricated by Ago et al is illustrated in Fig 11a Thisshowed a clear diode characteristic in the dark andupon illumination through the aluminium electrode aphotocurrent was observed A quantum efficiency approxi-mately twice that of the standard indiumndashtin oxidedevice was reported

Kymakis and Amaratunga used a slightly differentapproach to prepare a nanotubeconjugated polymerphotovoltaic device72 In this work composite SWNTpoly(3-octylthiophene) films were drop or spin cast froma solution onto indiumndashtin oxide (ITO) and quartz sub-strates to produce the arrangement shown in Fig 11bDevices were also prepared using the pure P3OT poly-mer A considerable enhancement of the photovoltaic

effect was observed with the P3OTSWNTs blend devicecompared with the pure polymer device

Non-linear optical propertiesAs mentioned above in the section lsquoStructure and pro-perties of carbon nanotubesrsquo it has been demonstratedthat suspensions of both SWNTs and MWNTs havenon-linear optical properties Nanotubepolymer com-posites may therefore have applications in optical devicesTo date however there has been little work on theoptical limiting properties of nanotube containing com-posites In one of the first studies in this area Chen et aldemonstrated third order optical non-linearity in a SWNTpolyimide composite73 Recently OrsquoFlaherty et al pre-pared a composite in which MWNTs were dispersed inthe polymer poly(99-di-n-octylfluorenyl-279-diyl) (PFO)74

Suspensions of the nanotubePFO composite in toluenewere prepared and the optical limiting properties of thesolutions were measured using nanosecond laser pulsesMaximum optical limiting was observed at carbonnanotube loadings in excess of 38 wt- relative to thepolymer

Carbon nanotubeceramic compositesCeramics have high stiffnesses and thermal stabilities butrelatively low breaking strengths Incorporating carbonnanotubes into a ceramic matrix might be expected toproduce a composite with both toughness and high tem-perature stability However achieving a homogeneousdispersion of tubes in an oxide with strong bondingbetween tubes and matrix presents rather more of achallenge than incorporating tubes into a polymer Variousapproaches have been adopted for incorporating nano-tubes into oxides such as alumina and silica and thesewill now be summarised

Peigney and co-workers have been among the pioneersin the synthesis of nanotubeoxide composites75ndash77 Theyhave developed an ingenious technique which involvesthe use of catalysts such as iron-containing a-aluminaPassing a CH4ndashH2 mixture over the catalysts resultedin the in situ growth of nanotubes on iron particlesproducing the precursor for a nanotubeoxide compo-site The resulting powders can then be hot pressed toform the final composite material The mechanicalproperties of these composites were initially disappoint-ing72 Thus the fracture strengths of the nanotubeFendashAl2O3 composites were only marginally higher than thatof Al2O3 and were generally markedly lower than thoseof the carbon-free FendashAl2O3 composites The fracturetoughness values were lower than or similar to that ofAl2O3 However SEM observations of composite frac-tures indicate that the nanotube bundles which are veryflexible could dissipate some fracture energy A similarmethod was used by An and Lim to prepare nanotubeAl2O3 composites in this case using C2H2 as the carbonsource78 Nanotube contents of up to 125 wt- wereachieved and a significant reinforcement effect wasreported

Achieving alignment of nanotubes in a ceramic matrixis much more difficult than in a polymer since ceramicsdo not lend themselves so easily to processing Howevera number of groups have shown that aligned nanotubemetal oxide composites can indeed be produced Peigney

ITO5indiumndashtin oxide PPV5poly(p-phenylene vinylene)

P3OT5poly(3-octylthiophene)

11 a Illustration of photovoltaic device fabricated by Ago

et al71 and b photovoltaic device fabricated by

Kymakis and Amaratunga72

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 39

et al exploited the fact that when their grains are main-tained at sufficiently low sizes (less than about 1 mm)many polycrystalline ceramics become superplasticallowing the use of thermomechanical processing forthe forming of these materials A series of cobalt andiron containing oxide catalysts were prepared77 Theprimary grain size in the powders was smaller than100 nm but in some of them the as produced primarygrains were strongly aggregated and required ball millingin order to reduce aggregate size to less than 1 mmNanotubes were then prepared in situ as describedabove and the powders were then extruded at hightemperatures in a graphite die under vacuum The bestresults were achieved with nanotubes incorporated intoFendashAl2O3 and FeCondashMgAl2O4 matrices In both casesthe nanotubes were not damaged by the processing andalignment of the nanotubes was clearly observed Infact the presence of the nanotubes seemed to facilitatethe superplastic forming by inhibiting grain growth andacting as lubricating agents It was demonstrated thatthe aligned nanotubemetal oxide composites showed ananisotropy of electrical conductivity which could beadjusted by controlling the nanotube content

One of the first studies of a SWNTceramic compositewas described by Zhan et al79 This group used thetechnique of spark plasma sintering (SPS) to prepareAl2O3 composites containing 10 vol- SWNTs Theadvantage of spark plasma sintering is that it allowsceramic powders to be annealed at lower temperaturesand for much shorter times than other sintering pro-cesses leading to the fabrication of fully dense ceramicsor composites with nanocrystalline microstructures undermild conditions Unlike some of the other processes usedto prepare nanotubeoxide composites spark plasmasintering does not damage the nanotubes and the nano-tube bundles could be seen to be located mainly at theboundaries of the Al2O3 grains in good contact with theAl2O3 matrix Before sintering the powders had beenmixed by ball milling which produced a reasonablyhomogeneous dispersion again without damaging thenanotubes The fracture toughnesses of the final com-posites were remarkable ndash about twice that of the fullydensified but unreinforced Al2O3 matrix In contrast toprevious work on nanotubeceramic composites thefracture toughness was found to increase with nanotubedensity This is thought to be due to the formation ofentangled networks of single-walled carbon nanotubeswhich may inhibit crack propagation A subsequentstudy of the nanotubeAl2O3 composites showed thatthey had high electrical conductivities80

Various techniques have been used to prepare carbonnanotubesilica composites Seeger et al described amethod which involved preparing a composite gel ofMWNTs with tetraethoxysilane (TEOS) and then sinter-ing this at 1150uC in argon81 A drawback of this tech-nique was that the sintering process led to a partialcrystallisation of the SiO2 resulting in a very inhomo-geneous matrix An alternative method which avoidsthese problems has been developed82 This involvesusing a Nd YAG laser to rapidly heat a TEOSnano-tube mixture resulting in partial melting of the matrixThis produced an amorphous silica matrix with nocrystallisation

A novel nanotubesilica composite material withpotentially useful optical properties has recently been

described by Han et al83 These workers synthesisedfilms consisting of silica spheres with diameters rangingfrom 200 to 650 nm Molybdenumcobalt catalyst parti-cles were then deposited onto the silica spheres and usedto catalyse the growth of SWNTs Since SWNTs arehighly non-linear and fast-switching materials they canbe incorporated into an optically confining environmentto achieve these characteristics at relatively low levels oflaser intensities

There have been a few studies on preparing nanotubecomposites using oxides other than alumina and silicaVincent et al reported the production of nanotubeTiO2

composites using solndashgel methods84 Possible applica-tions of the composites included optical non-linear waveguides and unidimensional conductive films Solndashgelmethods were also used by Sakamoto and Dunn toprepare SWNTV2O5 composites for use as electrodesin secondary lithium batteries85 Little work has beencarried out so far on incorporating nanotubes into non-oxide ceramics In one of the few published studies Maet al fabricated nanotubesilicon carbide composites bymixing SiC nanoparticles with 10 wt- MWNTs andhot pressing at 2000uC86 They found that the nanotubesplayed a strengthening and toughening role in thecomposite both bend strength and fracture toughnessincreased by about 10 compared to monolithic SiCThere have also been some initial studies on carbonnanotubecarbon fibre composites8788

Carbon nanotubemetal compositesComposite materials containing conventional carbonfibres in a metal matrix such as aluminium or mag-nesium are used in a number of specialist applica-tions8990 Such composites combine low density withhigh strength and modulus making them particularlyattractive to the aerospace industry There is growinginterest in the addition of carbon nanotubes to metalmatrices although only a small number of studies havebeen published to date Kuzumaki et al described thepreparation of a nanotubealuminium composite in 199891

Their method involved mixing a nanotube sample with afine aluminium powder mounting the mixture in a6 mm silver sheath and then drawing and heating thewire at 700uC in a vacuum furnace The result was acomposite wire in which the nanotubes were partiallyaligned along the axial direction The tensile strengths ofthe as prepared composite wires were comparable to thatof pure aluminium but the composite wires retained thisstrength after prolonged annealing at 600uC while thestrength of pure aluminium decreased by about 50after this treatment The same group subsequently pre-pared nanotubetitanium composites which showed alarge increase in hardness and modulus compared topure titanium92

Chen et al prepared nanotubenickelphosphorus com-posite coatings on stainless steel substrates using anelectroless plating method93 The coatings exhibited higherwear resistance and lower friction coefficients than SiCNiP and graphiteNiP composite coatings The samegroup prepared nanotubecopper composites by a powdermetallurgy technique and showed that the presence ofthe nanotubes again resulted in improved wear rates andfriction coefficients These results suggest that nanotube

Harris Carbon nanotube composites

40 International Materials Reviews 2004 VOL 49 NO 1

metal composites may have potential in tribologicalapplications

DiscussionThe subject of carbon nanotube composites has grownrapidly over the past 5 years or so into a vibrant andexciting area of research It is important to recognisehowever that this is still a new subject and the workcarried out so far may have hardly scratched the surfaceof what is possible For example the great majority ofthe studies that have been published to date haveinvolved the use of polymer matrices there is great scopefor more work on incorporating nanotubes into cera-mics metals and other forms of carbon Also most ofthe work up to now has been aimed at exploiting themechanical properties of carbon nanotubes rather thantheir electronic or optical properties although interest inthose is growing

When considering the mechanical performance ofcarbon nanotube composites it is clear that these dependstrongly on whether arc grown tubes or catalyticallyproduced tubes have been used In considering thecharacteristics of the two types of tube it is instructive tocompare the work of Qian et al34 with that of Watts andHsu57 Both incorporated the nanotubes into polymersand then used TEM to characterise the composites andstudy their fracture behaviour Qian et al used cataly-tically produced tubes and these were found to bewell dispersed throughout the polymer matrix with noclustering together of tubes However during compositefracture the tubes were often observed to break sug-gesting that they have relatively poor strength Wattsand Hsu on the other hand employed much moreperfect nanotubes produced by arc evaporation andTEM studies revealed superior mechanical propertieswith no observed breaking However the nanotubestended to cluster together so that during crack growththe tubes could quite easily slide past one another Thuswhile it is clear that the best mechanical performancewill be achieved with arc grown nanotubes it is impor-tant that the tubes be well dispersed throughout thematrix in order fully to exploit their properties

Equally important is achieving good bonding betweennanotubes and matrix Many studies have shown that thisinterfacial bonding can be enhanced by functionalising thenanotubes before embedding them in the matrix Theeffect of this can be seen in Fig 10 from the work ofGojny et al59 where bonding between MWNTs and anepoxy matrix is so strong that the outer shells of the tuberemain embedded in the matrix following pull-out Apossible disadvantage of functionalisation is that it couldweaken the nanotubes by disrupting the perfect cylindricalstructure This might be expected to be a particularproblem with single-walled tubes However functionali-sation of the nanotubes may not be necessary to producebonding with a polymer matrix Liao and Li havemodelled the interfacial characteristics of a nanotubepolystyrene composite assuming no covalent bondingbetween the tubes and the matrix94 It was found thatelectrostatic and van der Waals interactions contributedsignificantly to the interfacial stress transfer A nanotubepull-out simulation suggested that the interfacial shearstress of the nanotubepolystyrene system is about160 MPa significantly higher than for most carbon fibrereinforced polymer composite systems It is also worth

mentioning recent work by Star and co-workers9596

in which it was demonstrated that poly(5-alkoxy-m-phenylenevinylene)-co-[(25-dioctyloxy-p-phenylene)vinylene] (PAmPV) derivatives could be induced towrap around SWNTs forming strong non-covalentbonds

The quality of the nanotubes and the nature of theinterfacial bonding also has a strong influence on theelectrical properties of nanotubepolymer compositesArc grown tubes can have resistivities lower than that ofsingle crystal graphite but as already discussed theirelectronic properties depend on their precise structureAt present we do not have a reliable method of separat-ing metallic and semiconducting tubes so it is notpossible to exploit their full potential Catalytically pro-duced tubes are likely to be poorer conductors thanmetallic arc-grown tubes owing to their relatively dis-ordered structure Nevertheless catalytically grown tubesare sufficiently conducting to be useful in some com-mercial applications such as electrically conductive paintsThe influence of interfacial bonding on electronic pro-perties was demonstrated in the work of Cochet et al44

These workers prepared a MWNTpolyaniline com-posite by in situ polymerisation and found it to haveexcellent electrical transport properties This was attrib-uted to strong interaction between tubes and polymer

As has been shown in this review the possible appli-cations of carbon nanotube composites range widelyfrom electrically conductive paints to optical devicesOnly a relatively small number of these new materialshave been used commercially however and there arestill many problems to overcome before the full potentialof nanotube containing composites can be realised Notleast of these problems is the high cost of good qualitycarbon nanotubes Currently the commercial price ofunpurified arc-grownMWNTs is approximately $10000gwhile purified MWNTs cost around $30000g Cataly-tically grown MWNTs are much cheaper but as empha-sised throughout this review have generally inferiorproperties As far as SWNTs are concerned unpurifiedsamples cost around $5000g while purified SWNTs are$50000g As a comparison the price of gold is cur-rently about $1200g The development of new techni-ques for the low cost synthesis of high quality nanotubesis clearly essential This might be achieved throughimproved catalytic processes and there is currently alarge amount of work being carried out towards thisend Alternatively it may be possible to develop newnon-catalytic techniques for the synthesis of fullerene-related nanotubes In this connection it is interesting tonote that nanotubes can be produced by the high tem-perature heat treatment of conventional carbons such asmicroporous non-graphitising carbons9798 or carbonblack99100 Further work in this area would be of value

Acknowledgement

I thank Milo Shaffer for helpful comments

References1 P J F Harris lsquoCarbon nanotubes and related structures ndash new

materials for the twenty-first centuryrsquo 1999 Cambridge

Cambridge University Press

2 S Iijima Nature 1991 354 56ndash58

3 W Kratschmer L D Lamb K Fostiropoulos and D R

Huffman Nature 1990 347 354ndash358

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 41

4 S Iijima and T Ichihashi Nature 1993 363 603ndash605

5 D S Bethune C H Kiang M S De Vries G Gorman R Savoy

J Vasquez and R Beyers Nature 1993 363 605ndash607

6 H Dai A G Rinzler P Nikolaev A Thess D T Colbert and

R E Smalley Chem Phys Lett 1996 260 471ndash475

7 M M J Treacy T W Ebbesen and J M Gibson Nature 1996

381 678ndash680

8 E W Wong P E Sheehan and C M Lieber Science 1997 277

1971ndash1975

9 M R Falvo G J Clary R M Taylor V Chi F P Brooks

S Washburn and R Superfine Nature 1997 389 582ndash584

10 K Anazawa K Shimotani C Manabe H Watanabe and M

Shimizu Appl Phys Lett 2002 81 739ndash741

11 P Kim L Shi A Majumdar and P L McEuen Phys Rev Lett

2001 87 paper 215502

12 E T Thostenson Z F Ren and T W Chou Compos Sci

Technol 2001 61 1899ndash1912

13 K T Lau and D Hui Composites Part B Engineering 2002 33

263ndash277

14 Y Huh J Y Lee J Cheon Y K Hong J Y Koo T J Lee and

C J Lee J Mater Chem 2003 13 2297ndash2300

15 M Endo T Koyama and Y Hishiyama Jpn J Appl Phys 1976

15 2073ndash2076

16 S C Lyu B C Liu T J Lee Z Y Liu C W Yang C Y Park

and C J Lee Chem Commun 2003 734ndash735

17 M-F Yu O Lourie M J Dyer K Moloni T F Kelly and R S

Ruoff Science 2000 287 637ndash640

18 P Poncharal Z L Wang D Ugarte and W A De Heer Science

1999 283 1513ndash1516

19 J-P Salvetat A J Kulik J-M Bonard G A D Briggs

T Stockli K Metenier S Bonnamy F Beguin N A Burnham

and L Forro Adv Mater 1999 11 161ndash165

20 M-F Yu B S Files S Arepalli and R S Ruoff Phys Rev Lett

2000 84 5552ndash5555

21 M Terrones Int Mater Rev to be published

22 J W Mintmire B I Dunlap and C T White Phys Rev Lett

1992 68 631ndash634

23 R Saito M Fujita G Dresselhaus and M S Dresselhaus Phys

Rev B 1992 46 1804ndash1811

24 N Hamada S Sawada and A Oshiyama Phys Rev Lett 1992

68 1579ndash1581

25 T W Ebbesen H J Lezec H Hiura J W Bennett H F Ghaemi

and T Thio Nature 1996 382 54ndash56

26 J E Riggs D B Walker D L Carroll and Y-P Sun J Phys

Chem B 2000 104 7071ndash7076

27 S C Tsang Y K Chen P J F Harris and M L H Green

Nature 1994 372 159ndash162

28 R M Lago S C Tsang K L Lu Y K Chen and M L H

Green Chem Commun 1995 1355ndash1356

29 Y-P Sun K F Fu Y Lin and W Huang Acc Chem Res 2002

35 1096ndash1104

30 S B Sinnott J Nanosci Nanotechnol 2002 2113ndash123

31 M S P Shaffer X Fan and A H Windle Carbon 1998 36

1603ndash1612

32 M S P Shaffer and A H Windle Adv Mater 1999 11 937ndash941

33 D E Hill Y Lin A M Rao L F Allard and Y-P Sun

Macromolecules 2002 35 9466ndash9471

34 D Qian E C Dickey R Andrews and T Rantell Appl Phys

Lett 2000 76 2868ndash2870

35 R Andrews D Jacques D Qian and T Rantell Acc Chem Res

2002 35 1008ndash1017

36 R Haggenmueller H H Gommans A G Rinzler J E Fischer

and K I Winey Chem Phys Lett 2000 330 219ndash225

37 P Potschke T D Fornes and D R Paul Polymer 2002 43 3247ndash

3255

38 M Sennett E Welsh J B Wright W Z Li J G Wen and Z F

Ren Appl Phys A 2003 76 111ndash113

39 P M Ajayan O Stephan C Colliex and D Trauth Science 1994

265 1212ndash1214

40 J Sandler M S P Shaffer T Prasse W Bauhofer K Schulte

and H Windle Polymer 1999 40 5967ndash5971

41 A Allaoui S Bai H M Cheng and J B Bai Compos Sci

Technol 2002 62 1993ndash1998

42 X Xu M M Thwe C Shearwood and K Liao Appl Phys Lett

2002 81 2833ndash2835

43 D Puglia L Valentini and J M Kenny J Appl Polymer Sci

2003 88 452ndash458

44 M Cochet W K Maser A M Benito M A Callejas M T

Martınez J-M Benoit J Schreiber and O Chauvet Chem

Commun 2001 1450ndash1451

45 M S P Shaffer and K Koziol Chem Commun 2002 2074ndash

2075

46 C Park Z Ounaies K A Watson R E Crooks J Smith S E

Lowther J W Connell E J Siochi J S Harrison and T L St

Clair Chem Phys Lett 2002 364 303ndash308

47 K Jurewicz S Delpeux V Bertagna F Beguin and E

Frackowiak Chem Phys Lett 2001 347 36ndash40

48 M Hughes G Z Chen M S P Shaffer D J Fray and A H

Windle Chem Mater 2002 14 1610ndash1613

49 A A Mamedov N A Kotov M Prato D M Guldi J P

Wicksted and A Hirsch Nature Mater 2002 1 190ndash194

50 J H Rouse and P T Lillehei Nano Lett 2003 3 59ndash62

51 K Lozano and E V Barrera J Appl Polym Sci 2001 79 125ndash

133

52 L Valentini J Biagiotti J M Kenny and S Santucci Compos

Sci Technol 2003 63 1149ndash1153

53 J Sandler P Werner M S P Shaffer V Demchuk V Altstadt

and A H Windle Composites Part A Appl Sci Manuf 2002 33

1033ndash1039

54 R Czerw Z X Guo P M Ajayan Y P Sun and D L Carroll

Nano Lett 2001 1 423ndash427

55 B P Grady F Pompeo R L Shambaugh and D E Resasco

J Phys Chem B 2002 106 5852ndash5858

56 M Cadek J N Coleman V Barron K Hedicke and W J Blau

Appl Phys Lett 2002 81 5123ndash5125

57 P C P Watts and W K Hsu Nanotechnology 2003 14 L7ndashL10

58 L S Schadler S C Giannaris and P M Ajayan Appl Phys Lett

1998 73 3842ndash3844

59 F H Gojny J Nastalczyk Z Roslaniec and K Schulte Chem

Phys Lett 2003 370 820ndash824

60 P M Ajayan L S Schadler C Giannaris and A Rubio Adv

Mater 2000 12 750ndash753

61 M J Biercuk M C Llaguno M Radosavljevic J K Hyun A T

Johnson and J E Fischer Appl Phys Lett 2002 80 2767ndash2769

62 K Lozano J Bonilla-Rios and E V Barrera J Appl Polym Sci

2001 80 1162ndash1172

63 V Jamieson New Sci 15 Mar 2003 30

64 H Zengin W S Zhou J Y Jin W R Czerw D W Smith

L Echegoyen D L Carroll S H Foulger and J Ballato Adv

Mater 2002 14 1480ndash1483

65 J Deng X Ding W Zhang Y Peng J Wang X Long P Li and

A S C Chan Eur Polym J 2002 38 2497ndash2501

66 G B Blanchet C R Fincher and F Gao Appl Phys Lett 2003

82 1290ndash1292

67 S A Curran P M Ajayan W J Blau D L Carroll J N

Coleman A B Dalton A P Davey A Drury B McCarthy

S Maier and A A Strevens Adv Mater 1998 10 1091ndash1093

68 N S Sariciftci L Smilowitz A J Heeger and F Wudl Science

1992 258 1474ndash1476

69 N Camaioni L Garlaschelli A Geri M Maggini G Possamai

and G Ridolfi J Mater Chem 2002 12 2065ndash2070

70 D B Romero M Carrard W De Heer and L Zuppiroli

Adv Mater 1996 8 899ndash902

71 H Ago K Petritsch M S P Shaffer A H Windle and R H

Friend Adv Mater 1999 11 1281ndash1285

72 E Kymakis and G A J Amaratunga Appl Phys Lett 2002 80

112ndash114

73 Y-C Chen N R Raravikar L S Schadler P M Ajayan Y-P

Zhao T-M Lu G-C Wang and X-C Zhang Appl Phys Lett

2002 81 975ndash977

74 S A OrsquoFlaherty R Murphy S V Hold M Cadek J N

Coleman and W J Blau J Phys Chem B 2003 107 958ndash964

75 A Peigney C Laurent F Dobigeon and A Rousset J Mater

Res 1997 12 613ndash615

76 E Flahaut A Peigney C Laurent C Marliere F Chastel and

A Rousset Acta Mater 2000 48 3803ndash3812

77 A Peigney E Flahaut C Laurent F Chastel and A Rousset

Chem Phys Lett 2002 352 20ndash25

78 J W An and D S Lim J Ceram Process Res 2002 3 201ndash204

79 G-D Zhan J D Kuntz J L Wan and A K Mukherjee Nature

Mater 2003 2 38ndash42

80 G-D Zhan J D Kuntz J E Garay and A K Mukherjee Appl

Phys Lett 2003 83 1228ndash1230

81 T Seeger T Kohler T Frauenheim N Grobert M Ruhle

M Terrones and G Seifert Chem Commun 2002 34ndash35

82 T Seeger G De La Fuente W K Maser A M Benito M A

Callejas and M T Martınez Nanotechnology 2003 14 184ndash187

83 H Han S Vijayalakshmi A Lan Z Iqbal H Grebel E Lalanne

and A M Johnson Appl Phys Lett 2003 82 1458ndash1460

Harris Carbon nanotube composites

42 International Materials Reviews 2004 VOL 49 NO 1

84 P Vincent A Brioude C Journet S Rabaste S T Purcell J Le

Brusq and J C Plenet J Non-Cryst Sol 2002 311 130ndash137

85 J S Sakamoto and B Dunn J Electrochem Soc 2002 149 A26ndashA30

86 R Z Ma J Wu B Q Wei J Liang and D H Wu J Mater Sci

1998 33 5243ndash5246

87 R Andrews D Jacques A M Rao T Rantell F Derbyshire

Y Chen J Chen and R C Haddon Appl Phys Lett 1999 75

1329ndash1331

88 E T Thostenson W Z Li D Z Wang Z F Ren and T W

Chou J Appl Phys 2002 91 6034ndash6037

89 D D L Chung lsquoCarbon fiber compositesrsquo chap 7 1994 Boston

Butterworth-Heinemann

90 S Rawal JOM 2001 53 14ndash17

91 T Kuzumaki K Miyazawa H Ichinose and K Ito J Mater

Res 1998 13 2445ndash2449

92 T Kuzumaki O Ujiie H Ichinose and K Ito Adv Eng Mater

2000 2 416ndash418

93 W X Chen J P Tu L Y Wang H Y Gan Z D Xu and X B

Zhang Carbon 2003 41 215ndash222

94 K Liao and S Li Appl Phys Lett 2001 79 4225ndash4227

95 A Star J F Stoddart D Steuerman M Diehl A Boukai E W

Wong X Yang S W Chung H Choi and J R Heath

Angewandte Chemie-Int Ed 2001 40 1721ndash1725

96 A Star Y Liu K Grant L Ridvan J F Stoddart D W

Steuerman M R Diehl A Boukai and J R Heath

Macromolecules 2003 36 553ndash560

97 P J F Harris S C Tsang J B Claridge and M L H Green

J Chem Soc Faraday Trans 1994 90 2799ndash2802

98 A A Setlur S P Doherty J Y Dai and R P H Chang Appl

Phys Lett 2000 76 3008ndash3010

99 S P Doherty D B Buchholz B J Li and R P H Chang

J Mater Res 2003 18 941ndash949

100 D B Buchholz S P Doherty and R P H Chang Carbon 2003

41 1625ndash1634

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 43

Page 2: Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P. J. F. Harris * Carbon nanotubes are molecular-scale tubes of graphitic carbon with

Fig 1b The terms lsquozigzagrsquo and lsquoarmchairrsquo refer to thearrangement of hexagons around the circumferenceThere is a third class of structure lsquochiralrsquo in which thehexagons are arranged helically around the tube axis asshown in Fig 1c Experimentally the tubes are generallyless perfect than the idealised versions shown in Fig 1and as already noted may be either multiwalled orsingle-walled The abbreviations MWNT and SWNTare commonly used for multiwalled and single-wallednanotubes respectively Figure 2a shows a typicalsample of MWNT-containing cathodic soot at moderatemagnification As can be seen the nanotubes are accom-panied by other material including nanoparticles (hollowfullerene-related structures) and some disordered car-bon At high resolution the individual layers making upthe concentric tubes can be clearly imaged as in Fig 2bThe multiwalled tubes range in length from a few tensof nanometres to several micrometres and in outerdiameter from about 25 to 30 nm The end-caps of thetubes are sometimes symmetrical in shape but moreoften asymmetric Conical structures of the type shownin Fig 3 are commonly observed This type of structureis believed to result from the presence of a singlepentagon (at the position indicated by the arrow) withfive further pentagons at the apex of the cone

Multiwalled tubes produced by catalytic methods aregenerally much more defective than those made by arcevaporation as can be seen in Fig 414 However theirproperties can be improved by high temperature heattreatment15 Some typical micrographs of SWNTs canbe seen in Fig 516 As a result of their very smalldiameters the tubes are much more flexible thanMWNTs and are therefore often observed to be curledand looped rather than straight Because SWNTs fre-quently have a narrow range of diameters they tend tocluster together to form lsquoropesrsquo

There have been a large number of studies boththeoretical and experimental of the mechanical proper-ties of carbon nanotubes The first measurements ofstiffness were made in 1996 by Treacy et al7 whocarried out in situ measurements of the intrinsic thermalvibrations of multiwalled carbon nanotubes in a TEMThese indicated that the tubes had extremely high elasticmoduli but the error on the measurements was largeestimates for the Youngrsquos modulus ranged from

a armchair (n m)5(5 5) b zigzag (n m)5(9 0) c chiral

(n m)5(10 5)

1 Three classes of carbon nanotube structure

2 a TEM image of nanotube-containing soot and b

higher magnification image of individual tubes

3 Image of typical multiwalled nanotube (MWNT) cap with arrow indicating approximate position of pentagonal ring

Harris Carbon nanotube composites

32 International Materials Reviews 2004 VOL 49 NO 1

410 GPa to 415 TPa with an average of 18 TPaRather more accurate measurements were made a usinga lsquonanostressing stagersquo inside a scanning electron micro-scope (SEM)17 These experiments indicated that theYoungrsquos modulus of the outermost layers of multiwalledcarbon nanotubes varied from 270 to 950 GPa Thetensile strengths of individual nanotubes were also deter-mined in this work and values ranging from 11 to63 GPa were found Measurements of Youngrsquos modulushave also been carried out using atomic force micro-scopy Wong et al8 described a method of fixing nano-tubes at one end and then determining the bendingforce F of the fibres as a function of displacement d Forsmall deflections linear Fndashd curves were observed andthe results implied a value of y128 TPa for theYoungrsquos modulus However for deflections larger than10u an abrupt change in the slope of the Fndashd curve wasseen This was attributed to elastic buckling of the tubesIn TEM studies of nanotubes buckled tubes are quiteoften seen as a result of stresses imposed by the way thetubes are supported on the grid In some cases thebuckles are regularly spaced as shown in Fig 618 It hasbeen demonstrated that when the stress constraininga tube is released the tube can return to its originalstraight form This behaviour sets nanotubes apart fromconventional carbon fibres and all other currently knownfibres which are much more susceptible to fracture whensubject to stress beyond the elastic limit It also raisesthe fascinating possibility that composites could beproduced which would snap back into place followingdeformation

As already noted catalytically produced MWNTs aremuch less perfect in structure than arc grown tubes andare therefore considerably less stiff Salvetat et al usedatomic force microscopy to record forcendashdisplacementcurves for nanotubes prepared by the catalytic decom-position of acetylene19 This produced an averagevalue for the Youngrsquos modulus of 27 GPa The tensile

strengths of catalytically produced MWNTs have notyet been measured However it is possible that thesemight be relatively high the defects in these imperfectstructures might be expected to assist stress transferbetween layers

Determining the mechanical properties of SWNTspresents an even greater challenge than for MWNTsHowever some measurements have been made using ananostressing stage20 These produced values for theYoungrsquos modulus in the range 320ndash1470 GPa (mean1002 GPa) and an average breaking strength of 30 GPa

There have also been a substantial number of studiesof the electronic properties of carbon nanotubes121

Early theoretical work showed that nanotubes could bemetallic or semiconducting depending on their structureand diameter22ndash24 Ebbesen et al confirmed experimen-tally that the electronic properties of nanotubes couldvary widely from tube to tube25 both semiconductingand metallic behaviour has been observed Measuredresistivities fall into the approximate range 005 mV mndash10 mV m This compares with a value of approximately04 mV m for high quality single crystal graphite(parallel to the c axis) and 0017 mV m for copper Theresistivities of catalytically produced nanotubes arelikely to be similar to those of disordered carbons ieof the order of 10ndash100 mV m but experimental measure-ments on this type of tube have not yet been carried outThe thermal and optical properties of carbon nanotubes

a low magnification TEM image b HRTEM image of

typical nanotube

4 TEM images of multiwalled carbon nanotubes synthe-

sised by decomposition of C2H2 at 850uC over cobalt

catalyst14

a low magnification SEM image b high magnification

TEM image

5 Images of single-walled carbon nanotubes prepared

by decomposition of C2H2 at 950uC over FeMoAl2O3

catalyst16

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 33

have been studied less intensively than the electronicproperties but some interesting results have been obtainedThe high thermal conductivities of nanotubes have alreadybeen mentioned11 It has also been reported that carbonnanotube suspensions have non-linear optical proper-ties suggesting possible applications as optical limiters26

Another area which has been the subject of muchresearch is the chemical modification of carbon nano-tubes Chemical treatments of nanotubes were first carriedout in an attempt to selectively remove the end-caps anduse them as molecular containers It was demonstratedin 1994 that nanotube caps could be removed by treat-ment with hot nitric acid apparently as a result ofselective oxidation at the pentagonal rings27 In sub-sequent work it was demonstrated by acidndashbase titra-tion and by temperature programmed decompositionthat this acid treatment resulted in the formation ofsurface carboxylic acid and other groups on the tubesurfaces28 These surface acid groups can then be furthermodified in a variety of ways29 Functionalisation hasbeen widely used to facilitate the dispersion of nano-tubes in solvents as a prelude to the preparation ofcomposite materials and can also help to improvebonding between tubes and matrix30

Preparation of carbon nanotubepolymercompositesA commonly used method for preparing nanotubepolymer composites has involved mixing nanotube dis-persions with solutions of the polymer and then eva-porating the solvents in a controlled way As noted inthe previous section the nanotubes are often pretreatedchemically to facilitate solubilisation Shaffer et al forexample have demonstrated that acid treatments enablestable aqueous solutions of catalytically producedMWNTsto be prepared31 These workers then showed that ananotubePVA (polyvinyl alcohol) composite could be

prepared simply by mixing one of these aqueous nano-tube dispersions with an aqueous solution of the poly-mer and then casting the mixtures as films andevaporating the water32 Solution based methods havealso been used to produce nanotubepolystyrene com-posites Thus Hill et al have solubilised both SWNTsand MWNTs by functionalising with a polystyrenecopolymer33 This was achieved by first acid treating thetubes and then carrying out esterification of the surface-bound carboxylic acids The polymer-modified carbonnanotubes were shown to be soluble in common organicsolvents To prepare composites polystyrene was dissolvedin the nanotube solution and nanotubepolystyrene thinfilms prepared using wet casting Functionalisation ofthe nanotubes is not always necessary in order to pre-pare a polymer composite Qian et al used a high energyultrasonic probe to disperse MWNTs in toluene andthen mixed the dispersed suspension with a dilute solu-tion of polystyrene in toluene again with ultrasonicagitation34 The low viscosity of the polymer solutionallowed the nanotubes to move freely through thematrix The mixture was cast on glass and the solventremoved to yield MWNT-doped films Specimens of thecomposite were prepared for TEM by placing drops ofthe mixed solution on copper TEM grids On evapora-tion of the toluene a thin film of the composite was leftsuspended across the grid bars A micrograph of one ofthese films is shown in Fig 7 and this illustrates theexcellent dispersion that was achieved Note that in situpolymerisation has also been used to prepare nanotubepolystyrene composites as is discussed below

The solution mixing approach is limited to polymersthat freely dissolve in common solvents An alternativeis to use thermoplastic polymers (ie polymers thatsoften and melt when heated) and then apply meltprocessing techniques Thus shear mixing can be used toproduce a homogeneous dispersion of nanotubes andextrusion to produce nanotube alignment or to fabricateartefacts in the required form by injection moulding

7 TEM image of MWNTpolystyrene thin film (arrows

indicate regions of polymer shrinkage) inset shows

distribution of nanotube lengths346 High resolution TEM image of deformed nanotube

showing periodic buckling18

Harris Carbon nanotube composites

34 International Materials Reviews 2004 VOL 49 NO 1

Andrews et al used shear mixing to disperse catalytically-produced nanotubes in a range of polymers includinghigh impact polystyrene acrylonitrile butadiene styreneand polypropylene35 The dispersion of nanotubes wasdetermined as a function of mixing energy and tem-perature and the composites were formed into fibres andthin films Haggenmueller et al used a combination ofsolvent casting and melt mixing methods to disperseSWNTs in polymethylmethacrylate (PMMA)36 Compositefibres with a high degree of nanotube orientation wereproduced by melt spinning

Polycarbonate is another thermoplastic polymer thathas been used as a matrix for nanotube composites3738

Sennet et al have used melt processing techniques todisperse and align carbon nanotubes in polycarbonate38

Dispersion was achieved by mixing the catalyticallyproduced MWNTs and SWNTs with polycarbonateresin in a conical twin-screw extruder and alignment wascarried out using a fibre spinning apparatus By opti-mising mixing time and fibre draw rates excellent dis-persion and alignment were accomplished

Nanotubeepoxy composites have also been preparedby a number of groups39ndash43 The earliest work in thisarea was carried out by Ajayan et al39 In 1994 thisgroup described a study in which purified tubes wereembedded into an epoxy resin which was then cut intothin slices with a diamond knife Their initial aim indoing this was not to produce a composite material butto obtain thin TEM specimens which it was hoped wouldcontain cross-sectional images of nanotubes However nocross-sections were observed following sectioning insteadthe nanotubes were found to have become aligned in thedirection of the knife movement as shown in Fig 8According to Ajayan et al the alignment is primarily aconsequence of extensional or shear flow of the matrixproduced by the cutting Of course alignment can bemuch more readily achieved when the epoxy is in a liquidstate Xu et al showed that some alignment of nanotubesin nanotubeepoxy composites can been achieved by usingspin coating methods42 The cure reaction of a SWNT

epoxy resin composite has been analysed using thermalanalysis and Raman spectroscopy43 The SWNTs werefound to have a catalysing effect on the cure reaction

An alternative method for preparing nanotubepoly-mer composites is to use the monomer rather than thepolymer as a starting material and then carry out in situpolymerisation Cochet et al were among the first to usethis method to prepare a MWNTpolyaniline compo-site44 Nanotubes prepared by arc evaporation wereused and were sonicated in a solution of HCl to achievea dispersion The aniline monomer again in HCl wasadded to the suspension and a solution of an oxidantwas slowly added with constant sonication and coolingSonication was continued in an ice bath for 2 h and thecomposite was then obtained by filtering rinsing anddrying In this way it was possible to prepare com-posites with high MWNT loadings (up to 50 wt-)Transport measurements on the composite revealedmajor changes in the electronic behaviour confirmingstrong interaction between nanotubes and polymer asdiscussed below A number of other nanotubepolymercomposites have been prepared using in situ poly-merisation including MWNTpolystyrene45 and SWNTpolyimide46 Several groups have used electrochemicalpolymerisation to grow porous composite films ofMWNTand polypyrrole for use as supercapacitors4748

In the techniques described so far the aim has been toproduce composites in which the nanotubes are dis-tributed evenly throughout the polymer (either ran-domly oriented or aligned) For some applicationshowever a layered arrangement is advantageous Photo-voltaic devices containing layers of nanotubes or nanotubepolymer composites have been prepared as discussed inthe section lsquoElectrical and optical properties of carbonnanotubepolymer compositesrsquo below The preparationof multilayer SWNTpolymer composites with excellentmechanical properties has also been described49 Themethod involved the layer-by-layer deposition of SWNTsand polymer onto a substrate followed by crosslinkingIn this way composites with SWNT loadings as high as

8 Alignment of nanotubes in polymer matrix following cutting with microtome arrows indicate buckled nanotubes39

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 35

50 wt- could be obtained Coating was carried out byalternate dipping of a glass slide or silicon wafer intodispersions of SWNTs (stabilised by acid treatment) andpolymer solutions The layers were held together by vander Waals forces and by electrostatic attraction betweenthe negatively charged SWNTs and a positively chargedpolyelectrolyte such as branched polyethyleneimine (PEI)Coatings containing up to 50 SWNTPEI bilayers couldbe built up in this way When the procedure was com-plete the multilayer films were heated to 120uC topromote crosslinking The films could then be lifted offthe substrate to obtain uniform free-standing mem-branes A rather similar method was used to synthesisemultilayer composites films of the polyelectrolyte poly-diallyldimethylammonium chloride and SWNTs50

Before proceeding to discuss the properties of carbonnanotubepolymer composites in detail it is worth brieflyconsidering the effect that the inclusion of nanotubescan have on polymer microstructure Although thistopic has not yet been widely studied a number ofgroups have demonstrated that the crystallisation andmorphology of the polymers can be strongly affected bysmall additions of nanotubes For example Lozano andBarrera51 showed that the addition of catalytically growncarbon nanofibres can promote nucleation of polypro-pylene Similar observations were made by Valentiniet al52 The crystallinity of a polymers has also beenshown to increase with the addition of carbon nano-fibres5354 Changes in the crystal structure of polymerscan also be induced by nanotubes55 More work on theseeffects would be welcome

Mechanical properties of carbonnanotubepolymer compositesMuch of the work on the preparation of nanotubepolymer composites has been driven by a desire toexploit the tubesrsquo stiffness and strength Even where theinterest has been focused on other properties the abilityof nanotubes to improve the mechanical characteristicsof a polymer has often been a valuable added benefit Inearly work the measured properties were often dis-appointing but such studies were of value in under-standing the reasons for composite failure and inidentifying the critical issues which need to be addressed

Shaffer and Windle were among the first to carry outa systematic study of the mechanical properties ofnanotubepolymer composites Their technique for pre-paring MWNTPVA composites was described in theprevious section32 The tensile elastic moduli of thecomposite films were assessed in a dynamic mechanicalthermal analyser as a function of nanotube loading andtemperature The stiffness of the composites at roomtemperature was relatively low From the theory deve-loped for short-fibre composites a nanotube elasticmodulus of 150 MPa was obtained from the roomtemperature experimental data This value is well belowthe values reported for isolated nanotubes even defec-tive catalytically grown tubes have measured stiffness ofthe order of 30 GPa19 The low value may have more todo with poor stress transfer than with the weakness ofthe nanotubes themselves Above the glass transitiontemperature of the polymer (y85uC) the nanotubeshad a more significant effect on the properties of thecomposite

Cadek et al achieved rather better results in 2002with composites made from PVA and arc producedMWNTs56 The composites were prepared by solutionmixing and films were formed on glass substrates bydrop casting The presence of 1 wt- nanotubes increasedthe Youngrsquos modulus and hardness of the PVA byfactors of 18 and 16 respectively This improvementover the early results of Shaffer and Windle reflects boththe superior quality of the tubes and stronger interfacialbonding Evidence for the latter was provided by TEMstudies

The mechanical properties of nanotubepolystyrenecomposites were studied by Qian and co-workers3435

Two samples of catalytically produced MWNTs wereused one with an average length of y15 mm the otherwith an average length of y50 mm With the addition of1 wt- nanotubes they achieved a 36 and 42 increasein the elastic stiffness of the polymer for short and longtubes respectively (the neat polymer modulus wasy12 GPa) In both cases a 25 increase in the tensilestrength was observed Theoretical calculations carriedout assuming a nanotube modulus of 450 GPa pro-duced increases of a 48 and 62 in the stiffness of thepolymer for short and long tubes respectively The closeagreement within y10 between the experimental andtheoretically predicted composite moduli indicated thatthe external tensile loads were successfully transmittedto the nanotubes across the tubepolymer interface Athigher nanotube concentrations the changes in themechanical properties of the composites were morepronounced35 For MWNTpolystyrene composites con-taining from 25 to 25 vol- nanotubes Youngrsquosmodulus increased progressively from 19 to 45 GPawith the major increases occurring when the MWNTcontent was at or above about 10 vol- However thedependence of tensile strength on nanotube concentra-tion was more complex At the lower concentrations((10 vol-) the tensile strength decreased from theneat polymer value ofy40 MPa only exceeding it whenthe MWNT content was above 15 vol-

In an attempt to understand the tensile fracturemechanisms of the nanotubepolystyrene compositesQian et al performed deformation studies inside aTEM34 Their elegant method of preparing TEM speci-mens has been described above Focusing the electronbeam onto the thin film resulted in local thermal stresseswhich initiated cracks in the composite The propaga-tion speed of the crack could be controlled by varyingthe beam flux onto the sample These in situ TEMobservations showed that the cracks tended to nucleateat low nanotube density areas then propagate alongweak nanotubepolymer interfaces or relatively low nano-tube density regions The nanotubes became alignedperpendicular to the crack direction and bridged thecrack faces in the wake as shown in Fig 9 When thecrack opening displacement exceeded y800 nm thenanotubes began to break andor pull out of the matrixThe fact that some of the pulled-out tubes do not appearto be coated with polymer suggests that there is room forimprovement in nanotubendashpolymer bonding

Watts and Hsu have also used TEM to study thefracture behaviour of a nanotubepolymer composite57

Their composite consisted of arc grown MWNTsembedded in a diblock copolymer known as MPCndashDEA To prepare specimens of the composite for TEM

Harris Carbon nanotube composites

36 International Materials Reviews 2004 VOL 49 NO 1

the tubes were mixed with the polymer in an acidifiedaqueous solution and drops of this were applied to TEMgrids which were then dried in vacuum As in the workof Qian et al cracks were produced in the material byirradiation with an electron beam but a low acceleratingvoltage (75 kV) was used to avoid massive disruption ofthe thin composite film Individual tubes and bundles oftubes were observed to become stretched as the cracksgrew wider and pull-out was often seen In contrast tothe work of Qian et al however the tubes did notbreak This was probably because high quality arcgrown nanotubes were used rather than catalyticallyproduced tubes

The early work on nanotubeepoxy composites byAjayan et al has been described in the previoussection39 In 1998 Ajayan and co-workers described astudy of the mechanical behaviour of MWNTepoxycomposites in both tension and compression58 It wasfound that the compression modulus was higher thanthe tensile modulus indicating that load transfer to thenanotubes in the composite was much higher in com-pression The authors suggested that during load trans-fer to MWNTs only the outer layers are stressed intension whereas all the layers respond in compression

Nanotubeepoxy thin films prepared by Xu et alshowed excellent mechanical properties42 Compared toneat resin thin films for which the elastic modulus was42 GPa a 20 increase in modulus was seen when01 wt- MWNTs were added This was attributed topartial alignment of the MWNTs induced by spincoating Fracture behaviour of the films was investigatedby SEM and showed that pulled-out tubes were oftencovered with polymer suggesting strong interfacial adhe-sion Similar results were reported by Gojny et al59 Thisgroup showed that functionalising arc produced MWNTswith amine groups greatly enhanced bonding with theepoxy matrix As a result the outer shells of the tubesoften remained embedded in the matrix following pull-out as shown in Fig 10

All of the work described so far in this section hasinvolved multiwalled carbon nanotubes Ajayan et al

have described a study of SWNT-containing compo-sites60 Two types of composite were prepared first anSWNTepoxy composite and second a pressed pelletcontaining SWNTs and carbonaceous soot materialformed during nanotube synthesis The SWNTs wereused as-prepared ie generally grouped into bundlesrather than separated into individual tubes The com-posites were loaded axially in tension until failureoccurred and the fracture surfaces were then examinedin detail using SEM In some cases the tube bundleswere found to have been pulled out of the matrix duringthe deformation and fracture of the composites while inother cases the nanotubes were not entirely pulled outbut were stretched between two fracture surfaces Theauthors believed that during pull-out the nanotubeswere sliding axially within the ropes and that the failureobserved at large crack distances was not failure ofindividual tubes but the bundles pulling apart Studiesby micro-Raman spectroscopy supported this hypoth-esis It was found that little shift in the second order A1g

band occurred as a result of the application of axialtension showing that the individual nanotubes were notbeing significantly stretched To take full advantage ofthe high Youngrsquos modulus of SWNTs in polymer com-posites it is clear that load must be transferred effec-tively from the matrix to the nanotubes Ajayan et alsuggested that load transfer could be improved by break-ing bundles down into individual tube fragments beforedispersing in the matrix Alternatively the bundles them-selves could be reinforced by crosslinking the tubes withinbundles (eg by chemical treatments or irradiation)

Single-walled nanotubeepoxy composites have alsobeen prepared by Biercuk et al61 The Vickers hardnessof the polymer was found to increase monotonicallywith addition of SWNTs up to a factor of 35 at 2 wt-nanotube loading Greatly enhanced thermal conducti-vities were also observed Thus samples loaded with1 wt- unpurified SWNT material showed a 70increase in thermal conductivity at 40 K rising to125 at room temperature

The preparation of SWNTpolyelectrolyte compositesby the layer-by-layer method has been described in theprevious section49 The films produced in this way wereshown to have exceptional mechanical properties Thusthe average ultimate tensile strength of the films wasfound to be 220 MPa with some measurements as highas 325 MPa This is several orders of magnitude greater

10 Image of epoxy coated MWNT subject to tensile

stress as result of electron irradiation telescopic

pull-out of outer tube layer has occurred59

9 TEM image of crack in MWNTpolystyrene thin film

induced by thermal stresses34

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 37

than the tensile strength of strong industrial plastics andindicates that the layer-by-layer method has great pro-mise for the production of strong nanotube-containingcomposites

Electrical and optical properties ofcarbon nanotubepolymer compositesThe effect of adding nanotubes on the electrical pro-perties of polymers has been studied by many groupsThese studies fall into a number of distinct categories Insome cases the nanotubes have been used to increase theconductivity of relatively low cost polymers as analternative to currently used fillers such as carbon blackOther studies have involved the incorporation of nano-tubes into conducting polymers such as polyaniline Stillfurther work has been aimed at specific applicationssuch as photovoltaic devices or supercapacitors In manycases improved mechanical properties have been avaluable by-product of the inclusion of nanotubes Theaddition of nanotubes to low cost bulk polymers isconsidered first

Electrostatic dissipationImproving the electrical conductivity of bulk polymers isimportant in a number of applications For example insome aircraft components enhanced conductivity isrequired to provide electrostatic discharge and electro-magnetic radio frequency interference protection Staticelectrical dissipation is also needed in other applicationsincluding computer housings and exterior automotiveparts With these types of application in mind Sandleret al investigated the electrical properties of nanotubeepoxy composites40 Matrix resistivities of about 100 V mwith filler volume fractions as low as 01 vol- wereachieved These figures represented an advance on thebest conductivity values previously obtained with carbonblack in the same epoxy matrix

Lozano and co-workers prepared composites ofcatalytically grown carbon nanofibres in a polypropy-lene matrix5162 The nanofibres were purified and func-tionalised to facilitate bonding with the matrix and highshear processing was used to disperse the fibres in thepolymer matrix Rheological and microscopic analysisshowed that a homogeneous dispersion of nanofibreswas formed and a percolation threshold for electricalconduction of 9ndash18 wt- was observed

These and other published studies suggest that carbonnanotubes have great promise in reducing electrostaticcharging of bulk polymers Nanotubes also have otheradvantages over conventional fillers such as carbonblack and carbon fibres in that they are more amenableto processing and can be more easily dispersed through-out the matrix In addition to the studies available in theopen literature there is undoubtedly a great deal ofcommercial work being carried out on the use of nano-tubes to reduce the electrostatic charging of plastics Infact nanotube containing plastics are already being usedin commercial products63 These include fuel lines inautomobiles where the nanotubes help to dissipate anydangerous charge which may build up Thermoplasticpolymers containing nanotubes are also used in someexterior automobile parts so that they can be earthedduring electrostatic painting

Conducting polymer compositesPolyaniline is a conducting polymer with many attrac-tive characteristics such as processability and environ-mental stability There has been interest in addingnanotubes to this polymer to provide both enhancedconductivity and improved mechanical properties Asmentioned above in the section lsquoPreparation of carbonnanotubepolymer compositesrsquo Cochet et al have pre-pared a MWNTpolyaniline composite by in situ poly-merisation44 The nanotubes were found to have astrong influence on the transport properties of thepolymer Thus the room temperature resistivity of thecomposite was found to be an order of magnitude lowerthan that of pure polyaniline while the low temperatureresistivity was much smaller than that of either poly-aniline or of MWNTs The temperature dependence ofresistivity was also weaker than that of polyanilinealone These observations were explained by assumingthat in situ polymerisation favours charge transfer betweenpolyaniline and MWNTs resulting in an overall materialwhich is more conducting than the starting componentsRaman studies indicated that these charge transferprocesses involved a site-selective interaction betweenthe quinoid ring of the polymer and the MWNTs Othergroups have also demonstrated excellent conductivity inMWNTpolyaniline composites6465 while a group fromDuPont have shown that SWNTpolyaniline compositescan be used as printable conductors for organic elec-tronics devices66

Optoelectronic propertiesSeveral groups have prepared nanotube compositesusing the conjugated polymer poly(p-phenylene viny-lene) (PPV) and its derivatives Polymers of this classhave electroluminescent properties and are widely usedin light-emitting diodes They are also used in photo-voltaic devices In an early study workers from Irelandand the USA described the synthesis of a compositematerial containing arc produced MWNTs in a PPVderivative PmPV67 The composite was prepared usinga solution mixing method and good wetting of the tubesby the polymer was achieved One of the aims of addingnanotubes to the polymer was to increase its electricalconductivity Previous attempts to achieve this by dopinghad resulted in degradation of the optical properties ofthe polymer In the case of nanotubes an increase in theelectrical conductivity of the polymer by up to eightorders of magnitude was seen with no concomitantdegradation of optical properties This was apparentlybecause the tubes acted as nanometric heat sinks pre-venting the build up of large thermal effects whichdamage these conjugated systems It was also shown inthis work that the composite could be used as theemissive layer in an organic light-emitting diode (LED)

As mentioned above several groups have been inter-ested in using nanotubepolymer composites in photo-voltaic devices The photovoltaic effect involves thegeneration of electron and hole pairs and their sub-sequent collection at opposite electrodes In organicmaterials the photon absorption excites an electron intothe conduction band which leads to a generation ofbound electronndashhole pairs or excitons These excitonshave to dissociate into free charges in order to be trans-ported to the electrodes Exciton dissociation can occurat a junction between a polymer and an electron acceptor

Harris Carbon nanotube composites

38 International Materials Reviews 2004 VOL 49 NO 1

molecule This allows the preferential transfer of the elec-trons into the electron acceptor molecule while leavingthe holes to be preferentially transported through thepolymer a process known as photoinduced chargetransfer In 1992 photoinduced charge transfer from aconducting polymer to C60 was discovered68 and thishas led to the fabrication of several efficient photo-voltaic systems using fullerenepolymer combinations69

Photovoltaic devices based on nanotubepolymer com-binations have now been fabricated by several groupsand this work will now be summarised

The earliest attempt to fabricate a nanotubepolymerphotovoltaic device appears to be the 1996 work ofRomero et al70 In this work a layer of MWNTs wasdeposited on Teflon and a PPV derivative was cast ontothis with a thin aluminium layer on top The devicedisplayed clear diode behaviour but only a small photo-current was observed when the junction was irradiatedwith light from a HendashNe laser This may have been dueto the relatively thick PPV active layer Ago et alfabricated a nanotube containing photovoltaic deviceby depositing a layer of PPV onto a glass-supportedMWNT film71 The interaction between MWNTs andPPV was investigated by means of photoluminescence(PL) measurements The photoluminescence efficiencyof the PPV was found to be markedly reduced by thepresence of the MWNT layer Since photoluminescenceis a result of radiative recombination a reduction mayimply improved exciton separation which is what onewishes to achieve in a photovoltaic device The devicefabricated by Ago et al is illustrated in Fig 11a Thisshowed a clear diode characteristic in the dark andupon illumination through the aluminium electrode aphotocurrent was observed A quantum efficiency approxi-mately twice that of the standard indiumndashtin oxidedevice was reported

Kymakis and Amaratunga used a slightly differentapproach to prepare a nanotubeconjugated polymerphotovoltaic device72 In this work composite SWNTpoly(3-octylthiophene) films were drop or spin cast froma solution onto indiumndashtin oxide (ITO) and quartz sub-strates to produce the arrangement shown in Fig 11bDevices were also prepared using the pure P3OT poly-mer A considerable enhancement of the photovoltaic

effect was observed with the P3OTSWNTs blend devicecompared with the pure polymer device

Non-linear optical propertiesAs mentioned above in the section lsquoStructure and pro-perties of carbon nanotubesrsquo it has been demonstratedthat suspensions of both SWNTs and MWNTs havenon-linear optical properties Nanotubepolymer com-posites may therefore have applications in optical devicesTo date however there has been little work on theoptical limiting properties of nanotube containing com-posites In one of the first studies in this area Chen et aldemonstrated third order optical non-linearity in a SWNTpolyimide composite73 Recently OrsquoFlaherty et al pre-pared a composite in which MWNTs were dispersed inthe polymer poly(99-di-n-octylfluorenyl-279-diyl) (PFO)74

Suspensions of the nanotubePFO composite in toluenewere prepared and the optical limiting properties of thesolutions were measured using nanosecond laser pulsesMaximum optical limiting was observed at carbonnanotube loadings in excess of 38 wt- relative to thepolymer

Carbon nanotubeceramic compositesCeramics have high stiffnesses and thermal stabilities butrelatively low breaking strengths Incorporating carbonnanotubes into a ceramic matrix might be expected toproduce a composite with both toughness and high tem-perature stability However achieving a homogeneousdispersion of tubes in an oxide with strong bondingbetween tubes and matrix presents rather more of achallenge than incorporating tubes into a polymer Variousapproaches have been adopted for incorporating nano-tubes into oxides such as alumina and silica and thesewill now be summarised

Peigney and co-workers have been among the pioneersin the synthesis of nanotubeoxide composites75ndash77 Theyhave developed an ingenious technique which involvesthe use of catalysts such as iron-containing a-aluminaPassing a CH4ndashH2 mixture over the catalysts resultedin the in situ growth of nanotubes on iron particlesproducing the precursor for a nanotubeoxide compo-site The resulting powders can then be hot pressed toform the final composite material The mechanicalproperties of these composites were initially disappoint-ing72 Thus the fracture strengths of the nanotubeFendashAl2O3 composites were only marginally higher than thatof Al2O3 and were generally markedly lower than thoseof the carbon-free FendashAl2O3 composites The fracturetoughness values were lower than or similar to that ofAl2O3 However SEM observations of composite frac-tures indicate that the nanotube bundles which are veryflexible could dissipate some fracture energy A similarmethod was used by An and Lim to prepare nanotubeAl2O3 composites in this case using C2H2 as the carbonsource78 Nanotube contents of up to 125 wt- wereachieved and a significant reinforcement effect wasreported

Achieving alignment of nanotubes in a ceramic matrixis much more difficult than in a polymer since ceramicsdo not lend themselves so easily to processing Howevera number of groups have shown that aligned nanotubemetal oxide composites can indeed be produced Peigney

ITO5indiumndashtin oxide PPV5poly(p-phenylene vinylene)

P3OT5poly(3-octylthiophene)

11 a Illustration of photovoltaic device fabricated by Ago

et al71 and b photovoltaic device fabricated by

Kymakis and Amaratunga72

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 39

et al exploited the fact that when their grains are main-tained at sufficiently low sizes (less than about 1 mm)many polycrystalline ceramics become superplasticallowing the use of thermomechanical processing forthe forming of these materials A series of cobalt andiron containing oxide catalysts were prepared77 Theprimary grain size in the powders was smaller than100 nm but in some of them the as produced primarygrains were strongly aggregated and required ball millingin order to reduce aggregate size to less than 1 mmNanotubes were then prepared in situ as describedabove and the powders were then extruded at hightemperatures in a graphite die under vacuum The bestresults were achieved with nanotubes incorporated intoFendashAl2O3 and FeCondashMgAl2O4 matrices In both casesthe nanotubes were not damaged by the processing andalignment of the nanotubes was clearly observed Infact the presence of the nanotubes seemed to facilitatethe superplastic forming by inhibiting grain growth andacting as lubricating agents It was demonstrated thatthe aligned nanotubemetal oxide composites showed ananisotropy of electrical conductivity which could beadjusted by controlling the nanotube content

One of the first studies of a SWNTceramic compositewas described by Zhan et al79 This group used thetechnique of spark plasma sintering (SPS) to prepareAl2O3 composites containing 10 vol- SWNTs Theadvantage of spark plasma sintering is that it allowsceramic powders to be annealed at lower temperaturesand for much shorter times than other sintering pro-cesses leading to the fabrication of fully dense ceramicsor composites with nanocrystalline microstructures undermild conditions Unlike some of the other processes usedto prepare nanotubeoxide composites spark plasmasintering does not damage the nanotubes and the nano-tube bundles could be seen to be located mainly at theboundaries of the Al2O3 grains in good contact with theAl2O3 matrix Before sintering the powders had beenmixed by ball milling which produced a reasonablyhomogeneous dispersion again without damaging thenanotubes The fracture toughnesses of the final com-posites were remarkable ndash about twice that of the fullydensified but unreinforced Al2O3 matrix In contrast toprevious work on nanotubeceramic composites thefracture toughness was found to increase with nanotubedensity This is thought to be due to the formation ofentangled networks of single-walled carbon nanotubeswhich may inhibit crack propagation A subsequentstudy of the nanotubeAl2O3 composites showed thatthey had high electrical conductivities80

Various techniques have been used to prepare carbonnanotubesilica composites Seeger et al described amethod which involved preparing a composite gel ofMWNTs with tetraethoxysilane (TEOS) and then sinter-ing this at 1150uC in argon81 A drawback of this tech-nique was that the sintering process led to a partialcrystallisation of the SiO2 resulting in a very inhomo-geneous matrix An alternative method which avoidsthese problems has been developed82 This involvesusing a Nd YAG laser to rapidly heat a TEOSnano-tube mixture resulting in partial melting of the matrixThis produced an amorphous silica matrix with nocrystallisation

A novel nanotubesilica composite material withpotentially useful optical properties has recently been

described by Han et al83 These workers synthesisedfilms consisting of silica spheres with diameters rangingfrom 200 to 650 nm Molybdenumcobalt catalyst parti-cles were then deposited onto the silica spheres and usedto catalyse the growth of SWNTs Since SWNTs arehighly non-linear and fast-switching materials they canbe incorporated into an optically confining environmentto achieve these characteristics at relatively low levels oflaser intensities

There have been a few studies on preparing nanotubecomposites using oxides other than alumina and silicaVincent et al reported the production of nanotubeTiO2

composites using solndashgel methods84 Possible applica-tions of the composites included optical non-linear waveguides and unidimensional conductive films Solndashgelmethods were also used by Sakamoto and Dunn toprepare SWNTV2O5 composites for use as electrodesin secondary lithium batteries85 Little work has beencarried out so far on incorporating nanotubes into non-oxide ceramics In one of the few published studies Maet al fabricated nanotubesilicon carbide composites bymixing SiC nanoparticles with 10 wt- MWNTs andhot pressing at 2000uC86 They found that the nanotubesplayed a strengthening and toughening role in thecomposite both bend strength and fracture toughnessincreased by about 10 compared to monolithic SiCThere have also been some initial studies on carbonnanotubecarbon fibre composites8788

Carbon nanotubemetal compositesComposite materials containing conventional carbonfibres in a metal matrix such as aluminium or mag-nesium are used in a number of specialist applica-tions8990 Such composites combine low density withhigh strength and modulus making them particularlyattractive to the aerospace industry There is growinginterest in the addition of carbon nanotubes to metalmatrices although only a small number of studies havebeen published to date Kuzumaki et al described thepreparation of a nanotubealuminium composite in 199891

Their method involved mixing a nanotube sample with afine aluminium powder mounting the mixture in a6 mm silver sheath and then drawing and heating thewire at 700uC in a vacuum furnace The result was acomposite wire in which the nanotubes were partiallyaligned along the axial direction The tensile strengths ofthe as prepared composite wires were comparable to thatof pure aluminium but the composite wires retained thisstrength after prolonged annealing at 600uC while thestrength of pure aluminium decreased by about 50after this treatment The same group subsequently pre-pared nanotubetitanium composites which showed alarge increase in hardness and modulus compared topure titanium92

Chen et al prepared nanotubenickelphosphorus com-posite coatings on stainless steel substrates using anelectroless plating method93 The coatings exhibited higherwear resistance and lower friction coefficients than SiCNiP and graphiteNiP composite coatings The samegroup prepared nanotubecopper composites by a powdermetallurgy technique and showed that the presence ofthe nanotubes again resulted in improved wear rates andfriction coefficients These results suggest that nanotube

Harris Carbon nanotube composites

40 International Materials Reviews 2004 VOL 49 NO 1

metal composites may have potential in tribologicalapplications

DiscussionThe subject of carbon nanotube composites has grownrapidly over the past 5 years or so into a vibrant andexciting area of research It is important to recognisehowever that this is still a new subject and the workcarried out so far may have hardly scratched the surfaceof what is possible For example the great majority ofthe studies that have been published to date haveinvolved the use of polymer matrices there is great scopefor more work on incorporating nanotubes into cera-mics metals and other forms of carbon Also most ofthe work up to now has been aimed at exploiting themechanical properties of carbon nanotubes rather thantheir electronic or optical properties although interest inthose is growing

When considering the mechanical performance ofcarbon nanotube composites it is clear that these dependstrongly on whether arc grown tubes or catalyticallyproduced tubes have been used In considering thecharacteristics of the two types of tube it is instructive tocompare the work of Qian et al34 with that of Watts andHsu57 Both incorporated the nanotubes into polymersand then used TEM to characterise the composites andstudy their fracture behaviour Qian et al used cataly-tically produced tubes and these were found to bewell dispersed throughout the polymer matrix with noclustering together of tubes However during compositefracture the tubes were often observed to break sug-gesting that they have relatively poor strength Wattsand Hsu on the other hand employed much moreperfect nanotubes produced by arc evaporation andTEM studies revealed superior mechanical propertieswith no observed breaking However the nanotubestended to cluster together so that during crack growththe tubes could quite easily slide past one another Thuswhile it is clear that the best mechanical performancewill be achieved with arc grown nanotubes it is impor-tant that the tubes be well dispersed throughout thematrix in order fully to exploit their properties

Equally important is achieving good bonding betweennanotubes and matrix Many studies have shown that thisinterfacial bonding can be enhanced by functionalising thenanotubes before embedding them in the matrix Theeffect of this can be seen in Fig 10 from the work ofGojny et al59 where bonding between MWNTs and anepoxy matrix is so strong that the outer shells of the tuberemain embedded in the matrix following pull-out Apossible disadvantage of functionalisation is that it couldweaken the nanotubes by disrupting the perfect cylindricalstructure This might be expected to be a particularproblem with single-walled tubes However functionali-sation of the nanotubes may not be necessary to producebonding with a polymer matrix Liao and Li havemodelled the interfacial characteristics of a nanotubepolystyrene composite assuming no covalent bondingbetween the tubes and the matrix94 It was found thatelectrostatic and van der Waals interactions contributedsignificantly to the interfacial stress transfer A nanotubepull-out simulation suggested that the interfacial shearstress of the nanotubepolystyrene system is about160 MPa significantly higher than for most carbon fibrereinforced polymer composite systems It is also worth

mentioning recent work by Star and co-workers9596

in which it was demonstrated that poly(5-alkoxy-m-phenylenevinylene)-co-[(25-dioctyloxy-p-phenylene)vinylene] (PAmPV) derivatives could be induced towrap around SWNTs forming strong non-covalentbonds

The quality of the nanotubes and the nature of theinterfacial bonding also has a strong influence on theelectrical properties of nanotubepolymer compositesArc grown tubes can have resistivities lower than that ofsingle crystal graphite but as already discussed theirelectronic properties depend on their precise structureAt present we do not have a reliable method of separat-ing metallic and semiconducting tubes so it is notpossible to exploit their full potential Catalytically pro-duced tubes are likely to be poorer conductors thanmetallic arc-grown tubes owing to their relatively dis-ordered structure Nevertheless catalytically grown tubesare sufficiently conducting to be useful in some com-mercial applications such as electrically conductive paintsThe influence of interfacial bonding on electronic pro-perties was demonstrated in the work of Cochet et al44

These workers prepared a MWNTpolyaniline com-posite by in situ polymerisation and found it to haveexcellent electrical transport properties This was attrib-uted to strong interaction between tubes and polymer

As has been shown in this review the possible appli-cations of carbon nanotube composites range widelyfrom electrically conductive paints to optical devicesOnly a relatively small number of these new materialshave been used commercially however and there arestill many problems to overcome before the full potentialof nanotube containing composites can be realised Notleast of these problems is the high cost of good qualitycarbon nanotubes Currently the commercial price ofunpurified arc-grownMWNTs is approximately $10000gwhile purified MWNTs cost around $30000g Cataly-tically grown MWNTs are much cheaper but as empha-sised throughout this review have generally inferiorproperties As far as SWNTs are concerned unpurifiedsamples cost around $5000g while purified SWNTs are$50000g As a comparison the price of gold is cur-rently about $1200g The development of new techni-ques for the low cost synthesis of high quality nanotubesis clearly essential This might be achieved throughimproved catalytic processes and there is currently alarge amount of work being carried out towards thisend Alternatively it may be possible to develop newnon-catalytic techniques for the synthesis of fullerene-related nanotubes In this connection it is interesting tonote that nanotubes can be produced by the high tem-perature heat treatment of conventional carbons such asmicroporous non-graphitising carbons9798 or carbonblack99100 Further work in this area would be of value

Acknowledgement

I thank Milo Shaffer for helpful comments

References1 P J F Harris lsquoCarbon nanotubes and related structures ndash new

materials for the twenty-first centuryrsquo 1999 Cambridge

Cambridge University Press

2 S Iijima Nature 1991 354 56ndash58

3 W Kratschmer L D Lamb K Fostiropoulos and D R

Huffman Nature 1990 347 354ndash358

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 41

4 S Iijima and T Ichihashi Nature 1993 363 603ndash605

5 D S Bethune C H Kiang M S De Vries G Gorman R Savoy

J Vasquez and R Beyers Nature 1993 363 605ndash607

6 H Dai A G Rinzler P Nikolaev A Thess D T Colbert and

R E Smalley Chem Phys Lett 1996 260 471ndash475

7 M M J Treacy T W Ebbesen and J M Gibson Nature 1996

381 678ndash680

8 E W Wong P E Sheehan and C M Lieber Science 1997 277

1971ndash1975

9 M R Falvo G J Clary R M Taylor V Chi F P Brooks

S Washburn and R Superfine Nature 1997 389 582ndash584

10 K Anazawa K Shimotani C Manabe H Watanabe and M

Shimizu Appl Phys Lett 2002 81 739ndash741

11 P Kim L Shi A Majumdar and P L McEuen Phys Rev Lett

2001 87 paper 215502

12 E T Thostenson Z F Ren and T W Chou Compos Sci

Technol 2001 61 1899ndash1912

13 K T Lau and D Hui Composites Part B Engineering 2002 33

263ndash277

14 Y Huh J Y Lee J Cheon Y K Hong J Y Koo T J Lee and

C J Lee J Mater Chem 2003 13 2297ndash2300

15 M Endo T Koyama and Y Hishiyama Jpn J Appl Phys 1976

15 2073ndash2076

16 S C Lyu B C Liu T J Lee Z Y Liu C W Yang C Y Park

and C J Lee Chem Commun 2003 734ndash735

17 M-F Yu O Lourie M J Dyer K Moloni T F Kelly and R S

Ruoff Science 2000 287 637ndash640

18 P Poncharal Z L Wang D Ugarte and W A De Heer Science

1999 283 1513ndash1516

19 J-P Salvetat A J Kulik J-M Bonard G A D Briggs

T Stockli K Metenier S Bonnamy F Beguin N A Burnham

and L Forro Adv Mater 1999 11 161ndash165

20 M-F Yu B S Files S Arepalli and R S Ruoff Phys Rev Lett

2000 84 5552ndash5555

21 M Terrones Int Mater Rev to be published

22 J W Mintmire B I Dunlap and C T White Phys Rev Lett

1992 68 631ndash634

23 R Saito M Fujita G Dresselhaus and M S Dresselhaus Phys

Rev B 1992 46 1804ndash1811

24 N Hamada S Sawada and A Oshiyama Phys Rev Lett 1992

68 1579ndash1581

25 T W Ebbesen H J Lezec H Hiura J W Bennett H F Ghaemi

and T Thio Nature 1996 382 54ndash56

26 J E Riggs D B Walker D L Carroll and Y-P Sun J Phys

Chem B 2000 104 7071ndash7076

27 S C Tsang Y K Chen P J F Harris and M L H Green

Nature 1994 372 159ndash162

28 R M Lago S C Tsang K L Lu Y K Chen and M L H

Green Chem Commun 1995 1355ndash1356

29 Y-P Sun K F Fu Y Lin and W Huang Acc Chem Res 2002

35 1096ndash1104

30 S B Sinnott J Nanosci Nanotechnol 2002 2113ndash123

31 M S P Shaffer X Fan and A H Windle Carbon 1998 36

1603ndash1612

32 M S P Shaffer and A H Windle Adv Mater 1999 11 937ndash941

33 D E Hill Y Lin A M Rao L F Allard and Y-P Sun

Macromolecules 2002 35 9466ndash9471

34 D Qian E C Dickey R Andrews and T Rantell Appl Phys

Lett 2000 76 2868ndash2870

35 R Andrews D Jacques D Qian and T Rantell Acc Chem Res

2002 35 1008ndash1017

36 R Haggenmueller H H Gommans A G Rinzler J E Fischer

and K I Winey Chem Phys Lett 2000 330 219ndash225

37 P Potschke T D Fornes and D R Paul Polymer 2002 43 3247ndash

3255

38 M Sennett E Welsh J B Wright W Z Li J G Wen and Z F

Ren Appl Phys A 2003 76 111ndash113

39 P M Ajayan O Stephan C Colliex and D Trauth Science 1994

265 1212ndash1214

40 J Sandler M S P Shaffer T Prasse W Bauhofer K Schulte

and H Windle Polymer 1999 40 5967ndash5971

41 A Allaoui S Bai H M Cheng and J B Bai Compos Sci

Technol 2002 62 1993ndash1998

42 X Xu M M Thwe C Shearwood and K Liao Appl Phys Lett

2002 81 2833ndash2835

43 D Puglia L Valentini and J M Kenny J Appl Polymer Sci

2003 88 452ndash458

44 M Cochet W K Maser A M Benito M A Callejas M T

Martınez J-M Benoit J Schreiber and O Chauvet Chem

Commun 2001 1450ndash1451

45 M S P Shaffer and K Koziol Chem Commun 2002 2074ndash

2075

46 C Park Z Ounaies K A Watson R E Crooks J Smith S E

Lowther J W Connell E J Siochi J S Harrison and T L St

Clair Chem Phys Lett 2002 364 303ndash308

47 K Jurewicz S Delpeux V Bertagna F Beguin and E

Frackowiak Chem Phys Lett 2001 347 36ndash40

48 M Hughes G Z Chen M S P Shaffer D J Fray and A H

Windle Chem Mater 2002 14 1610ndash1613

49 A A Mamedov N A Kotov M Prato D M Guldi J P

Wicksted and A Hirsch Nature Mater 2002 1 190ndash194

50 J H Rouse and P T Lillehei Nano Lett 2003 3 59ndash62

51 K Lozano and E V Barrera J Appl Polym Sci 2001 79 125ndash

133

52 L Valentini J Biagiotti J M Kenny and S Santucci Compos

Sci Technol 2003 63 1149ndash1153

53 J Sandler P Werner M S P Shaffer V Demchuk V Altstadt

and A H Windle Composites Part A Appl Sci Manuf 2002 33

1033ndash1039

54 R Czerw Z X Guo P M Ajayan Y P Sun and D L Carroll

Nano Lett 2001 1 423ndash427

55 B P Grady F Pompeo R L Shambaugh and D E Resasco

J Phys Chem B 2002 106 5852ndash5858

56 M Cadek J N Coleman V Barron K Hedicke and W J Blau

Appl Phys Lett 2002 81 5123ndash5125

57 P C P Watts and W K Hsu Nanotechnology 2003 14 L7ndashL10

58 L S Schadler S C Giannaris and P M Ajayan Appl Phys Lett

1998 73 3842ndash3844

59 F H Gojny J Nastalczyk Z Roslaniec and K Schulte Chem

Phys Lett 2003 370 820ndash824

60 P M Ajayan L S Schadler C Giannaris and A Rubio Adv

Mater 2000 12 750ndash753

61 M J Biercuk M C Llaguno M Radosavljevic J K Hyun A T

Johnson and J E Fischer Appl Phys Lett 2002 80 2767ndash2769

62 K Lozano J Bonilla-Rios and E V Barrera J Appl Polym Sci

2001 80 1162ndash1172

63 V Jamieson New Sci 15 Mar 2003 30

64 H Zengin W S Zhou J Y Jin W R Czerw D W Smith

L Echegoyen D L Carroll S H Foulger and J Ballato Adv

Mater 2002 14 1480ndash1483

65 J Deng X Ding W Zhang Y Peng J Wang X Long P Li and

A S C Chan Eur Polym J 2002 38 2497ndash2501

66 G B Blanchet C R Fincher and F Gao Appl Phys Lett 2003

82 1290ndash1292

67 S A Curran P M Ajayan W J Blau D L Carroll J N

Coleman A B Dalton A P Davey A Drury B McCarthy

S Maier and A A Strevens Adv Mater 1998 10 1091ndash1093

68 N S Sariciftci L Smilowitz A J Heeger and F Wudl Science

1992 258 1474ndash1476

69 N Camaioni L Garlaschelli A Geri M Maggini G Possamai

and G Ridolfi J Mater Chem 2002 12 2065ndash2070

70 D B Romero M Carrard W De Heer and L Zuppiroli

Adv Mater 1996 8 899ndash902

71 H Ago K Petritsch M S P Shaffer A H Windle and R H

Friend Adv Mater 1999 11 1281ndash1285

72 E Kymakis and G A J Amaratunga Appl Phys Lett 2002 80

112ndash114

73 Y-C Chen N R Raravikar L S Schadler P M Ajayan Y-P

Zhao T-M Lu G-C Wang and X-C Zhang Appl Phys Lett

2002 81 975ndash977

74 S A OrsquoFlaherty R Murphy S V Hold M Cadek J N

Coleman and W J Blau J Phys Chem B 2003 107 958ndash964

75 A Peigney C Laurent F Dobigeon and A Rousset J Mater

Res 1997 12 613ndash615

76 E Flahaut A Peigney C Laurent C Marliere F Chastel and

A Rousset Acta Mater 2000 48 3803ndash3812

77 A Peigney E Flahaut C Laurent F Chastel and A Rousset

Chem Phys Lett 2002 352 20ndash25

78 J W An and D S Lim J Ceram Process Res 2002 3 201ndash204

79 G-D Zhan J D Kuntz J L Wan and A K Mukherjee Nature

Mater 2003 2 38ndash42

80 G-D Zhan J D Kuntz J E Garay and A K Mukherjee Appl

Phys Lett 2003 83 1228ndash1230

81 T Seeger T Kohler T Frauenheim N Grobert M Ruhle

M Terrones and G Seifert Chem Commun 2002 34ndash35

82 T Seeger G De La Fuente W K Maser A M Benito M A

Callejas and M T Martınez Nanotechnology 2003 14 184ndash187

83 H Han S Vijayalakshmi A Lan Z Iqbal H Grebel E Lalanne

and A M Johnson Appl Phys Lett 2003 82 1458ndash1460

Harris Carbon nanotube composites

42 International Materials Reviews 2004 VOL 49 NO 1

84 P Vincent A Brioude C Journet S Rabaste S T Purcell J Le

Brusq and J C Plenet J Non-Cryst Sol 2002 311 130ndash137

85 J S Sakamoto and B Dunn J Electrochem Soc 2002 149 A26ndashA30

86 R Z Ma J Wu B Q Wei J Liang and D H Wu J Mater Sci

1998 33 5243ndash5246

87 R Andrews D Jacques A M Rao T Rantell F Derbyshire

Y Chen J Chen and R C Haddon Appl Phys Lett 1999 75

1329ndash1331

88 E T Thostenson W Z Li D Z Wang Z F Ren and T W

Chou J Appl Phys 2002 91 6034ndash6037

89 D D L Chung lsquoCarbon fiber compositesrsquo chap 7 1994 Boston

Butterworth-Heinemann

90 S Rawal JOM 2001 53 14ndash17

91 T Kuzumaki K Miyazawa H Ichinose and K Ito J Mater

Res 1998 13 2445ndash2449

92 T Kuzumaki O Ujiie H Ichinose and K Ito Adv Eng Mater

2000 2 416ndash418

93 W X Chen J P Tu L Y Wang H Y Gan Z D Xu and X B

Zhang Carbon 2003 41 215ndash222

94 K Liao and S Li Appl Phys Lett 2001 79 4225ndash4227

95 A Star J F Stoddart D Steuerman M Diehl A Boukai E W

Wong X Yang S W Chung H Choi and J R Heath

Angewandte Chemie-Int Ed 2001 40 1721ndash1725

96 A Star Y Liu K Grant L Ridvan J F Stoddart D W

Steuerman M R Diehl A Boukai and J R Heath

Macromolecules 2003 36 553ndash560

97 P J F Harris S C Tsang J B Claridge and M L H Green

J Chem Soc Faraday Trans 1994 90 2799ndash2802

98 A A Setlur S P Doherty J Y Dai and R P H Chang Appl

Phys Lett 2000 76 3008ndash3010

99 S P Doherty D B Buchholz B J Li and R P H Chang

J Mater Res 2003 18 941ndash949

100 D B Buchholz S P Doherty and R P H Chang Carbon 2003

41 1625ndash1634

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 43

Page 3: Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P. J. F. Harris * Carbon nanotubes are molecular-scale tubes of graphitic carbon with

410 GPa to 415 TPa with an average of 18 TPaRather more accurate measurements were made a usinga lsquonanostressing stagersquo inside a scanning electron micro-scope (SEM)17 These experiments indicated that theYoungrsquos modulus of the outermost layers of multiwalledcarbon nanotubes varied from 270 to 950 GPa Thetensile strengths of individual nanotubes were also deter-mined in this work and values ranging from 11 to63 GPa were found Measurements of Youngrsquos modulushave also been carried out using atomic force micro-scopy Wong et al8 described a method of fixing nano-tubes at one end and then determining the bendingforce F of the fibres as a function of displacement d Forsmall deflections linear Fndashd curves were observed andthe results implied a value of y128 TPa for theYoungrsquos modulus However for deflections larger than10u an abrupt change in the slope of the Fndashd curve wasseen This was attributed to elastic buckling of the tubesIn TEM studies of nanotubes buckled tubes are quiteoften seen as a result of stresses imposed by the way thetubes are supported on the grid In some cases thebuckles are regularly spaced as shown in Fig 618 It hasbeen demonstrated that when the stress constraininga tube is released the tube can return to its originalstraight form This behaviour sets nanotubes apart fromconventional carbon fibres and all other currently knownfibres which are much more susceptible to fracture whensubject to stress beyond the elastic limit It also raisesthe fascinating possibility that composites could beproduced which would snap back into place followingdeformation

As already noted catalytically produced MWNTs aremuch less perfect in structure than arc grown tubes andare therefore considerably less stiff Salvetat et al usedatomic force microscopy to record forcendashdisplacementcurves for nanotubes prepared by the catalytic decom-position of acetylene19 This produced an averagevalue for the Youngrsquos modulus of 27 GPa The tensile

strengths of catalytically produced MWNTs have notyet been measured However it is possible that thesemight be relatively high the defects in these imperfectstructures might be expected to assist stress transferbetween layers

Determining the mechanical properties of SWNTspresents an even greater challenge than for MWNTsHowever some measurements have been made using ananostressing stage20 These produced values for theYoungrsquos modulus in the range 320ndash1470 GPa (mean1002 GPa) and an average breaking strength of 30 GPa

There have also been a substantial number of studiesof the electronic properties of carbon nanotubes121

Early theoretical work showed that nanotubes could bemetallic or semiconducting depending on their structureand diameter22ndash24 Ebbesen et al confirmed experimen-tally that the electronic properties of nanotubes couldvary widely from tube to tube25 both semiconductingand metallic behaviour has been observed Measuredresistivities fall into the approximate range 005 mV mndash10 mV m This compares with a value of approximately04 mV m for high quality single crystal graphite(parallel to the c axis) and 0017 mV m for copper Theresistivities of catalytically produced nanotubes arelikely to be similar to those of disordered carbons ieof the order of 10ndash100 mV m but experimental measure-ments on this type of tube have not yet been carried outThe thermal and optical properties of carbon nanotubes

a low magnification TEM image b HRTEM image of

typical nanotube

4 TEM images of multiwalled carbon nanotubes synthe-

sised by decomposition of C2H2 at 850uC over cobalt

catalyst14

a low magnification SEM image b high magnification

TEM image

5 Images of single-walled carbon nanotubes prepared

by decomposition of C2H2 at 950uC over FeMoAl2O3

catalyst16

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 33

have been studied less intensively than the electronicproperties but some interesting results have been obtainedThe high thermal conductivities of nanotubes have alreadybeen mentioned11 It has also been reported that carbonnanotube suspensions have non-linear optical proper-ties suggesting possible applications as optical limiters26

Another area which has been the subject of muchresearch is the chemical modification of carbon nano-tubes Chemical treatments of nanotubes were first carriedout in an attempt to selectively remove the end-caps anduse them as molecular containers It was demonstratedin 1994 that nanotube caps could be removed by treat-ment with hot nitric acid apparently as a result ofselective oxidation at the pentagonal rings27 In sub-sequent work it was demonstrated by acidndashbase titra-tion and by temperature programmed decompositionthat this acid treatment resulted in the formation ofsurface carboxylic acid and other groups on the tubesurfaces28 These surface acid groups can then be furthermodified in a variety of ways29 Functionalisation hasbeen widely used to facilitate the dispersion of nano-tubes in solvents as a prelude to the preparation ofcomposite materials and can also help to improvebonding between tubes and matrix30

Preparation of carbon nanotubepolymercompositesA commonly used method for preparing nanotubepolymer composites has involved mixing nanotube dis-persions with solutions of the polymer and then eva-porating the solvents in a controlled way As noted inthe previous section the nanotubes are often pretreatedchemically to facilitate solubilisation Shaffer et al forexample have demonstrated that acid treatments enablestable aqueous solutions of catalytically producedMWNTsto be prepared31 These workers then showed that ananotubePVA (polyvinyl alcohol) composite could be

prepared simply by mixing one of these aqueous nano-tube dispersions with an aqueous solution of the poly-mer and then casting the mixtures as films andevaporating the water32 Solution based methods havealso been used to produce nanotubepolystyrene com-posites Thus Hill et al have solubilised both SWNTsand MWNTs by functionalising with a polystyrenecopolymer33 This was achieved by first acid treating thetubes and then carrying out esterification of the surface-bound carboxylic acids The polymer-modified carbonnanotubes were shown to be soluble in common organicsolvents To prepare composites polystyrene was dissolvedin the nanotube solution and nanotubepolystyrene thinfilms prepared using wet casting Functionalisation ofthe nanotubes is not always necessary in order to pre-pare a polymer composite Qian et al used a high energyultrasonic probe to disperse MWNTs in toluene andthen mixed the dispersed suspension with a dilute solu-tion of polystyrene in toluene again with ultrasonicagitation34 The low viscosity of the polymer solutionallowed the nanotubes to move freely through thematrix The mixture was cast on glass and the solventremoved to yield MWNT-doped films Specimens of thecomposite were prepared for TEM by placing drops ofthe mixed solution on copper TEM grids On evapora-tion of the toluene a thin film of the composite was leftsuspended across the grid bars A micrograph of one ofthese films is shown in Fig 7 and this illustrates theexcellent dispersion that was achieved Note that in situpolymerisation has also been used to prepare nanotubepolystyrene composites as is discussed below

The solution mixing approach is limited to polymersthat freely dissolve in common solvents An alternativeis to use thermoplastic polymers (ie polymers thatsoften and melt when heated) and then apply meltprocessing techniques Thus shear mixing can be used toproduce a homogeneous dispersion of nanotubes andextrusion to produce nanotube alignment or to fabricateartefacts in the required form by injection moulding

7 TEM image of MWNTpolystyrene thin film (arrows

indicate regions of polymer shrinkage) inset shows

distribution of nanotube lengths346 High resolution TEM image of deformed nanotube

showing periodic buckling18

Harris Carbon nanotube composites

34 International Materials Reviews 2004 VOL 49 NO 1

Andrews et al used shear mixing to disperse catalytically-produced nanotubes in a range of polymers includinghigh impact polystyrene acrylonitrile butadiene styreneand polypropylene35 The dispersion of nanotubes wasdetermined as a function of mixing energy and tem-perature and the composites were formed into fibres andthin films Haggenmueller et al used a combination ofsolvent casting and melt mixing methods to disperseSWNTs in polymethylmethacrylate (PMMA)36 Compositefibres with a high degree of nanotube orientation wereproduced by melt spinning

Polycarbonate is another thermoplastic polymer thathas been used as a matrix for nanotube composites3738

Sennet et al have used melt processing techniques todisperse and align carbon nanotubes in polycarbonate38

Dispersion was achieved by mixing the catalyticallyproduced MWNTs and SWNTs with polycarbonateresin in a conical twin-screw extruder and alignment wascarried out using a fibre spinning apparatus By opti-mising mixing time and fibre draw rates excellent dis-persion and alignment were accomplished

Nanotubeepoxy composites have also been preparedby a number of groups39ndash43 The earliest work in thisarea was carried out by Ajayan et al39 In 1994 thisgroup described a study in which purified tubes wereembedded into an epoxy resin which was then cut intothin slices with a diamond knife Their initial aim indoing this was not to produce a composite material butto obtain thin TEM specimens which it was hoped wouldcontain cross-sectional images of nanotubes However nocross-sections were observed following sectioning insteadthe nanotubes were found to have become aligned in thedirection of the knife movement as shown in Fig 8According to Ajayan et al the alignment is primarily aconsequence of extensional or shear flow of the matrixproduced by the cutting Of course alignment can bemuch more readily achieved when the epoxy is in a liquidstate Xu et al showed that some alignment of nanotubesin nanotubeepoxy composites can been achieved by usingspin coating methods42 The cure reaction of a SWNT

epoxy resin composite has been analysed using thermalanalysis and Raman spectroscopy43 The SWNTs werefound to have a catalysing effect on the cure reaction

An alternative method for preparing nanotubepoly-mer composites is to use the monomer rather than thepolymer as a starting material and then carry out in situpolymerisation Cochet et al were among the first to usethis method to prepare a MWNTpolyaniline compo-site44 Nanotubes prepared by arc evaporation wereused and were sonicated in a solution of HCl to achievea dispersion The aniline monomer again in HCl wasadded to the suspension and a solution of an oxidantwas slowly added with constant sonication and coolingSonication was continued in an ice bath for 2 h and thecomposite was then obtained by filtering rinsing anddrying In this way it was possible to prepare com-posites with high MWNT loadings (up to 50 wt-)Transport measurements on the composite revealedmajor changes in the electronic behaviour confirmingstrong interaction between nanotubes and polymer asdiscussed below A number of other nanotubepolymercomposites have been prepared using in situ poly-merisation including MWNTpolystyrene45 and SWNTpolyimide46 Several groups have used electrochemicalpolymerisation to grow porous composite films ofMWNTand polypyrrole for use as supercapacitors4748

In the techniques described so far the aim has been toproduce composites in which the nanotubes are dis-tributed evenly throughout the polymer (either ran-domly oriented or aligned) For some applicationshowever a layered arrangement is advantageous Photo-voltaic devices containing layers of nanotubes or nanotubepolymer composites have been prepared as discussed inthe section lsquoElectrical and optical properties of carbonnanotubepolymer compositesrsquo below The preparationof multilayer SWNTpolymer composites with excellentmechanical properties has also been described49 Themethod involved the layer-by-layer deposition of SWNTsand polymer onto a substrate followed by crosslinkingIn this way composites with SWNT loadings as high as

8 Alignment of nanotubes in polymer matrix following cutting with microtome arrows indicate buckled nanotubes39

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 35

50 wt- could be obtained Coating was carried out byalternate dipping of a glass slide or silicon wafer intodispersions of SWNTs (stabilised by acid treatment) andpolymer solutions The layers were held together by vander Waals forces and by electrostatic attraction betweenthe negatively charged SWNTs and a positively chargedpolyelectrolyte such as branched polyethyleneimine (PEI)Coatings containing up to 50 SWNTPEI bilayers couldbe built up in this way When the procedure was com-plete the multilayer films were heated to 120uC topromote crosslinking The films could then be lifted offthe substrate to obtain uniform free-standing mem-branes A rather similar method was used to synthesisemultilayer composites films of the polyelectrolyte poly-diallyldimethylammonium chloride and SWNTs50

Before proceeding to discuss the properties of carbonnanotubepolymer composites in detail it is worth brieflyconsidering the effect that the inclusion of nanotubescan have on polymer microstructure Although thistopic has not yet been widely studied a number ofgroups have demonstrated that the crystallisation andmorphology of the polymers can be strongly affected bysmall additions of nanotubes For example Lozano andBarrera51 showed that the addition of catalytically growncarbon nanofibres can promote nucleation of polypro-pylene Similar observations were made by Valentiniet al52 The crystallinity of a polymers has also beenshown to increase with the addition of carbon nano-fibres5354 Changes in the crystal structure of polymerscan also be induced by nanotubes55 More work on theseeffects would be welcome

Mechanical properties of carbonnanotubepolymer compositesMuch of the work on the preparation of nanotubepolymer composites has been driven by a desire toexploit the tubesrsquo stiffness and strength Even where theinterest has been focused on other properties the abilityof nanotubes to improve the mechanical characteristicsof a polymer has often been a valuable added benefit Inearly work the measured properties were often dis-appointing but such studies were of value in under-standing the reasons for composite failure and inidentifying the critical issues which need to be addressed

Shaffer and Windle were among the first to carry outa systematic study of the mechanical properties ofnanotubepolymer composites Their technique for pre-paring MWNTPVA composites was described in theprevious section32 The tensile elastic moduli of thecomposite films were assessed in a dynamic mechanicalthermal analyser as a function of nanotube loading andtemperature The stiffness of the composites at roomtemperature was relatively low From the theory deve-loped for short-fibre composites a nanotube elasticmodulus of 150 MPa was obtained from the roomtemperature experimental data This value is well belowthe values reported for isolated nanotubes even defec-tive catalytically grown tubes have measured stiffness ofthe order of 30 GPa19 The low value may have more todo with poor stress transfer than with the weakness ofthe nanotubes themselves Above the glass transitiontemperature of the polymer (y85uC) the nanotubeshad a more significant effect on the properties of thecomposite

Cadek et al achieved rather better results in 2002with composites made from PVA and arc producedMWNTs56 The composites were prepared by solutionmixing and films were formed on glass substrates bydrop casting The presence of 1 wt- nanotubes increasedthe Youngrsquos modulus and hardness of the PVA byfactors of 18 and 16 respectively This improvementover the early results of Shaffer and Windle reflects boththe superior quality of the tubes and stronger interfacialbonding Evidence for the latter was provided by TEMstudies

The mechanical properties of nanotubepolystyrenecomposites were studied by Qian and co-workers3435

Two samples of catalytically produced MWNTs wereused one with an average length of y15 mm the otherwith an average length of y50 mm With the addition of1 wt- nanotubes they achieved a 36 and 42 increasein the elastic stiffness of the polymer for short and longtubes respectively (the neat polymer modulus wasy12 GPa) In both cases a 25 increase in the tensilestrength was observed Theoretical calculations carriedout assuming a nanotube modulus of 450 GPa pro-duced increases of a 48 and 62 in the stiffness of thepolymer for short and long tubes respectively The closeagreement within y10 between the experimental andtheoretically predicted composite moduli indicated thatthe external tensile loads were successfully transmittedto the nanotubes across the tubepolymer interface Athigher nanotube concentrations the changes in themechanical properties of the composites were morepronounced35 For MWNTpolystyrene composites con-taining from 25 to 25 vol- nanotubes Youngrsquosmodulus increased progressively from 19 to 45 GPawith the major increases occurring when the MWNTcontent was at or above about 10 vol- However thedependence of tensile strength on nanotube concentra-tion was more complex At the lower concentrations((10 vol-) the tensile strength decreased from theneat polymer value ofy40 MPa only exceeding it whenthe MWNT content was above 15 vol-

In an attempt to understand the tensile fracturemechanisms of the nanotubepolystyrene compositesQian et al performed deformation studies inside aTEM34 Their elegant method of preparing TEM speci-mens has been described above Focusing the electronbeam onto the thin film resulted in local thermal stresseswhich initiated cracks in the composite The propaga-tion speed of the crack could be controlled by varyingthe beam flux onto the sample These in situ TEMobservations showed that the cracks tended to nucleateat low nanotube density areas then propagate alongweak nanotubepolymer interfaces or relatively low nano-tube density regions The nanotubes became alignedperpendicular to the crack direction and bridged thecrack faces in the wake as shown in Fig 9 When thecrack opening displacement exceeded y800 nm thenanotubes began to break andor pull out of the matrixThe fact that some of the pulled-out tubes do not appearto be coated with polymer suggests that there is room forimprovement in nanotubendashpolymer bonding

Watts and Hsu have also used TEM to study thefracture behaviour of a nanotubepolymer composite57

Their composite consisted of arc grown MWNTsembedded in a diblock copolymer known as MPCndashDEA To prepare specimens of the composite for TEM

Harris Carbon nanotube composites

36 International Materials Reviews 2004 VOL 49 NO 1

the tubes were mixed with the polymer in an acidifiedaqueous solution and drops of this were applied to TEMgrids which were then dried in vacuum As in the workof Qian et al cracks were produced in the material byirradiation with an electron beam but a low acceleratingvoltage (75 kV) was used to avoid massive disruption ofthe thin composite film Individual tubes and bundles oftubes were observed to become stretched as the cracksgrew wider and pull-out was often seen In contrast tothe work of Qian et al however the tubes did notbreak This was probably because high quality arcgrown nanotubes were used rather than catalyticallyproduced tubes

The early work on nanotubeepoxy composites byAjayan et al has been described in the previoussection39 In 1998 Ajayan and co-workers described astudy of the mechanical behaviour of MWNTepoxycomposites in both tension and compression58 It wasfound that the compression modulus was higher thanthe tensile modulus indicating that load transfer to thenanotubes in the composite was much higher in com-pression The authors suggested that during load trans-fer to MWNTs only the outer layers are stressed intension whereas all the layers respond in compression

Nanotubeepoxy thin films prepared by Xu et alshowed excellent mechanical properties42 Compared toneat resin thin films for which the elastic modulus was42 GPa a 20 increase in modulus was seen when01 wt- MWNTs were added This was attributed topartial alignment of the MWNTs induced by spincoating Fracture behaviour of the films was investigatedby SEM and showed that pulled-out tubes were oftencovered with polymer suggesting strong interfacial adhe-sion Similar results were reported by Gojny et al59 Thisgroup showed that functionalising arc produced MWNTswith amine groups greatly enhanced bonding with theepoxy matrix As a result the outer shells of the tubesoften remained embedded in the matrix following pull-out as shown in Fig 10

All of the work described so far in this section hasinvolved multiwalled carbon nanotubes Ajayan et al

have described a study of SWNT-containing compo-sites60 Two types of composite were prepared first anSWNTepoxy composite and second a pressed pelletcontaining SWNTs and carbonaceous soot materialformed during nanotube synthesis The SWNTs wereused as-prepared ie generally grouped into bundlesrather than separated into individual tubes The com-posites were loaded axially in tension until failureoccurred and the fracture surfaces were then examinedin detail using SEM In some cases the tube bundleswere found to have been pulled out of the matrix duringthe deformation and fracture of the composites while inother cases the nanotubes were not entirely pulled outbut were stretched between two fracture surfaces Theauthors believed that during pull-out the nanotubeswere sliding axially within the ropes and that the failureobserved at large crack distances was not failure ofindividual tubes but the bundles pulling apart Studiesby micro-Raman spectroscopy supported this hypoth-esis It was found that little shift in the second order A1g

band occurred as a result of the application of axialtension showing that the individual nanotubes were notbeing significantly stretched To take full advantage ofthe high Youngrsquos modulus of SWNTs in polymer com-posites it is clear that load must be transferred effec-tively from the matrix to the nanotubes Ajayan et alsuggested that load transfer could be improved by break-ing bundles down into individual tube fragments beforedispersing in the matrix Alternatively the bundles them-selves could be reinforced by crosslinking the tubes withinbundles (eg by chemical treatments or irradiation)

Single-walled nanotubeepoxy composites have alsobeen prepared by Biercuk et al61 The Vickers hardnessof the polymer was found to increase monotonicallywith addition of SWNTs up to a factor of 35 at 2 wt-nanotube loading Greatly enhanced thermal conducti-vities were also observed Thus samples loaded with1 wt- unpurified SWNT material showed a 70increase in thermal conductivity at 40 K rising to125 at room temperature

The preparation of SWNTpolyelectrolyte compositesby the layer-by-layer method has been described in theprevious section49 The films produced in this way wereshown to have exceptional mechanical properties Thusthe average ultimate tensile strength of the films wasfound to be 220 MPa with some measurements as highas 325 MPa This is several orders of magnitude greater

10 Image of epoxy coated MWNT subject to tensile

stress as result of electron irradiation telescopic

pull-out of outer tube layer has occurred59

9 TEM image of crack in MWNTpolystyrene thin film

induced by thermal stresses34

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 37

than the tensile strength of strong industrial plastics andindicates that the layer-by-layer method has great pro-mise for the production of strong nanotube-containingcomposites

Electrical and optical properties ofcarbon nanotubepolymer compositesThe effect of adding nanotubes on the electrical pro-perties of polymers has been studied by many groupsThese studies fall into a number of distinct categories Insome cases the nanotubes have been used to increase theconductivity of relatively low cost polymers as analternative to currently used fillers such as carbon blackOther studies have involved the incorporation of nano-tubes into conducting polymers such as polyaniline Stillfurther work has been aimed at specific applicationssuch as photovoltaic devices or supercapacitors In manycases improved mechanical properties have been avaluable by-product of the inclusion of nanotubes Theaddition of nanotubes to low cost bulk polymers isconsidered first

Electrostatic dissipationImproving the electrical conductivity of bulk polymers isimportant in a number of applications For example insome aircraft components enhanced conductivity isrequired to provide electrostatic discharge and electro-magnetic radio frequency interference protection Staticelectrical dissipation is also needed in other applicationsincluding computer housings and exterior automotiveparts With these types of application in mind Sandleret al investigated the electrical properties of nanotubeepoxy composites40 Matrix resistivities of about 100 V mwith filler volume fractions as low as 01 vol- wereachieved These figures represented an advance on thebest conductivity values previously obtained with carbonblack in the same epoxy matrix

Lozano and co-workers prepared composites ofcatalytically grown carbon nanofibres in a polypropy-lene matrix5162 The nanofibres were purified and func-tionalised to facilitate bonding with the matrix and highshear processing was used to disperse the fibres in thepolymer matrix Rheological and microscopic analysisshowed that a homogeneous dispersion of nanofibreswas formed and a percolation threshold for electricalconduction of 9ndash18 wt- was observed

These and other published studies suggest that carbonnanotubes have great promise in reducing electrostaticcharging of bulk polymers Nanotubes also have otheradvantages over conventional fillers such as carbonblack and carbon fibres in that they are more amenableto processing and can be more easily dispersed through-out the matrix In addition to the studies available in theopen literature there is undoubtedly a great deal ofcommercial work being carried out on the use of nano-tubes to reduce the electrostatic charging of plastics Infact nanotube containing plastics are already being usedin commercial products63 These include fuel lines inautomobiles where the nanotubes help to dissipate anydangerous charge which may build up Thermoplasticpolymers containing nanotubes are also used in someexterior automobile parts so that they can be earthedduring electrostatic painting

Conducting polymer compositesPolyaniline is a conducting polymer with many attrac-tive characteristics such as processability and environ-mental stability There has been interest in addingnanotubes to this polymer to provide both enhancedconductivity and improved mechanical properties Asmentioned above in the section lsquoPreparation of carbonnanotubepolymer compositesrsquo Cochet et al have pre-pared a MWNTpolyaniline composite by in situ poly-merisation44 The nanotubes were found to have astrong influence on the transport properties of thepolymer Thus the room temperature resistivity of thecomposite was found to be an order of magnitude lowerthan that of pure polyaniline while the low temperatureresistivity was much smaller than that of either poly-aniline or of MWNTs The temperature dependence ofresistivity was also weaker than that of polyanilinealone These observations were explained by assumingthat in situ polymerisation favours charge transfer betweenpolyaniline and MWNTs resulting in an overall materialwhich is more conducting than the starting componentsRaman studies indicated that these charge transferprocesses involved a site-selective interaction betweenthe quinoid ring of the polymer and the MWNTs Othergroups have also demonstrated excellent conductivity inMWNTpolyaniline composites6465 while a group fromDuPont have shown that SWNTpolyaniline compositescan be used as printable conductors for organic elec-tronics devices66

Optoelectronic propertiesSeveral groups have prepared nanotube compositesusing the conjugated polymer poly(p-phenylene viny-lene) (PPV) and its derivatives Polymers of this classhave electroluminescent properties and are widely usedin light-emitting diodes They are also used in photo-voltaic devices In an early study workers from Irelandand the USA described the synthesis of a compositematerial containing arc produced MWNTs in a PPVderivative PmPV67 The composite was prepared usinga solution mixing method and good wetting of the tubesby the polymer was achieved One of the aims of addingnanotubes to the polymer was to increase its electricalconductivity Previous attempts to achieve this by dopinghad resulted in degradation of the optical properties ofthe polymer In the case of nanotubes an increase in theelectrical conductivity of the polymer by up to eightorders of magnitude was seen with no concomitantdegradation of optical properties This was apparentlybecause the tubes acted as nanometric heat sinks pre-venting the build up of large thermal effects whichdamage these conjugated systems It was also shown inthis work that the composite could be used as theemissive layer in an organic light-emitting diode (LED)

As mentioned above several groups have been inter-ested in using nanotubepolymer composites in photo-voltaic devices The photovoltaic effect involves thegeneration of electron and hole pairs and their sub-sequent collection at opposite electrodes In organicmaterials the photon absorption excites an electron intothe conduction band which leads to a generation ofbound electronndashhole pairs or excitons These excitonshave to dissociate into free charges in order to be trans-ported to the electrodes Exciton dissociation can occurat a junction between a polymer and an electron acceptor

Harris Carbon nanotube composites

38 International Materials Reviews 2004 VOL 49 NO 1

molecule This allows the preferential transfer of the elec-trons into the electron acceptor molecule while leavingthe holes to be preferentially transported through thepolymer a process known as photoinduced chargetransfer In 1992 photoinduced charge transfer from aconducting polymer to C60 was discovered68 and thishas led to the fabrication of several efficient photo-voltaic systems using fullerenepolymer combinations69

Photovoltaic devices based on nanotubepolymer com-binations have now been fabricated by several groupsand this work will now be summarised

The earliest attempt to fabricate a nanotubepolymerphotovoltaic device appears to be the 1996 work ofRomero et al70 In this work a layer of MWNTs wasdeposited on Teflon and a PPV derivative was cast ontothis with a thin aluminium layer on top The devicedisplayed clear diode behaviour but only a small photo-current was observed when the junction was irradiatedwith light from a HendashNe laser This may have been dueto the relatively thick PPV active layer Ago et alfabricated a nanotube containing photovoltaic deviceby depositing a layer of PPV onto a glass-supportedMWNT film71 The interaction between MWNTs andPPV was investigated by means of photoluminescence(PL) measurements The photoluminescence efficiencyof the PPV was found to be markedly reduced by thepresence of the MWNT layer Since photoluminescenceis a result of radiative recombination a reduction mayimply improved exciton separation which is what onewishes to achieve in a photovoltaic device The devicefabricated by Ago et al is illustrated in Fig 11a Thisshowed a clear diode characteristic in the dark andupon illumination through the aluminium electrode aphotocurrent was observed A quantum efficiency approxi-mately twice that of the standard indiumndashtin oxidedevice was reported

Kymakis and Amaratunga used a slightly differentapproach to prepare a nanotubeconjugated polymerphotovoltaic device72 In this work composite SWNTpoly(3-octylthiophene) films were drop or spin cast froma solution onto indiumndashtin oxide (ITO) and quartz sub-strates to produce the arrangement shown in Fig 11bDevices were also prepared using the pure P3OT poly-mer A considerable enhancement of the photovoltaic

effect was observed with the P3OTSWNTs blend devicecompared with the pure polymer device

Non-linear optical propertiesAs mentioned above in the section lsquoStructure and pro-perties of carbon nanotubesrsquo it has been demonstratedthat suspensions of both SWNTs and MWNTs havenon-linear optical properties Nanotubepolymer com-posites may therefore have applications in optical devicesTo date however there has been little work on theoptical limiting properties of nanotube containing com-posites In one of the first studies in this area Chen et aldemonstrated third order optical non-linearity in a SWNTpolyimide composite73 Recently OrsquoFlaherty et al pre-pared a composite in which MWNTs were dispersed inthe polymer poly(99-di-n-octylfluorenyl-279-diyl) (PFO)74

Suspensions of the nanotubePFO composite in toluenewere prepared and the optical limiting properties of thesolutions were measured using nanosecond laser pulsesMaximum optical limiting was observed at carbonnanotube loadings in excess of 38 wt- relative to thepolymer

Carbon nanotubeceramic compositesCeramics have high stiffnesses and thermal stabilities butrelatively low breaking strengths Incorporating carbonnanotubes into a ceramic matrix might be expected toproduce a composite with both toughness and high tem-perature stability However achieving a homogeneousdispersion of tubes in an oxide with strong bondingbetween tubes and matrix presents rather more of achallenge than incorporating tubes into a polymer Variousapproaches have been adopted for incorporating nano-tubes into oxides such as alumina and silica and thesewill now be summarised

Peigney and co-workers have been among the pioneersin the synthesis of nanotubeoxide composites75ndash77 Theyhave developed an ingenious technique which involvesthe use of catalysts such as iron-containing a-aluminaPassing a CH4ndashH2 mixture over the catalysts resultedin the in situ growth of nanotubes on iron particlesproducing the precursor for a nanotubeoxide compo-site The resulting powders can then be hot pressed toform the final composite material The mechanicalproperties of these composites were initially disappoint-ing72 Thus the fracture strengths of the nanotubeFendashAl2O3 composites were only marginally higher than thatof Al2O3 and were generally markedly lower than thoseof the carbon-free FendashAl2O3 composites The fracturetoughness values were lower than or similar to that ofAl2O3 However SEM observations of composite frac-tures indicate that the nanotube bundles which are veryflexible could dissipate some fracture energy A similarmethod was used by An and Lim to prepare nanotubeAl2O3 composites in this case using C2H2 as the carbonsource78 Nanotube contents of up to 125 wt- wereachieved and a significant reinforcement effect wasreported

Achieving alignment of nanotubes in a ceramic matrixis much more difficult than in a polymer since ceramicsdo not lend themselves so easily to processing Howevera number of groups have shown that aligned nanotubemetal oxide composites can indeed be produced Peigney

ITO5indiumndashtin oxide PPV5poly(p-phenylene vinylene)

P3OT5poly(3-octylthiophene)

11 a Illustration of photovoltaic device fabricated by Ago

et al71 and b photovoltaic device fabricated by

Kymakis and Amaratunga72

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 39

et al exploited the fact that when their grains are main-tained at sufficiently low sizes (less than about 1 mm)many polycrystalline ceramics become superplasticallowing the use of thermomechanical processing forthe forming of these materials A series of cobalt andiron containing oxide catalysts were prepared77 Theprimary grain size in the powders was smaller than100 nm but in some of them the as produced primarygrains were strongly aggregated and required ball millingin order to reduce aggregate size to less than 1 mmNanotubes were then prepared in situ as describedabove and the powders were then extruded at hightemperatures in a graphite die under vacuum The bestresults were achieved with nanotubes incorporated intoFendashAl2O3 and FeCondashMgAl2O4 matrices In both casesthe nanotubes were not damaged by the processing andalignment of the nanotubes was clearly observed Infact the presence of the nanotubes seemed to facilitatethe superplastic forming by inhibiting grain growth andacting as lubricating agents It was demonstrated thatthe aligned nanotubemetal oxide composites showed ananisotropy of electrical conductivity which could beadjusted by controlling the nanotube content

One of the first studies of a SWNTceramic compositewas described by Zhan et al79 This group used thetechnique of spark plasma sintering (SPS) to prepareAl2O3 composites containing 10 vol- SWNTs Theadvantage of spark plasma sintering is that it allowsceramic powders to be annealed at lower temperaturesand for much shorter times than other sintering pro-cesses leading to the fabrication of fully dense ceramicsor composites with nanocrystalline microstructures undermild conditions Unlike some of the other processes usedto prepare nanotubeoxide composites spark plasmasintering does not damage the nanotubes and the nano-tube bundles could be seen to be located mainly at theboundaries of the Al2O3 grains in good contact with theAl2O3 matrix Before sintering the powders had beenmixed by ball milling which produced a reasonablyhomogeneous dispersion again without damaging thenanotubes The fracture toughnesses of the final com-posites were remarkable ndash about twice that of the fullydensified but unreinforced Al2O3 matrix In contrast toprevious work on nanotubeceramic composites thefracture toughness was found to increase with nanotubedensity This is thought to be due to the formation ofentangled networks of single-walled carbon nanotubeswhich may inhibit crack propagation A subsequentstudy of the nanotubeAl2O3 composites showed thatthey had high electrical conductivities80

Various techniques have been used to prepare carbonnanotubesilica composites Seeger et al described amethod which involved preparing a composite gel ofMWNTs with tetraethoxysilane (TEOS) and then sinter-ing this at 1150uC in argon81 A drawback of this tech-nique was that the sintering process led to a partialcrystallisation of the SiO2 resulting in a very inhomo-geneous matrix An alternative method which avoidsthese problems has been developed82 This involvesusing a Nd YAG laser to rapidly heat a TEOSnano-tube mixture resulting in partial melting of the matrixThis produced an amorphous silica matrix with nocrystallisation

A novel nanotubesilica composite material withpotentially useful optical properties has recently been

described by Han et al83 These workers synthesisedfilms consisting of silica spheres with diameters rangingfrom 200 to 650 nm Molybdenumcobalt catalyst parti-cles were then deposited onto the silica spheres and usedto catalyse the growth of SWNTs Since SWNTs arehighly non-linear and fast-switching materials they canbe incorporated into an optically confining environmentto achieve these characteristics at relatively low levels oflaser intensities

There have been a few studies on preparing nanotubecomposites using oxides other than alumina and silicaVincent et al reported the production of nanotubeTiO2

composites using solndashgel methods84 Possible applica-tions of the composites included optical non-linear waveguides and unidimensional conductive films Solndashgelmethods were also used by Sakamoto and Dunn toprepare SWNTV2O5 composites for use as electrodesin secondary lithium batteries85 Little work has beencarried out so far on incorporating nanotubes into non-oxide ceramics In one of the few published studies Maet al fabricated nanotubesilicon carbide composites bymixing SiC nanoparticles with 10 wt- MWNTs andhot pressing at 2000uC86 They found that the nanotubesplayed a strengthening and toughening role in thecomposite both bend strength and fracture toughnessincreased by about 10 compared to monolithic SiCThere have also been some initial studies on carbonnanotubecarbon fibre composites8788

Carbon nanotubemetal compositesComposite materials containing conventional carbonfibres in a metal matrix such as aluminium or mag-nesium are used in a number of specialist applica-tions8990 Such composites combine low density withhigh strength and modulus making them particularlyattractive to the aerospace industry There is growinginterest in the addition of carbon nanotubes to metalmatrices although only a small number of studies havebeen published to date Kuzumaki et al described thepreparation of a nanotubealuminium composite in 199891

Their method involved mixing a nanotube sample with afine aluminium powder mounting the mixture in a6 mm silver sheath and then drawing and heating thewire at 700uC in a vacuum furnace The result was acomposite wire in which the nanotubes were partiallyaligned along the axial direction The tensile strengths ofthe as prepared composite wires were comparable to thatof pure aluminium but the composite wires retained thisstrength after prolonged annealing at 600uC while thestrength of pure aluminium decreased by about 50after this treatment The same group subsequently pre-pared nanotubetitanium composites which showed alarge increase in hardness and modulus compared topure titanium92

Chen et al prepared nanotubenickelphosphorus com-posite coatings on stainless steel substrates using anelectroless plating method93 The coatings exhibited higherwear resistance and lower friction coefficients than SiCNiP and graphiteNiP composite coatings The samegroup prepared nanotubecopper composites by a powdermetallurgy technique and showed that the presence ofthe nanotubes again resulted in improved wear rates andfriction coefficients These results suggest that nanotube

Harris Carbon nanotube composites

40 International Materials Reviews 2004 VOL 49 NO 1

metal composites may have potential in tribologicalapplications

DiscussionThe subject of carbon nanotube composites has grownrapidly over the past 5 years or so into a vibrant andexciting area of research It is important to recognisehowever that this is still a new subject and the workcarried out so far may have hardly scratched the surfaceof what is possible For example the great majority ofthe studies that have been published to date haveinvolved the use of polymer matrices there is great scopefor more work on incorporating nanotubes into cera-mics metals and other forms of carbon Also most ofthe work up to now has been aimed at exploiting themechanical properties of carbon nanotubes rather thantheir electronic or optical properties although interest inthose is growing

When considering the mechanical performance ofcarbon nanotube composites it is clear that these dependstrongly on whether arc grown tubes or catalyticallyproduced tubes have been used In considering thecharacteristics of the two types of tube it is instructive tocompare the work of Qian et al34 with that of Watts andHsu57 Both incorporated the nanotubes into polymersand then used TEM to characterise the composites andstudy their fracture behaviour Qian et al used cataly-tically produced tubes and these were found to bewell dispersed throughout the polymer matrix with noclustering together of tubes However during compositefracture the tubes were often observed to break sug-gesting that they have relatively poor strength Wattsand Hsu on the other hand employed much moreperfect nanotubes produced by arc evaporation andTEM studies revealed superior mechanical propertieswith no observed breaking However the nanotubestended to cluster together so that during crack growththe tubes could quite easily slide past one another Thuswhile it is clear that the best mechanical performancewill be achieved with arc grown nanotubes it is impor-tant that the tubes be well dispersed throughout thematrix in order fully to exploit their properties

Equally important is achieving good bonding betweennanotubes and matrix Many studies have shown that thisinterfacial bonding can be enhanced by functionalising thenanotubes before embedding them in the matrix Theeffect of this can be seen in Fig 10 from the work ofGojny et al59 where bonding between MWNTs and anepoxy matrix is so strong that the outer shells of the tuberemain embedded in the matrix following pull-out Apossible disadvantage of functionalisation is that it couldweaken the nanotubes by disrupting the perfect cylindricalstructure This might be expected to be a particularproblem with single-walled tubes However functionali-sation of the nanotubes may not be necessary to producebonding with a polymer matrix Liao and Li havemodelled the interfacial characteristics of a nanotubepolystyrene composite assuming no covalent bondingbetween the tubes and the matrix94 It was found thatelectrostatic and van der Waals interactions contributedsignificantly to the interfacial stress transfer A nanotubepull-out simulation suggested that the interfacial shearstress of the nanotubepolystyrene system is about160 MPa significantly higher than for most carbon fibrereinforced polymer composite systems It is also worth

mentioning recent work by Star and co-workers9596

in which it was demonstrated that poly(5-alkoxy-m-phenylenevinylene)-co-[(25-dioctyloxy-p-phenylene)vinylene] (PAmPV) derivatives could be induced towrap around SWNTs forming strong non-covalentbonds

The quality of the nanotubes and the nature of theinterfacial bonding also has a strong influence on theelectrical properties of nanotubepolymer compositesArc grown tubes can have resistivities lower than that ofsingle crystal graphite but as already discussed theirelectronic properties depend on their precise structureAt present we do not have a reliable method of separat-ing metallic and semiconducting tubes so it is notpossible to exploit their full potential Catalytically pro-duced tubes are likely to be poorer conductors thanmetallic arc-grown tubes owing to their relatively dis-ordered structure Nevertheless catalytically grown tubesare sufficiently conducting to be useful in some com-mercial applications such as electrically conductive paintsThe influence of interfacial bonding on electronic pro-perties was demonstrated in the work of Cochet et al44

These workers prepared a MWNTpolyaniline com-posite by in situ polymerisation and found it to haveexcellent electrical transport properties This was attrib-uted to strong interaction between tubes and polymer

As has been shown in this review the possible appli-cations of carbon nanotube composites range widelyfrom electrically conductive paints to optical devicesOnly a relatively small number of these new materialshave been used commercially however and there arestill many problems to overcome before the full potentialof nanotube containing composites can be realised Notleast of these problems is the high cost of good qualitycarbon nanotubes Currently the commercial price ofunpurified arc-grownMWNTs is approximately $10000gwhile purified MWNTs cost around $30000g Cataly-tically grown MWNTs are much cheaper but as empha-sised throughout this review have generally inferiorproperties As far as SWNTs are concerned unpurifiedsamples cost around $5000g while purified SWNTs are$50000g As a comparison the price of gold is cur-rently about $1200g The development of new techni-ques for the low cost synthesis of high quality nanotubesis clearly essential This might be achieved throughimproved catalytic processes and there is currently alarge amount of work being carried out towards thisend Alternatively it may be possible to develop newnon-catalytic techniques for the synthesis of fullerene-related nanotubes In this connection it is interesting tonote that nanotubes can be produced by the high tem-perature heat treatment of conventional carbons such asmicroporous non-graphitising carbons9798 or carbonblack99100 Further work in this area would be of value

Acknowledgement

I thank Milo Shaffer for helpful comments

References1 P J F Harris lsquoCarbon nanotubes and related structures ndash new

materials for the twenty-first centuryrsquo 1999 Cambridge

Cambridge University Press

2 S Iijima Nature 1991 354 56ndash58

3 W Kratschmer L D Lamb K Fostiropoulos and D R

Huffman Nature 1990 347 354ndash358

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 41

4 S Iijima and T Ichihashi Nature 1993 363 603ndash605

5 D S Bethune C H Kiang M S De Vries G Gorman R Savoy

J Vasquez and R Beyers Nature 1993 363 605ndash607

6 H Dai A G Rinzler P Nikolaev A Thess D T Colbert and

R E Smalley Chem Phys Lett 1996 260 471ndash475

7 M M J Treacy T W Ebbesen and J M Gibson Nature 1996

381 678ndash680

8 E W Wong P E Sheehan and C M Lieber Science 1997 277

1971ndash1975

9 M R Falvo G J Clary R M Taylor V Chi F P Brooks

S Washburn and R Superfine Nature 1997 389 582ndash584

10 K Anazawa K Shimotani C Manabe H Watanabe and M

Shimizu Appl Phys Lett 2002 81 739ndash741

11 P Kim L Shi A Majumdar and P L McEuen Phys Rev Lett

2001 87 paper 215502

12 E T Thostenson Z F Ren and T W Chou Compos Sci

Technol 2001 61 1899ndash1912

13 K T Lau and D Hui Composites Part B Engineering 2002 33

263ndash277

14 Y Huh J Y Lee J Cheon Y K Hong J Y Koo T J Lee and

C J Lee J Mater Chem 2003 13 2297ndash2300

15 M Endo T Koyama and Y Hishiyama Jpn J Appl Phys 1976

15 2073ndash2076

16 S C Lyu B C Liu T J Lee Z Y Liu C W Yang C Y Park

and C J Lee Chem Commun 2003 734ndash735

17 M-F Yu O Lourie M J Dyer K Moloni T F Kelly and R S

Ruoff Science 2000 287 637ndash640

18 P Poncharal Z L Wang D Ugarte and W A De Heer Science

1999 283 1513ndash1516

19 J-P Salvetat A J Kulik J-M Bonard G A D Briggs

T Stockli K Metenier S Bonnamy F Beguin N A Burnham

and L Forro Adv Mater 1999 11 161ndash165

20 M-F Yu B S Files S Arepalli and R S Ruoff Phys Rev Lett

2000 84 5552ndash5555

21 M Terrones Int Mater Rev to be published

22 J W Mintmire B I Dunlap and C T White Phys Rev Lett

1992 68 631ndash634

23 R Saito M Fujita G Dresselhaus and M S Dresselhaus Phys

Rev B 1992 46 1804ndash1811

24 N Hamada S Sawada and A Oshiyama Phys Rev Lett 1992

68 1579ndash1581

25 T W Ebbesen H J Lezec H Hiura J W Bennett H F Ghaemi

and T Thio Nature 1996 382 54ndash56

26 J E Riggs D B Walker D L Carroll and Y-P Sun J Phys

Chem B 2000 104 7071ndash7076

27 S C Tsang Y K Chen P J F Harris and M L H Green

Nature 1994 372 159ndash162

28 R M Lago S C Tsang K L Lu Y K Chen and M L H

Green Chem Commun 1995 1355ndash1356

29 Y-P Sun K F Fu Y Lin and W Huang Acc Chem Res 2002

35 1096ndash1104

30 S B Sinnott J Nanosci Nanotechnol 2002 2113ndash123

31 M S P Shaffer X Fan and A H Windle Carbon 1998 36

1603ndash1612

32 M S P Shaffer and A H Windle Adv Mater 1999 11 937ndash941

33 D E Hill Y Lin A M Rao L F Allard and Y-P Sun

Macromolecules 2002 35 9466ndash9471

34 D Qian E C Dickey R Andrews and T Rantell Appl Phys

Lett 2000 76 2868ndash2870

35 R Andrews D Jacques D Qian and T Rantell Acc Chem Res

2002 35 1008ndash1017

36 R Haggenmueller H H Gommans A G Rinzler J E Fischer

and K I Winey Chem Phys Lett 2000 330 219ndash225

37 P Potschke T D Fornes and D R Paul Polymer 2002 43 3247ndash

3255

38 M Sennett E Welsh J B Wright W Z Li J G Wen and Z F

Ren Appl Phys A 2003 76 111ndash113

39 P M Ajayan O Stephan C Colliex and D Trauth Science 1994

265 1212ndash1214

40 J Sandler M S P Shaffer T Prasse W Bauhofer K Schulte

and H Windle Polymer 1999 40 5967ndash5971

41 A Allaoui S Bai H M Cheng and J B Bai Compos Sci

Technol 2002 62 1993ndash1998

42 X Xu M M Thwe C Shearwood and K Liao Appl Phys Lett

2002 81 2833ndash2835

43 D Puglia L Valentini and J M Kenny J Appl Polymer Sci

2003 88 452ndash458

44 M Cochet W K Maser A M Benito M A Callejas M T

Martınez J-M Benoit J Schreiber and O Chauvet Chem

Commun 2001 1450ndash1451

45 M S P Shaffer and K Koziol Chem Commun 2002 2074ndash

2075

46 C Park Z Ounaies K A Watson R E Crooks J Smith S E

Lowther J W Connell E J Siochi J S Harrison and T L St

Clair Chem Phys Lett 2002 364 303ndash308

47 K Jurewicz S Delpeux V Bertagna F Beguin and E

Frackowiak Chem Phys Lett 2001 347 36ndash40

48 M Hughes G Z Chen M S P Shaffer D J Fray and A H

Windle Chem Mater 2002 14 1610ndash1613

49 A A Mamedov N A Kotov M Prato D M Guldi J P

Wicksted and A Hirsch Nature Mater 2002 1 190ndash194

50 J H Rouse and P T Lillehei Nano Lett 2003 3 59ndash62

51 K Lozano and E V Barrera J Appl Polym Sci 2001 79 125ndash

133

52 L Valentini J Biagiotti J M Kenny and S Santucci Compos

Sci Technol 2003 63 1149ndash1153

53 J Sandler P Werner M S P Shaffer V Demchuk V Altstadt

and A H Windle Composites Part A Appl Sci Manuf 2002 33

1033ndash1039

54 R Czerw Z X Guo P M Ajayan Y P Sun and D L Carroll

Nano Lett 2001 1 423ndash427

55 B P Grady F Pompeo R L Shambaugh and D E Resasco

J Phys Chem B 2002 106 5852ndash5858

56 M Cadek J N Coleman V Barron K Hedicke and W J Blau

Appl Phys Lett 2002 81 5123ndash5125

57 P C P Watts and W K Hsu Nanotechnology 2003 14 L7ndashL10

58 L S Schadler S C Giannaris and P M Ajayan Appl Phys Lett

1998 73 3842ndash3844

59 F H Gojny J Nastalczyk Z Roslaniec and K Schulte Chem

Phys Lett 2003 370 820ndash824

60 P M Ajayan L S Schadler C Giannaris and A Rubio Adv

Mater 2000 12 750ndash753

61 M J Biercuk M C Llaguno M Radosavljevic J K Hyun A T

Johnson and J E Fischer Appl Phys Lett 2002 80 2767ndash2769

62 K Lozano J Bonilla-Rios and E V Barrera J Appl Polym Sci

2001 80 1162ndash1172

63 V Jamieson New Sci 15 Mar 2003 30

64 H Zengin W S Zhou J Y Jin W R Czerw D W Smith

L Echegoyen D L Carroll S H Foulger and J Ballato Adv

Mater 2002 14 1480ndash1483

65 J Deng X Ding W Zhang Y Peng J Wang X Long P Li and

A S C Chan Eur Polym J 2002 38 2497ndash2501

66 G B Blanchet C R Fincher and F Gao Appl Phys Lett 2003

82 1290ndash1292

67 S A Curran P M Ajayan W J Blau D L Carroll J N

Coleman A B Dalton A P Davey A Drury B McCarthy

S Maier and A A Strevens Adv Mater 1998 10 1091ndash1093

68 N S Sariciftci L Smilowitz A J Heeger and F Wudl Science

1992 258 1474ndash1476

69 N Camaioni L Garlaschelli A Geri M Maggini G Possamai

and G Ridolfi J Mater Chem 2002 12 2065ndash2070

70 D B Romero M Carrard W De Heer and L Zuppiroli

Adv Mater 1996 8 899ndash902

71 H Ago K Petritsch M S P Shaffer A H Windle and R H

Friend Adv Mater 1999 11 1281ndash1285

72 E Kymakis and G A J Amaratunga Appl Phys Lett 2002 80

112ndash114

73 Y-C Chen N R Raravikar L S Schadler P M Ajayan Y-P

Zhao T-M Lu G-C Wang and X-C Zhang Appl Phys Lett

2002 81 975ndash977

74 S A OrsquoFlaherty R Murphy S V Hold M Cadek J N

Coleman and W J Blau J Phys Chem B 2003 107 958ndash964

75 A Peigney C Laurent F Dobigeon and A Rousset J Mater

Res 1997 12 613ndash615

76 E Flahaut A Peigney C Laurent C Marliere F Chastel and

A Rousset Acta Mater 2000 48 3803ndash3812

77 A Peigney E Flahaut C Laurent F Chastel and A Rousset

Chem Phys Lett 2002 352 20ndash25

78 J W An and D S Lim J Ceram Process Res 2002 3 201ndash204

79 G-D Zhan J D Kuntz J L Wan and A K Mukherjee Nature

Mater 2003 2 38ndash42

80 G-D Zhan J D Kuntz J E Garay and A K Mukherjee Appl

Phys Lett 2003 83 1228ndash1230

81 T Seeger T Kohler T Frauenheim N Grobert M Ruhle

M Terrones and G Seifert Chem Commun 2002 34ndash35

82 T Seeger G De La Fuente W K Maser A M Benito M A

Callejas and M T Martınez Nanotechnology 2003 14 184ndash187

83 H Han S Vijayalakshmi A Lan Z Iqbal H Grebel E Lalanne

and A M Johnson Appl Phys Lett 2003 82 1458ndash1460

Harris Carbon nanotube composites

42 International Materials Reviews 2004 VOL 49 NO 1

84 P Vincent A Brioude C Journet S Rabaste S T Purcell J Le

Brusq and J C Plenet J Non-Cryst Sol 2002 311 130ndash137

85 J S Sakamoto and B Dunn J Electrochem Soc 2002 149 A26ndashA30

86 R Z Ma J Wu B Q Wei J Liang and D H Wu J Mater Sci

1998 33 5243ndash5246

87 R Andrews D Jacques A M Rao T Rantell F Derbyshire

Y Chen J Chen and R C Haddon Appl Phys Lett 1999 75

1329ndash1331

88 E T Thostenson W Z Li D Z Wang Z F Ren and T W

Chou J Appl Phys 2002 91 6034ndash6037

89 D D L Chung lsquoCarbon fiber compositesrsquo chap 7 1994 Boston

Butterworth-Heinemann

90 S Rawal JOM 2001 53 14ndash17

91 T Kuzumaki K Miyazawa H Ichinose and K Ito J Mater

Res 1998 13 2445ndash2449

92 T Kuzumaki O Ujiie H Ichinose and K Ito Adv Eng Mater

2000 2 416ndash418

93 W X Chen J P Tu L Y Wang H Y Gan Z D Xu and X B

Zhang Carbon 2003 41 215ndash222

94 K Liao and S Li Appl Phys Lett 2001 79 4225ndash4227

95 A Star J F Stoddart D Steuerman M Diehl A Boukai E W

Wong X Yang S W Chung H Choi and J R Heath

Angewandte Chemie-Int Ed 2001 40 1721ndash1725

96 A Star Y Liu K Grant L Ridvan J F Stoddart D W

Steuerman M R Diehl A Boukai and J R Heath

Macromolecules 2003 36 553ndash560

97 P J F Harris S C Tsang J B Claridge and M L H Green

J Chem Soc Faraday Trans 1994 90 2799ndash2802

98 A A Setlur S P Doherty J Y Dai and R P H Chang Appl

Phys Lett 2000 76 3008ndash3010

99 S P Doherty D B Buchholz B J Li and R P H Chang

J Mater Res 2003 18 941ndash949

100 D B Buchholz S P Doherty and R P H Chang Carbon 2003

41 1625ndash1634

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 43

Page 4: Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P. J. F. Harris * Carbon nanotubes are molecular-scale tubes of graphitic carbon with

have been studied less intensively than the electronicproperties but some interesting results have been obtainedThe high thermal conductivities of nanotubes have alreadybeen mentioned11 It has also been reported that carbonnanotube suspensions have non-linear optical proper-ties suggesting possible applications as optical limiters26

Another area which has been the subject of muchresearch is the chemical modification of carbon nano-tubes Chemical treatments of nanotubes were first carriedout in an attempt to selectively remove the end-caps anduse them as molecular containers It was demonstratedin 1994 that nanotube caps could be removed by treat-ment with hot nitric acid apparently as a result ofselective oxidation at the pentagonal rings27 In sub-sequent work it was demonstrated by acidndashbase titra-tion and by temperature programmed decompositionthat this acid treatment resulted in the formation ofsurface carboxylic acid and other groups on the tubesurfaces28 These surface acid groups can then be furthermodified in a variety of ways29 Functionalisation hasbeen widely used to facilitate the dispersion of nano-tubes in solvents as a prelude to the preparation ofcomposite materials and can also help to improvebonding between tubes and matrix30

Preparation of carbon nanotubepolymercompositesA commonly used method for preparing nanotubepolymer composites has involved mixing nanotube dis-persions with solutions of the polymer and then eva-porating the solvents in a controlled way As noted inthe previous section the nanotubes are often pretreatedchemically to facilitate solubilisation Shaffer et al forexample have demonstrated that acid treatments enablestable aqueous solutions of catalytically producedMWNTsto be prepared31 These workers then showed that ananotubePVA (polyvinyl alcohol) composite could be

prepared simply by mixing one of these aqueous nano-tube dispersions with an aqueous solution of the poly-mer and then casting the mixtures as films andevaporating the water32 Solution based methods havealso been used to produce nanotubepolystyrene com-posites Thus Hill et al have solubilised both SWNTsand MWNTs by functionalising with a polystyrenecopolymer33 This was achieved by first acid treating thetubes and then carrying out esterification of the surface-bound carboxylic acids The polymer-modified carbonnanotubes were shown to be soluble in common organicsolvents To prepare composites polystyrene was dissolvedin the nanotube solution and nanotubepolystyrene thinfilms prepared using wet casting Functionalisation ofthe nanotubes is not always necessary in order to pre-pare a polymer composite Qian et al used a high energyultrasonic probe to disperse MWNTs in toluene andthen mixed the dispersed suspension with a dilute solu-tion of polystyrene in toluene again with ultrasonicagitation34 The low viscosity of the polymer solutionallowed the nanotubes to move freely through thematrix The mixture was cast on glass and the solventremoved to yield MWNT-doped films Specimens of thecomposite were prepared for TEM by placing drops ofthe mixed solution on copper TEM grids On evapora-tion of the toluene a thin film of the composite was leftsuspended across the grid bars A micrograph of one ofthese films is shown in Fig 7 and this illustrates theexcellent dispersion that was achieved Note that in situpolymerisation has also been used to prepare nanotubepolystyrene composites as is discussed below

The solution mixing approach is limited to polymersthat freely dissolve in common solvents An alternativeis to use thermoplastic polymers (ie polymers thatsoften and melt when heated) and then apply meltprocessing techniques Thus shear mixing can be used toproduce a homogeneous dispersion of nanotubes andextrusion to produce nanotube alignment or to fabricateartefacts in the required form by injection moulding

7 TEM image of MWNTpolystyrene thin film (arrows

indicate regions of polymer shrinkage) inset shows

distribution of nanotube lengths346 High resolution TEM image of deformed nanotube

showing periodic buckling18

Harris Carbon nanotube composites

34 International Materials Reviews 2004 VOL 49 NO 1

Andrews et al used shear mixing to disperse catalytically-produced nanotubes in a range of polymers includinghigh impact polystyrene acrylonitrile butadiene styreneand polypropylene35 The dispersion of nanotubes wasdetermined as a function of mixing energy and tem-perature and the composites were formed into fibres andthin films Haggenmueller et al used a combination ofsolvent casting and melt mixing methods to disperseSWNTs in polymethylmethacrylate (PMMA)36 Compositefibres with a high degree of nanotube orientation wereproduced by melt spinning

Polycarbonate is another thermoplastic polymer thathas been used as a matrix for nanotube composites3738

Sennet et al have used melt processing techniques todisperse and align carbon nanotubes in polycarbonate38

Dispersion was achieved by mixing the catalyticallyproduced MWNTs and SWNTs with polycarbonateresin in a conical twin-screw extruder and alignment wascarried out using a fibre spinning apparatus By opti-mising mixing time and fibre draw rates excellent dis-persion and alignment were accomplished

Nanotubeepoxy composites have also been preparedby a number of groups39ndash43 The earliest work in thisarea was carried out by Ajayan et al39 In 1994 thisgroup described a study in which purified tubes wereembedded into an epoxy resin which was then cut intothin slices with a diamond knife Their initial aim indoing this was not to produce a composite material butto obtain thin TEM specimens which it was hoped wouldcontain cross-sectional images of nanotubes However nocross-sections were observed following sectioning insteadthe nanotubes were found to have become aligned in thedirection of the knife movement as shown in Fig 8According to Ajayan et al the alignment is primarily aconsequence of extensional or shear flow of the matrixproduced by the cutting Of course alignment can bemuch more readily achieved when the epoxy is in a liquidstate Xu et al showed that some alignment of nanotubesin nanotubeepoxy composites can been achieved by usingspin coating methods42 The cure reaction of a SWNT

epoxy resin composite has been analysed using thermalanalysis and Raman spectroscopy43 The SWNTs werefound to have a catalysing effect on the cure reaction

An alternative method for preparing nanotubepoly-mer composites is to use the monomer rather than thepolymer as a starting material and then carry out in situpolymerisation Cochet et al were among the first to usethis method to prepare a MWNTpolyaniline compo-site44 Nanotubes prepared by arc evaporation wereused and were sonicated in a solution of HCl to achievea dispersion The aniline monomer again in HCl wasadded to the suspension and a solution of an oxidantwas slowly added with constant sonication and coolingSonication was continued in an ice bath for 2 h and thecomposite was then obtained by filtering rinsing anddrying In this way it was possible to prepare com-posites with high MWNT loadings (up to 50 wt-)Transport measurements on the composite revealedmajor changes in the electronic behaviour confirmingstrong interaction between nanotubes and polymer asdiscussed below A number of other nanotubepolymercomposites have been prepared using in situ poly-merisation including MWNTpolystyrene45 and SWNTpolyimide46 Several groups have used electrochemicalpolymerisation to grow porous composite films ofMWNTand polypyrrole for use as supercapacitors4748

In the techniques described so far the aim has been toproduce composites in which the nanotubes are dis-tributed evenly throughout the polymer (either ran-domly oriented or aligned) For some applicationshowever a layered arrangement is advantageous Photo-voltaic devices containing layers of nanotubes or nanotubepolymer composites have been prepared as discussed inthe section lsquoElectrical and optical properties of carbonnanotubepolymer compositesrsquo below The preparationof multilayer SWNTpolymer composites with excellentmechanical properties has also been described49 Themethod involved the layer-by-layer deposition of SWNTsand polymer onto a substrate followed by crosslinkingIn this way composites with SWNT loadings as high as

8 Alignment of nanotubes in polymer matrix following cutting with microtome arrows indicate buckled nanotubes39

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 35

50 wt- could be obtained Coating was carried out byalternate dipping of a glass slide or silicon wafer intodispersions of SWNTs (stabilised by acid treatment) andpolymer solutions The layers were held together by vander Waals forces and by electrostatic attraction betweenthe negatively charged SWNTs and a positively chargedpolyelectrolyte such as branched polyethyleneimine (PEI)Coatings containing up to 50 SWNTPEI bilayers couldbe built up in this way When the procedure was com-plete the multilayer films were heated to 120uC topromote crosslinking The films could then be lifted offthe substrate to obtain uniform free-standing mem-branes A rather similar method was used to synthesisemultilayer composites films of the polyelectrolyte poly-diallyldimethylammonium chloride and SWNTs50

Before proceeding to discuss the properties of carbonnanotubepolymer composites in detail it is worth brieflyconsidering the effect that the inclusion of nanotubescan have on polymer microstructure Although thistopic has not yet been widely studied a number ofgroups have demonstrated that the crystallisation andmorphology of the polymers can be strongly affected bysmall additions of nanotubes For example Lozano andBarrera51 showed that the addition of catalytically growncarbon nanofibres can promote nucleation of polypro-pylene Similar observations were made by Valentiniet al52 The crystallinity of a polymers has also beenshown to increase with the addition of carbon nano-fibres5354 Changes in the crystal structure of polymerscan also be induced by nanotubes55 More work on theseeffects would be welcome

Mechanical properties of carbonnanotubepolymer compositesMuch of the work on the preparation of nanotubepolymer composites has been driven by a desire toexploit the tubesrsquo stiffness and strength Even where theinterest has been focused on other properties the abilityof nanotubes to improve the mechanical characteristicsof a polymer has often been a valuable added benefit Inearly work the measured properties were often dis-appointing but such studies were of value in under-standing the reasons for composite failure and inidentifying the critical issues which need to be addressed

Shaffer and Windle were among the first to carry outa systematic study of the mechanical properties ofnanotubepolymer composites Their technique for pre-paring MWNTPVA composites was described in theprevious section32 The tensile elastic moduli of thecomposite films were assessed in a dynamic mechanicalthermal analyser as a function of nanotube loading andtemperature The stiffness of the composites at roomtemperature was relatively low From the theory deve-loped for short-fibre composites a nanotube elasticmodulus of 150 MPa was obtained from the roomtemperature experimental data This value is well belowthe values reported for isolated nanotubes even defec-tive catalytically grown tubes have measured stiffness ofthe order of 30 GPa19 The low value may have more todo with poor stress transfer than with the weakness ofthe nanotubes themselves Above the glass transitiontemperature of the polymer (y85uC) the nanotubeshad a more significant effect on the properties of thecomposite

Cadek et al achieved rather better results in 2002with composites made from PVA and arc producedMWNTs56 The composites were prepared by solutionmixing and films were formed on glass substrates bydrop casting The presence of 1 wt- nanotubes increasedthe Youngrsquos modulus and hardness of the PVA byfactors of 18 and 16 respectively This improvementover the early results of Shaffer and Windle reflects boththe superior quality of the tubes and stronger interfacialbonding Evidence for the latter was provided by TEMstudies

The mechanical properties of nanotubepolystyrenecomposites were studied by Qian and co-workers3435

Two samples of catalytically produced MWNTs wereused one with an average length of y15 mm the otherwith an average length of y50 mm With the addition of1 wt- nanotubes they achieved a 36 and 42 increasein the elastic stiffness of the polymer for short and longtubes respectively (the neat polymer modulus wasy12 GPa) In both cases a 25 increase in the tensilestrength was observed Theoretical calculations carriedout assuming a nanotube modulus of 450 GPa pro-duced increases of a 48 and 62 in the stiffness of thepolymer for short and long tubes respectively The closeagreement within y10 between the experimental andtheoretically predicted composite moduli indicated thatthe external tensile loads were successfully transmittedto the nanotubes across the tubepolymer interface Athigher nanotube concentrations the changes in themechanical properties of the composites were morepronounced35 For MWNTpolystyrene composites con-taining from 25 to 25 vol- nanotubes Youngrsquosmodulus increased progressively from 19 to 45 GPawith the major increases occurring when the MWNTcontent was at or above about 10 vol- However thedependence of tensile strength on nanotube concentra-tion was more complex At the lower concentrations((10 vol-) the tensile strength decreased from theneat polymer value ofy40 MPa only exceeding it whenthe MWNT content was above 15 vol-

In an attempt to understand the tensile fracturemechanisms of the nanotubepolystyrene compositesQian et al performed deformation studies inside aTEM34 Their elegant method of preparing TEM speci-mens has been described above Focusing the electronbeam onto the thin film resulted in local thermal stresseswhich initiated cracks in the composite The propaga-tion speed of the crack could be controlled by varyingthe beam flux onto the sample These in situ TEMobservations showed that the cracks tended to nucleateat low nanotube density areas then propagate alongweak nanotubepolymer interfaces or relatively low nano-tube density regions The nanotubes became alignedperpendicular to the crack direction and bridged thecrack faces in the wake as shown in Fig 9 When thecrack opening displacement exceeded y800 nm thenanotubes began to break andor pull out of the matrixThe fact that some of the pulled-out tubes do not appearto be coated with polymer suggests that there is room forimprovement in nanotubendashpolymer bonding

Watts and Hsu have also used TEM to study thefracture behaviour of a nanotubepolymer composite57

Their composite consisted of arc grown MWNTsembedded in a diblock copolymer known as MPCndashDEA To prepare specimens of the composite for TEM

Harris Carbon nanotube composites

36 International Materials Reviews 2004 VOL 49 NO 1

the tubes were mixed with the polymer in an acidifiedaqueous solution and drops of this were applied to TEMgrids which were then dried in vacuum As in the workof Qian et al cracks were produced in the material byirradiation with an electron beam but a low acceleratingvoltage (75 kV) was used to avoid massive disruption ofthe thin composite film Individual tubes and bundles oftubes were observed to become stretched as the cracksgrew wider and pull-out was often seen In contrast tothe work of Qian et al however the tubes did notbreak This was probably because high quality arcgrown nanotubes were used rather than catalyticallyproduced tubes

The early work on nanotubeepoxy composites byAjayan et al has been described in the previoussection39 In 1998 Ajayan and co-workers described astudy of the mechanical behaviour of MWNTepoxycomposites in both tension and compression58 It wasfound that the compression modulus was higher thanthe tensile modulus indicating that load transfer to thenanotubes in the composite was much higher in com-pression The authors suggested that during load trans-fer to MWNTs only the outer layers are stressed intension whereas all the layers respond in compression

Nanotubeepoxy thin films prepared by Xu et alshowed excellent mechanical properties42 Compared toneat resin thin films for which the elastic modulus was42 GPa a 20 increase in modulus was seen when01 wt- MWNTs were added This was attributed topartial alignment of the MWNTs induced by spincoating Fracture behaviour of the films was investigatedby SEM and showed that pulled-out tubes were oftencovered with polymer suggesting strong interfacial adhe-sion Similar results were reported by Gojny et al59 Thisgroup showed that functionalising arc produced MWNTswith amine groups greatly enhanced bonding with theepoxy matrix As a result the outer shells of the tubesoften remained embedded in the matrix following pull-out as shown in Fig 10

All of the work described so far in this section hasinvolved multiwalled carbon nanotubes Ajayan et al

have described a study of SWNT-containing compo-sites60 Two types of composite were prepared first anSWNTepoxy composite and second a pressed pelletcontaining SWNTs and carbonaceous soot materialformed during nanotube synthesis The SWNTs wereused as-prepared ie generally grouped into bundlesrather than separated into individual tubes The com-posites were loaded axially in tension until failureoccurred and the fracture surfaces were then examinedin detail using SEM In some cases the tube bundleswere found to have been pulled out of the matrix duringthe deformation and fracture of the composites while inother cases the nanotubes were not entirely pulled outbut were stretched between two fracture surfaces Theauthors believed that during pull-out the nanotubeswere sliding axially within the ropes and that the failureobserved at large crack distances was not failure ofindividual tubes but the bundles pulling apart Studiesby micro-Raman spectroscopy supported this hypoth-esis It was found that little shift in the second order A1g

band occurred as a result of the application of axialtension showing that the individual nanotubes were notbeing significantly stretched To take full advantage ofthe high Youngrsquos modulus of SWNTs in polymer com-posites it is clear that load must be transferred effec-tively from the matrix to the nanotubes Ajayan et alsuggested that load transfer could be improved by break-ing bundles down into individual tube fragments beforedispersing in the matrix Alternatively the bundles them-selves could be reinforced by crosslinking the tubes withinbundles (eg by chemical treatments or irradiation)

Single-walled nanotubeepoxy composites have alsobeen prepared by Biercuk et al61 The Vickers hardnessof the polymer was found to increase monotonicallywith addition of SWNTs up to a factor of 35 at 2 wt-nanotube loading Greatly enhanced thermal conducti-vities were also observed Thus samples loaded with1 wt- unpurified SWNT material showed a 70increase in thermal conductivity at 40 K rising to125 at room temperature

The preparation of SWNTpolyelectrolyte compositesby the layer-by-layer method has been described in theprevious section49 The films produced in this way wereshown to have exceptional mechanical properties Thusthe average ultimate tensile strength of the films wasfound to be 220 MPa with some measurements as highas 325 MPa This is several orders of magnitude greater

10 Image of epoxy coated MWNT subject to tensile

stress as result of electron irradiation telescopic

pull-out of outer tube layer has occurred59

9 TEM image of crack in MWNTpolystyrene thin film

induced by thermal stresses34

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 37

than the tensile strength of strong industrial plastics andindicates that the layer-by-layer method has great pro-mise for the production of strong nanotube-containingcomposites

Electrical and optical properties ofcarbon nanotubepolymer compositesThe effect of adding nanotubes on the electrical pro-perties of polymers has been studied by many groupsThese studies fall into a number of distinct categories Insome cases the nanotubes have been used to increase theconductivity of relatively low cost polymers as analternative to currently used fillers such as carbon blackOther studies have involved the incorporation of nano-tubes into conducting polymers such as polyaniline Stillfurther work has been aimed at specific applicationssuch as photovoltaic devices or supercapacitors In manycases improved mechanical properties have been avaluable by-product of the inclusion of nanotubes Theaddition of nanotubes to low cost bulk polymers isconsidered first

Electrostatic dissipationImproving the electrical conductivity of bulk polymers isimportant in a number of applications For example insome aircraft components enhanced conductivity isrequired to provide electrostatic discharge and electro-magnetic radio frequency interference protection Staticelectrical dissipation is also needed in other applicationsincluding computer housings and exterior automotiveparts With these types of application in mind Sandleret al investigated the electrical properties of nanotubeepoxy composites40 Matrix resistivities of about 100 V mwith filler volume fractions as low as 01 vol- wereachieved These figures represented an advance on thebest conductivity values previously obtained with carbonblack in the same epoxy matrix

Lozano and co-workers prepared composites ofcatalytically grown carbon nanofibres in a polypropy-lene matrix5162 The nanofibres were purified and func-tionalised to facilitate bonding with the matrix and highshear processing was used to disperse the fibres in thepolymer matrix Rheological and microscopic analysisshowed that a homogeneous dispersion of nanofibreswas formed and a percolation threshold for electricalconduction of 9ndash18 wt- was observed

These and other published studies suggest that carbonnanotubes have great promise in reducing electrostaticcharging of bulk polymers Nanotubes also have otheradvantages over conventional fillers such as carbonblack and carbon fibres in that they are more amenableto processing and can be more easily dispersed through-out the matrix In addition to the studies available in theopen literature there is undoubtedly a great deal ofcommercial work being carried out on the use of nano-tubes to reduce the electrostatic charging of plastics Infact nanotube containing plastics are already being usedin commercial products63 These include fuel lines inautomobiles where the nanotubes help to dissipate anydangerous charge which may build up Thermoplasticpolymers containing nanotubes are also used in someexterior automobile parts so that they can be earthedduring electrostatic painting

Conducting polymer compositesPolyaniline is a conducting polymer with many attrac-tive characteristics such as processability and environ-mental stability There has been interest in addingnanotubes to this polymer to provide both enhancedconductivity and improved mechanical properties Asmentioned above in the section lsquoPreparation of carbonnanotubepolymer compositesrsquo Cochet et al have pre-pared a MWNTpolyaniline composite by in situ poly-merisation44 The nanotubes were found to have astrong influence on the transport properties of thepolymer Thus the room temperature resistivity of thecomposite was found to be an order of magnitude lowerthan that of pure polyaniline while the low temperatureresistivity was much smaller than that of either poly-aniline or of MWNTs The temperature dependence ofresistivity was also weaker than that of polyanilinealone These observations were explained by assumingthat in situ polymerisation favours charge transfer betweenpolyaniline and MWNTs resulting in an overall materialwhich is more conducting than the starting componentsRaman studies indicated that these charge transferprocesses involved a site-selective interaction betweenthe quinoid ring of the polymer and the MWNTs Othergroups have also demonstrated excellent conductivity inMWNTpolyaniline composites6465 while a group fromDuPont have shown that SWNTpolyaniline compositescan be used as printable conductors for organic elec-tronics devices66

Optoelectronic propertiesSeveral groups have prepared nanotube compositesusing the conjugated polymer poly(p-phenylene viny-lene) (PPV) and its derivatives Polymers of this classhave electroluminescent properties and are widely usedin light-emitting diodes They are also used in photo-voltaic devices In an early study workers from Irelandand the USA described the synthesis of a compositematerial containing arc produced MWNTs in a PPVderivative PmPV67 The composite was prepared usinga solution mixing method and good wetting of the tubesby the polymer was achieved One of the aims of addingnanotubes to the polymer was to increase its electricalconductivity Previous attempts to achieve this by dopinghad resulted in degradation of the optical properties ofthe polymer In the case of nanotubes an increase in theelectrical conductivity of the polymer by up to eightorders of magnitude was seen with no concomitantdegradation of optical properties This was apparentlybecause the tubes acted as nanometric heat sinks pre-venting the build up of large thermal effects whichdamage these conjugated systems It was also shown inthis work that the composite could be used as theemissive layer in an organic light-emitting diode (LED)

As mentioned above several groups have been inter-ested in using nanotubepolymer composites in photo-voltaic devices The photovoltaic effect involves thegeneration of electron and hole pairs and their sub-sequent collection at opposite electrodes In organicmaterials the photon absorption excites an electron intothe conduction band which leads to a generation ofbound electronndashhole pairs or excitons These excitonshave to dissociate into free charges in order to be trans-ported to the electrodes Exciton dissociation can occurat a junction between a polymer and an electron acceptor

Harris Carbon nanotube composites

38 International Materials Reviews 2004 VOL 49 NO 1

molecule This allows the preferential transfer of the elec-trons into the electron acceptor molecule while leavingthe holes to be preferentially transported through thepolymer a process known as photoinduced chargetransfer In 1992 photoinduced charge transfer from aconducting polymer to C60 was discovered68 and thishas led to the fabrication of several efficient photo-voltaic systems using fullerenepolymer combinations69

Photovoltaic devices based on nanotubepolymer com-binations have now been fabricated by several groupsand this work will now be summarised

The earliest attempt to fabricate a nanotubepolymerphotovoltaic device appears to be the 1996 work ofRomero et al70 In this work a layer of MWNTs wasdeposited on Teflon and a PPV derivative was cast ontothis with a thin aluminium layer on top The devicedisplayed clear diode behaviour but only a small photo-current was observed when the junction was irradiatedwith light from a HendashNe laser This may have been dueto the relatively thick PPV active layer Ago et alfabricated a nanotube containing photovoltaic deviceby depositing a layer of PPV onto a glass-supportedMWNT film71 The interaction between MWNTs andPPV was investigated by means of photoluminescence(PL) measurements The photoluminescence efficiencyof the PPV was found to be markedly reduced by thepresence of the MWNT layer Since photoluminescenceis a result of radiative recombination a reduction mayimply improved exciton separation which is what onewishes to achieve in a photovoltaic device The devicefabricated by Ago et al is illustrated in Fig 11a Thisshowed a clear diode characteristic in the dark andupon illumination through the aluminium electrode aphotocurrent was observed A quantum efficiency approxi-mately twice that of the standard indiumndashtin oxidedevice was reported

Kymakis and Amaratunga used a slightly differentapproach to prepare a nanotubeconjugated polymerphotovoltaic device72 In this work composite SWNTpoly(3-octylthiophene) films were drop or spin cast froma solution onto indiumndashtin oxide (ITO) and quartz sub-strates to produce the arrangement shown in Fig 11bDevices were also prepared using the pure P3OT poly-mer A considerable enhancement of the photovoltaic

effect was observed with the P3OTSWNTs blend devicecompared with the pure polymer device

Non-linear optical propertiesAs mentioned above in the section lsquoStructure and pro-perties of carbon nanotubesrsquo it has been demonstratedthat suspensions of both SWNTs and MWNTs havenon-linear optical properties Nanotubepolymer com-posites may therefore have applications in optical devicesTo date however there has been little work on theoptical limiting properties of nanotube containing com-posites In one of the first studies in this area Chen et aldemonstrated third order optical non-linearity in a SWNTpolyimide composite73 Recently OrsquoFlaherty et al pre-pared a composite in which MWNTs were dispersed inthe polymer poly(99-di-n-octylfluorenyl-279-diyl) (PFO)74

Suspensions of the nanotubePFO composite in toluenewere prepared and the optical limiting properties of thesolutions were measured using nanosecond laser pulsesMaximum optical limiting was observed at carbonnanotube loadings in excess of 38 wt- relative to thepolymer

Carbon nanotubeceramic compositesCeramics have high stiffnesses and thermal stabilities butrelatively low breaking strengths Incorporating carbonnanotubes into a ceramic matrix might be expected toproduce a composite with both toughness and high tem-perature stability However achieving a homogeneousdispersion of tubes in an oxide with strong bondingbetween tubes and matrix presents rather more of achallenge than incorporating tubes into a polymer Variousapproaches have been adopted for incorporating nano-tubes into oxides such as alumina and silica and thesewill now be summarised

Peigney and co-workers have been among the pioneersin the synthesis of nanotubeoxide composites75ndash77 Theyhave developed an ingenious technique which involvesthe use of catalysts such as iron-containing a-aluminaPassing a CH4ndashH2 mixture over the catalysts resultedin the in situ growth of nanotubes on iron particlesproducing the precursor for a nanotubeoxide compo-site The resulting powders can then be hot pressed toform the final composite material The mechanicalproperties of these composites were initially disappoint-ing72 Thus the fracture strengths of the nanotubeFendashAl2O3 composites were only marginally higher than thatof Al2O3 and were generally markedly lower than thoseof the carbon-free FendashAl2O3 composites The fracturetoughness values were lower than or similar to that ofAl2O3 However SEM observations of composite frac-tures indicate that the nanotube bundles which are veryflexible could dissipate some fracture energy A similarmethod was used by An and Lim to prepare nanotubeAl2O3 composites in this case using C2H2 as the carbonsource78 Nanotube contents of up to 125 wt- wereachieved and a significant reinforcement effect wasreported

Achieving alignment of nanotubes in a ceramic matrixis much more difficult than in a polymer since ceramicsdo not lend themselves so easily to processing Howevera number of groups have shown that aligned nanotubemetal oxide composites can indeed be produced Peigney

ITO5indiumndashtin oxide PPV5poly(p-phenylene vinylene)

P3OT5poly(3-octylthiophene)

11 a Illustration of photovoltaic device fabricated by Ago

et al71 and b photovoltaic device fabricated by

Kymakis and Amaratunga72

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 39

et al exploited the fact that when their grains are main-tained at sufficiently low sizes (less than about 1 mm)many polycrystalline ceramics become superplasticallowing the use of thermomechanical processing forthe forming of these materials A series of cobalt andiron containing oxide catalysts were prepared77 Theprimary grain size in the powders was smaller than100 nm but in some of them the as produced primarygrains were strongly aggregated and required ball millingin order to reduce aggregate size to less than 1 mmNanotubes were then prepared in situ as describedabove and the powders were then extruded at hightemperatures in a graphite die under vacuum The bestresults were achieved with nanotubes incorporated intoFendashAl2O3 and FeCondashMgAl2O4 matrices In both casesthe nanotubes were not damaged by the processing andalignment of the nanotubes was clearly observed Infact the presence of the nanotubes seemed to facilitatethe superplastic forming by inhibiting grain growth andacting as lubricating agents It was demonstrated thatthe aligned nanotubemetal oxide composites showed ananisotropy of electrical conductivity which could beadjusted by controlling the nanotube content

One of the first studies of a SWNTceramic compositewas described by Zhan et al79 This group used thetechnique of spark plasma sintering (SPS) to prepareAl2O3 composites containing 10 vol- SWNTs Theadvantage of spark plasma sintering is that it allowsceramic powders to be annealed at lower temperaturesand for much shorter times than other sintering pro-cesses leading to the fabrication of fully dense ceramicsor composites with nanocrystalline microstructures undermild conditions Unlike some of the other processes usedto prepare nanotubeoxide composites spark plasmasintering does not damage the nanotubes and the nano-tube bundles could be seen to be located mainly at theboundaries of the Al2O3 grains in good contact with theAl2O3 matrix Before sintering the powders had beenmixed by ball milling which produced a reasonablyhomogeneous dispersion again without damaging thenanotubes The fracture toughnesses of the final com-posites were remarkable ndash about twice that of the fullydensified but unreinforced Al2O3 matrix In contrast toprevious work on nanotubeceramic composites thefracture toughness was found to increase with nanotubedensity This is thought to be due to the formation ofentangled networks of single-walled carbon nanotubeswhich may inhibit crack propagation A subsequentstudy of the nanotubeAl2O3 composites showed thatthey had high electrical conductivities80

Various techniques have been used to prepare carbonnanotubesilica composites Seeger et al described amethod which involved preparing a composite gel ofMWNTs with tetraethoxysilane (TEOS) and then sinter-ing this at 1150uC in argon81 A drawback of this tech-nique was that the sintering process led to a partialcrystallisation of the SiO2 resulting in a very inhomo-geneous matrix An alternative method which avoidsthese problems has been developed82 This involvesusing a Nd YAG laser to rapidly heat a TEOSnano-tube mixture resulting in partial melting of the matrixThis produced an amorphous silica matrix with nocrystallisation

A novel nanotubesilica composite material withpotentially useful optical properties has recently been

described by Han et al83 These workers synthesisedfilms consisting of silica spheres with diameters rangingfrom 200 to 650 nm Molybdenumcobalt catalyst parti-cles were then deposited onto the silica spheres and usedto catalyse the growth of SWNTs Since SWNTs arehighly non-linear and fast-switching materials they canbe incorporated into an optically confining environmentto achieve these characteristics at relatively low levels oflaser intensities

There have been a few studies on preparing nanotubecomposites using oxides other than alumina and silicaVincent et al reported the production of nanotubeTiO2

composites using solndashgel methods84 Possible applica-tions of the composites included optical non-linear waveguides and unidimensional conductive films Solndashgelmethods were also used by Sakamoto and Dunn toprepare SWNTV2O5 composites for use as electrodesin secondary lithium batteries85 Little work has beencarried out so far on incorporating nanotubes into non-oxide ceramics In one of the few published studies Maet al fabricated nanotubesilicon carbide composites bymixing SiC nanoparticles with 10 wt- MWNTs andhot pressing at 2000uC86 They found that the nanotubesplayed a strengthening and toughening role in thecomposite both bend strength and fracture toughnessincreased by about 10 compared to monolithic SiCThere have also been some initial studies on carbonnanotubecarbon fibre composites8788

Carbon nanotubemetal compositesComposite materials containing conventional carbonfibres in a metal matrix such as aluminium or mag-nesium are used in a number of specialist applica-tions8990 Such composites combine low density withhigh strength and modulus making them particularlyattractive to the aerospace industry There is growinginterest in the addition of carbon nanotubes to metalmatrices although only a small number of studies havebeen published to date Kuzumaki et al described thepreparation of a nanotubealuminium composite in 199891

Their method involved mixing a nanotube sample with afine aluminium powder mounting the mixture in a6 mm silver sheath and then drawing and heating thewire at 700uC in a vacuum furnace The result was acomposite wire in which the nanotubes were partiallyaligned along the axial direction The tensile strengths ofthe as prepared composite wires were comparable to thatof pure aluminium but the composite wires retained thisstrength after prolonged annealing at 600uC while thestrength of pure aluminium decreased by about 50after this treatment The same group subsequently pre-pared nanotubetitanium composites which showed alarge increase in hardness and modulus compared topure titanium92

Chen et al prepared nanotubenickelphosphorus com-posite coatings on stainless steel substrates using anelectroless plating method93 The coatings exhibited higherwear resistance and lower friction coefficients than SiCNiP and graphiteNiP composite coatings The samegroup prepared nanotubecopper composites by a powdermetallurgy technique and showed that the presence ofthe nanotubes again resulted in improved wear rates andfriction coefficients These results suggest that nanotube

Harris Carbon nanotube composites

40 International Materials Reviews 2004 VOL 49 NO 1

metal composites may have potential in tribologicalapplications

DiscussionThe subject of carbon nanotube composites has grownrapidly over the past 5 years or so into a vibrant andexciting area of research It is important to recognisehowever that this is still a new subject and the workcarried out so far may have hardly scratched the surfaceof what is possible For example the great majority ofthe studies that have been published to date haveinvolved the use of polymer matrices there is great scopefor more work on incorporating nanotubes into cera-mics metals and other forms of carbon Also most ofthe work up to now has been aimed at exploiting themechanical properties of carbon nanotubes rather thantheir electronic or optical properties although interest inthose is growing

When considering the mechanical performance ofcarbon nanotube composites it is clear that these dependstrongly on whether arc grown tubes or catalyticallyproduced tubes have been used In considering thecharacteristics of the two types of tube it is instructive tocompare the work of Qian et al34 with that of Watts andHsu57 Both incorporated the nanotubes into polymersand then used TEM to characterise the composites andstudy their fracture behaviour Qian et al used cataly-tically produced tubes and these were found to bewell dispersed throughout the polymer matrix with noclustering together of tubes However during compositefracture the tubes were often observed to break sug-gesting that they have relatively poor strength Wattsand Hsu on the other hand employed much moreperfect nanotubes produced by arc evaporation andTEM studies revealed superior mechanical propertieswith no observed breaking However the nanotubestended to cluster together so that during crack growththe tubes could quite easily slide past one another Thuswhile it is clear that the best mechanical performancewill be achieved with arc grown nanotubes it is impor-tant that the tubes be well dispersed throughout thematrix in order fully to exploit their properties

Equally important is achieving good bonding betweennanotubes and matrix Many studies have shown that thisinterfacial bonding can be enhanced by functionalising thenanotubes before embedding them in the matrix Theeffect of this can be seen in Fig 10 from the work ofGojny et al59 where bonding between MWNTs and anepoxy matrix is so strong that the outer shells of the tuberemain embedded in the matrix following pull-out Apossible disadvantage of functionalisation is that it couldweaken the nanotubes by disrupting the perfect cylindricalstructure This might be expected to be a particularproblem with single-walled tubes However functionali-sation of the nanotubes may not be necessary to producebonding with a polymer matrix Liao and Li havemodelled the interfacial characteristics of a nanotubepolystyrene composite assuming no covalent bondingbetween the tubes and the matrix94 It was found thatelectrostatic and van der Waals interactions contributedsignificantly to the interfacial stress transfer A nanotubepull-out simulation suggested that the interfacial shearstress of the nanotubepolystyrene system is about160 MPa significantly higher than for most carbon fibrereinforced polymer composite systems It is also worth

mentioning recent work by Star and co-workers9596

in which it was demonstrated that poly(5-alkoxy-m-phenylenevinylene)-co-[(25-dioctyloxy-p-phenylene)vinylene] (PAmPV) derivatives could be induced towrap around SWNTs forming strong non-covalentbonds

The quality of the nanotubes and the nature of theinterfacial bonding also has a strong influence on theelectrical properties of nanotubepolymer compositesArc grown tubes can have resistivities lower than that ofsingle crystal graphite but as already discussed theirelectronic properties depend on their precise structureAt present we do not have a reliable method of separat-ing metallic and semiconducting tubes so it is notpossible to exploit their full potential Catalytically pro-duced tubes are likely to be poorer conductors thanmetallic arc-grown tubes owing to their relatively dis-ordered structure Nevertheless catalytically grown tubesare sufficiently conducting to be useful in some com-mercial applications such as electrically conductive paintsThe influence of interfacial bonding on electronic pro-perties was demonstrated in the work of Cochet et al44

These workers prepared a MWNTpolyaniline com-posite by in situ polymerisation and found it to haveexcellent electrical transport properties This was attrib-uted to strong interaction between tubes and polymer

As has been shown in this review the possible appli-cations of carbon nanotube composites range widelyfrom electrically conductive paints to optical devicesOnly a relatively small number of these new materialshave been used commercially however and there arestill many problems to overcome before the full potentialof nanotube containing composites can be realised Notleast of these problems is the high cost of good qualitycarbon nanotubes Currently the commercial price ofunpurified arc-grownMWNTs is approximately $10000gwhile purified MWNTs cost around $30000g Cataly-tically grown MWNTs are much cheaper but as empha-sised throughout this review have generally inferiorproperties As far as SWNTs are concerned unpurifiedsamples cost around $5000g while purified SWNTs are$50000g As a comparison the price of gold is cur-rently about $1200g The development of new techni-ques for the low cost synthesis of high quality nanotubesis clearly essential This might be achieved throughimproved catalytic processes and there is currently alarge amount of work being carried out towards thisend Alternatively it may be possible to develop newnon-catalytic techniques for the synthesis of fullerene-related nanotubes In this connection it is interesting tonote that nanotubes can be produced by the high tem-perature heat treatment of conventional carbons such asmicroporous non-graphitising carbons9798 or carbonblack99100 Further work in this area would be of value

Acknowledgement

I thank Milo Shaffer for helpful comments

References1 P J F Harris lsquoCarbon nanotubes and related structures ndash new

materials for the twenty-first centuryrsquo 1999 Cambridge

Cambridge University Press

2 S Iijima Nature 1991 354 56ndash58

3 W Kratschmer L D Lamb K Fostiropoulos and D R

Huffman Nature 1990 347 354ndash358

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 41

4 S Iijima and T Ichihashi Nature 1993 363 603ndash605

5 D S Bethune C H Kiang M S De Vries G Gorman R Savoy

J Vasquez and R Beyers Nature 1993 363 605ndash607

6 H Dai A G Rinzler P Nikolaev A Thess D T Colbert and

R E Smalley Chem Phys Lett 1996 260 471ndash475

7 M M J Treacy T W Ebbesen and J M Gibson Nature 1996

381 678ndash680

8 E W Wong P E Sheehan and C M Lieber Science 1997 277

1971ndash1975

9 M R Falvo G J Clary R M Taylor V Chi F P Brooks

S Washburn and R Superfine Nature 1997 389 582ndash584

10 K Anazawa K Shimotani C Manabe H Watanabe and M

Shimizu Appl Phys Lett 2002 81 739ndash741

11 P Kim L Shi A Majumdar and P L McEuen Phys Rev Lett

2001 87 paper 215502

12 E T Thostenson Z F Ren and T W Chou Compos Sci

Technol 2001 61 1899ndash1912

13 K T Lau and D Hui Composites Part B Engineering 2002 33

263ndash277

14 Y Huh J Y Lee J Cheon Y K Hong J Y Koo T J Lee and

C J Lee J Mater Chem 2003 13 2297ndash2300

15 M Endo T Koyama and Y Hishiyama Jpn J Appl Phys 1976

15 2073ndash2076

16 S C Lyu B C Liu T J Lee Z Y Liu C W Yang C Y Park

and C J Lee Chem Commun 2003 734ndash735

17 M-F Yu O Lourie M J Dyer K Moloni T F Kelly and R S

Ruoff Science 2000 287 637ndash640

18 P Poncharal Z L Wang D Ugarte and W A De Heer Science

1999 283 1513ndash1516

19 J-P Salvetat A J Kulik J-M Bonard G A D Briggs

T Stockli K Metenier S Bonnamy F Beguin N A Burnham

and L Forro Adv Mater 1999 11 161ndash165

20 M-F Yu B S Files S Arepalli and R S Ruoff Phys Rev Lett

2000 84 5552ndash5555

21 M Terrones Int Mater Rev to be published

22 J W Mintmire B I Dunlap and C T White Phys Rev Lett

1992 68 631ndash634

23 R Saito M Fujita G Dresselhaus and M S Dresselhaus Phys

Rev B 1992 46 1804ndash1811

24 N Hamada S Sawada and A Oshiyama Phys Rev Lett 1992

68 1579ndash1581

25 T W Ebbesen H J Lezec H Hiura J W Bennett H F Ghaemi

and T Thio Nature 1996 382 54ndash56

26 J E Riggs D B Walker D L Carroll and Y-P Sun J Phys

Chem B 2000 104 7071ndash7076

27 S C Tsang Y K Chen P J F Harris and M L H Green

Nature 1994 372 159ndash162

28 R M Lago S C Tsang K L Lu Y K Chen and M L H

Green Chem Commun 1995 1355ndash1356

29 Y-P Sun K F Fu Y Lin and W Huang Acc Chem Res 2002

35 1096ndash1104

30 S B Sinnott J Nanosci Nanotechnol 2002 2113ndash123

31 M S P Shaffer X Fan and A H Windle Carbon 1998 36

1603ndash1612

32 M S P Shaffer and A H Windle Adv Mater 1999 11 937ndash941

33 D E Hill Y Lin A M Rao L F Allard and Y-P Sun

Macromolecules 2002 35 9466ndash9471

34 D Qian E C Dickey R Andrews and T Rantell Appl Phys

Lett 2000 76 2868ndash2870

35 R Andrews D Jacques D Qian and T Rantell Acc Chem Res

2002 35 1008ndash1017

36 R Haggenmueller H H Gommans A G Rinzler J E Fischer

and K I Winey Chem Phys Lett 2000 330 219ndash225

37 P Potschke T D Fornes and D R Paul Polymer 2002 43 3247ndash

3255

38 M Sennett E Welsh J B Wright W Z Li J G Wen and Z F

Ren Appl Phys A 2003 76 111ndash113

39 P M Ajayan O Stephan C Colliex and D Trauth Science 1994

265 1212ndash1214

40 J Sandler M S P Shaffer T Prasse W Bauhofer K Schulte

and H Windle Polymer 1999 40 5967ndash5971

41 A Allaoui S Bai H M Cheng and J B Bai Compos Sci

Technol 2002 62 1993ndash1998

42 X Xu M M Thwe C Shearwood and K Liao Appl Phys Lett

2002 81 2833ndash2835

43 D Puglia L Valentini and J M Kenny J Appl Polymer Sci

2003 88 452ndash458

44 M Cochet W K Maser A M Benito M A Callejas M T

Martınez J-M Benoit J Schreiber and O Chauvet Chem

Commun 2001 1450ndash1451

45 M S P Shaffer and K Koziol Chem Commun 2002 2074ndash

2075

46 C Park Z Ounaies K A Watson R E Crooks J Smith S E

Lowther J W Connell E J Siochi J S Harrison and T L St

Clair Chem Phys Lett 2002 364 303ndash308

47 K Jurewicz S Delpeux V Bertagna F Beguin and E

Frackowiak Chem Phys Lett 2001 347 36ndash40

48 M Hughes G Z Chen M S P Shaffer D J Fray and A H

Windle Chem Mater 2002 14 1610ndash1613

49 A A Mamedov N A Kotov M Prato D M Guldi J P

Wicksted and A Hirsch Nature Mater 2002 1 190ndash194

50 J H Rouse and P T Lillehei Nano Lett 2003 3 59ndash62

51 K Lozano and E V Barrera J Appl Polym Sci 2001 79 125ndash

133

52 L Valentini J Biagiotti J M Kenny and S Santucci Compos

Sci Technol 2003 63 1149ndash1153

53 J Sandler P Werner M S P Shaffer V Demchuk V Altstadt

and A H Windle Composites Part A Appl Sci Manuf 2002 33

1033ndash1039

54 R Czerw Z X Guo P M Ajayan Y P Sun and D L Carroll

Nano Lett 2001 1 423ndash427

55 B P Grady F Pompeo R L Shambaugh and D E Resasco

J Phys Chem B 2002 106 5852ndash5858

56 M Cadek J N Coleman V Barron K Hedicke and W J Blau

Appl Phys Lett 2002 81 5123ndash5125

57 P C P Watts and W K Hsu Nanotechnology 2003 14 L7ndashL10

58 L S Schadler S C Giannaris and P M Ajayan Appl Phys Lett

1998 73 3842ndash3844

59 F H Gojny J Nastalczyk Z Roslaniec and K Schulte Chem

Phys Lett 2003 370 820ndash824

60 P M Ajayan L S Schadler C Giannaris and A Rubio Adv

Mater 2000 12 750ndash753

61 M J Biercuk M C Llaguno M Radosavljevic J K Hyun A T

Johnson and J E Fischer Appl Phys Lett 2002 80 2767ndash2769

62 K Lozano J Bonilla-Rios and E V Barrera J Appl Polym Sci

2001 80 1162ndash1172

63 V Jamieson New Sci 15 Mar 2003 30

64 H Zengin W S Zhou J Y Jin W R Czerw D W Smith

L Echegoyen D L Carroll S H Foulger and J Ballato Adv

Mater 2002 14 1480ndash1483

65 J Deng X Ding W Zhang Y Peng J Wang X Long P Li and

A S C Chan Eur Polym J 2002 38 2497ndash2501

66 G B Blanchet C R Fincher and F Gao Appl Phys Lett 2003

82 1290ndash1292

67 S A Curran P M Ajayan W J Blau D L Carroll J N

Coleman A B Dalton A P Davey A Drury B McCarthy

S Maier and A A Strevens Adv Mater 1998 10 1091ndash1093

68 N S Sariciftci L Smilowitz A J Heeger and F Wudl Science

1992 258 1474ndash1476

69 N Camaioni L Garlaschelli A Geri M Maggini G Possamai

and G Ridolfi J Mater Chem 2002 12 2065ndash2070

70 D B Romero M Carrard W De Heer and L Zuppiroli

Adv Mater 1996 8 899ndash902

71 H Ago K Petritsch M S P Shaffer A H Windle and R H

Friend Adv Mater 1999 11 1281ndash1285

72 E Kymakis and G A J Amaratunga Appl Phys Lett 2002 80

112ndash114

73 Y-C Chen N R Raravikar L S Schadler P M Ajayan Y-P

Zhao T-M Lu G-C Wang and X-C Zhang Appl Phys Lett

2002 81 975ndash977

74 S A OrsquoFlaherty R Murphy S V Hold M Cadek J N

Coleman and W J Blau J Phys Chem B 2003 107 958ndash964

75 A Peigney C Laurent F Dobigeon and A Rousset J Mater

Res 1997 12 613ndash615

76 E Flahaut A Peigney C Laurent C Marliere F Chastel and

A Rousset Acta Mater 2000 48 3803ndash3812

77 A Peigney E Flahaut C Laurent F Chastel and A Rousset

Chem Phys Lett 2002 352 20ndash25

78 J W An and D S Lim J Ceram Process Res 2002 3 201ndash204

79 G-D Zhan J D Kuntz J L Wan and A K Mukherjee Nature

Mater 2003 2 38ndash42

80 G-D Zhan J D Kuntz J E Garay and A K Mukherjee Appl

Phys Lett 2003 83 1228ndash1230

81 T Seeger T Kohler T Frauenheim N Grobert M Ruhle

M Terrones and G Seifert Chem Commun 2002 34ndash35

82 T Seeger G De La Fuente W K Maser A M Benito M A

Callejas and M T Martınez Nanotechnology 2003 14 184ndash187

83 H Han S Vijayalakshmi A Lan Z Iqbal H Grebel E Lalanne

and A M Johnson Appl Phys Lett 2003 82 1458ndash1460

Harris Carbon nanotube composites

42 International Materials Reviews 2004 VOL 49 NO 1

84 P Vincent A Brioude C Journet S Rabaste S T Purcell J Le

Brusq and J C Plenet J Non-Cryst Sol 2002 311 130ndash137

85 J S Sakamoto and B Dunn J Electrochem Soc 2002 149 A26ndashA30

86 R Z Ma J Wu B Q Wei J Liang and D H Wu J Mater Sci

1998 33 5243ndash5246

87 R Andrews D Jacques A M Rao T Rantell F Derbyshire

Y Chen J Chen and R C Haddon Appl Phys Lett 1999 75

1329ndash1331

88 E T Thostenson W Z Li D Z Wang Z F Ren and T W

Chou J Appl Phys 2002 91 6034ndash6037

89 D D L Chung lsquoCarbon fiber compositesrsquo chap 7 1994 Boston

Butterworth-Heinemann

90 S Rawal JOM 2001 53 14ndash17

91 T Kuzumaki K Miyazawa H Ichinose and K Ito J Mater

Res 1998 13 2445ndash2449

92 T Kuzumaki O Ujiie H Ichinose and K Ito Adv Eng Mater

2000 2 416ndash418

93 W X Chen J P Tu L Y Wang H Y Gan Z D Xu and X B

Zhang Carbon 2003 41 215ndash222

94 K Liao and S Li Appl Phys Lett 2001 79 4225ndash4227

95 A Star J F Stoddart D Steuerman M Diehl A Boukai E W

Wong X Yang S W Chung H Choi and J R Heath

Angewandte Chemie-Int Ed 2001 40 1721ndash1725

96 A Star Y Liu K Grant L Ridvan J F Stoddart D W

Steuerman M R Diehl A Boukai and J R Heath

Macromolecules 2003 36 553ndash560

97 P J F Harris S C Tsang J B Claridge and M L H Green

J Chem Soc Faraday Trans 1994 90 2799ndash2802

98 A A Setlur S P Doherty J Y Dai and R P H Chang Appl

Phys Lett 2000 76 3008ndash3010

99 S P Doherty D B Buchholz B J Li and R P H Chang

J Mater Res 2003 18 941ndash949

100 D B Buchholz S P Doherty and R P H Chang Carbon 2003

41 1625ndash1634

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 43

Page 5: Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P. J. F. Harris * Carbon nanotubes are molecular-scale tubes of graphitic carbon with

Andrews et al used shear mixing to disperse catalytically-produced nanotubes in a range of polymers includinghigh impact polystyrene acrylonitrile butadiene styreneand polypropylene35 The dispersion of nanotubes wasdetermined as a function of mixing energy and tem-perature and the composites were formed into fibres andthin films Haggenmueller et al used a combination ofsolvent casting and melt mixing methods to disperseSWNTs in polymethylmethacrylate (PMMA)36 Compositefibres with a high degree of nanotube orientation wereproduced by melt spinning

Polycarbonate is another thermoplastic polymer thathas been used as a matrix for nanotube composites3738

Sennet et al have used melt processing techniques todisperse and align carbon nanotubes in polycarbonate38

Dispersion was achieved by mixing the catalyticallyproduced MWNTs and SWNTs with polycarbonateresin in a conical twin-screw extruder and alignment wascarried out using a fibre spinning apparatus By opti-mising mixing time and fibre draw rates excellent dis-persion and alignment were accomplished

Nanotubeepoxy composites have also been preparedby a number of groups39ndash43 The earliest work in thisarea was carried out by Ajayan et al39 In 1994 thisgroup described a study in which purified tubes wereembedded into an epoxy resin which was then cut intothin slices with a diamond knife Their initial aim indoing this was not to produce a composite material butto obtain thin TEM specimens which it was hoped wouldcontain cross-sectional images of nanotubes However nocross-sections were observed following sectioning insteadthe nanotubes were found to have become aligned in thedirection of the knife movement as shown in Fig 8According to Ajayan et al the alignment is primarily aconsequence of extensional or shear flow of the matrixproduced by the cutting Of course alignment can bemuch more readily achieved when the epoxy is in a liquidstate Xu et al showed that some alignment of nanotubesin nanotubeepoxy composites can been achieved by usingspin coating methods42 The cure reaction of a SWNT

epoxy resin composite has been analysed using thermalanalysis and Raman spectroscopy43 The SWNTs werefound to have a catalysing effect on the cure reaction

An alternative method for preparing nanotubepoly-mer composites is to use the monomer rather than thepolymer as a starting material and then carry out in situpolymerisation Cochet et al were among the first to usethis method to prepare a MWNTpolyaniline compo-site44 Nanotubes prepared by arc evaporation wereused and were sonicated in a solution of HCl to achievea dispersion The aniline monomer again in HCl wasadded to the suspension and a solution of an oxidantwas slowly added with constant sonication and coolingSonication was continued in an ice bath for 2 h and thecomposite was then obtained by filtering rinsing anddrying In this way it was possible to prepare com-posites with high MWNT loadings (up to 50 wt-)Transport measurements on the composite revealedmajor changes in the electronic behaviour confirmingstrong interaction between nanotubes and polymer asdiscussed below A number of other nanotubepolymercomposites have been prepared using in situ poly-merisation including MWNTpolystyrene45 and SWNTpolyimide46 Several groups have used electrochemicalpolymerisation to grow porous composite films ofMWNTand polypyrrole for use as supercapacitors4748

In the techniques described so far the aim has been toproduce composites in which the nanotubes are dis-tributed evenly throughout the polymer (either ran-domly oriented or aligned) For some applicationshowever a layered arrangement is advantageous Photo-voltaic devices containing layers of nanotubes or nanotubepolymer composites have been prepared as discussed inthe section lsquoElectrical and optical properties of carbonnanotubepolymer compositesrsquo below The preparationof multilayer SWNTpolymer composites with excellentmechanical properties has also been described49 Themethod involved the layer-by-layer deposition of SWNTsand polymer onto a substrate followed by crosslinkingIn this way composites with SWNT loadings as high as

8 Alignment of nanotubes in polymer matrix following cutting with microtome arrows indicate buckled nanotubes39

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 35

50 wt- could be obtained Coating was carried out byalternate dipping of a glass slide or silicon wafer intodispersions of SWNTs (stabilised by acid treatment) andpolymer solutions The layers were held together by vander Waals forces and by electrostatic attraction betweenthe negatively charged SWNTs and a positively chargedpolyelectrolyte such as branched polyethyleneimine (PEI)Coatings containing up to 50 SWNTPEI bilayers couldbe built up in this way When the procedure was com-plete the multilayer films were heated to 120uC topromote crosslinking The films could then be lifted offthe substrate to obtain uniform free-standing mem-branes A rather similar method was used to synthesisemultilayer composites films of the polyelectrolyte poly-diallyldimethylammonium chloride and SWNTs50

Before proceeding to discuss the properties of carbonnanotubepolymer composites in detail it is worth brieflyconsidering the effect that the inclusion of nanotubescan have on polymer microstructure Although thistopic has not yet been widely studied a number ofgroups have demonstrated that the crystallisation andmorphology of the polymers can be strongly affected bysmall additions of nanotubes For example Lozano andBarrera51 showed that the addition of catalytically growncarbon nanofibres can promote nucleation of polypro-pylene Similar observations were made by Valentiniet al52 The crystallinity of a polymers has also beenshown to increase with the addition of carbon nano-fibres5354 Changes in the crystal structure of polymerscan also be induced by nanotubes55 More work on theseeffects would be welcome

Mechanical properties of carbonnanotubepolymer compositesMuch of the work on the preparation of nanotubepolymer composites has been driven by a desire toexploit the tubesrsquo stiffness and strength Even where theinterest has been focused on other properties the abilityof nanotubes to improve the mechanical characteristicsof a polymer has often been a valuable added benefit Inearly work the measured properties were often dis-appointing but such studies were of value in under-standing the reasons for composite failure and inidentifying the critical issues which need to be addressed

Shaffer and Windle were among the first to carry outa systematic study of the mechanical properties ofnanotubepolymer composites Their technique for pre-paring MWNTPVA composites was described in theprevious section32 The tensile elastic moduli of thecomposite films were assessed in a dynamic mechanicalthermal analyser as a function of nanotube loading andtemperature The stiffness of the composites at roomtemperature was relatively low From the theory deve-loped for short-fibre composites a nanotube elasticmodulus of 150 MPa was obtained from the roomtemperature experimental data This value is well belowthe values reported for isolated nanotubes even defec-tive catalytically grown tubes have measured stiffness ofthe order of 30 GPa19 The low value may have more todo with poor stress transfer than with the weakness ofthe nanotubes themselves Above the glass transitiontemperature of the polymer (y85uC) the nanotubeshad a more significant effect on the properties of thecomposite

Cadek et al achieved rather better results in 2002with composites made from PVA and arc producedMWNTs56 The composites were prepared by solutionmixing and films were formed on glass substrates bydrop casting The presence of 1 wt- nanotubes increasedthe Youngrsquos modulus and hardness of the PVA byfactors of 18 and 16 respectively This improvementover the early results of Shaffer and Windle reflects boththe superior quality of the tubes and stronger interfacialbonding Evidence for the latter was provided by TEMstudies

The mechanical properties of nanotubepolystyrenecomposites were studied by Qian and co-workers3435

Two samples of catalytically produced MWNTs wereused one with an average length of y15 mm the otherwith an average length of y50 mm With the addition of1 wt- nanotubes they achieved a 36 and 42 increasein the elastic stiffness of the polymer for short and longtubes respectively (the neat polymer modulus wasy12 GPa) In both cases a 25 increase in the tensilestrength was observed Theoretical calculations carriedout assuming a nanotube modulus of 450 GPa pro-duced increases of a 48 and 62 in the stiffness of thepolymer for short and long tubes respectively The closeagreement within y10 between the experimental andtheoretically predicted composite moduli indicated thatthe external tensile loads were successfully transmittedto the nanotubes across the tubepolymer interface Athigher nanotube concentrations the changes in themechanical properties of the composites were morepronounced35 For MWNTpolystyrene composites con-taining from 25 to 25 vol- nanotubes Youngrsquosmodulus increased progressively from 19 to 45 GPawith the major increases occurring when the MWNTcontent was at or above about 10 vol- However thedependence of tensile strength on nanotube concentra-tion was more complex At the lower concentrations((10 vol-) the tensile strength decreased from theneat polymer value ofy40 MPa only exceeding it whenthe MWNT content was above 15 vol-

In an attempt to understand the tensile fracturemechanisms of the nanotubepolystyrene compositesQian et al performed deformation studies inside aTEM34 Their elegant method of preparing TEM speci-mens has been described above Focusing the electronbeam onto the thin film resulted in local thermal stresseswhich initiated cracks in the composite The propaga-tion speed of the crack could be controlled by varyingthe beam flux onto the sample These in situ TEMobservations showed that the cracks tended to nucleateat low nanotube density areas then propagate alongweak nanotubepolymer interfaces or relatively low nano-tube density regions The nanotubes became alignedperpendicular to the crack direction and bridged thecrack faces in the wake as shown in Fig 9 When thecrack opening displacement exceeded y800 nm thenanotubes began to break andor pull out of the matrixThe fact that some of the pulled-out tubes do not appearto be coated with polymer suggests that there is room forimprovement in nanotubendashpolymer bonding

Watts and Hsu have also used TEM to study thefracture behaviour of a nanotubepolymer composite57

Their composite consisted of arc grown MWNTsembedded in a diblock copolymer known as MPCndashDEA To prepare specimens of the composite for TEM

Harris Carbon nanotube composites

36 International Materials Reviews 2004 VOL 49 NO 1

the tubes were mixed with the polymer in an acidifiedaqueous solution and drops of this were applied to TEMgrids which were then dried in vacuum As in the workof Qian et al cracks were produced in the material byirradiation with an electron beam but a low acceleratingvoltage (75 kV) was used to avoid massive disruption ofthe thin composite film Individual tubes and bundles oftubes were observed to become stretched as the cracksgrew wider and pull-out was often seen In contrast tothe work of Qian et al however the tubes did notbreak This was probably because high quality arcgrown nanotubes were used rather than catalyticallyproduced tubes

The early work on nanotubeepoxy composites byAjayan et al has been described in the previoussection39 In 1998 Ajayan and co-workers described astudy of the mechanical behaviour of MWNTepoxycomposites in both tension and compression58 It wasfound that the compression modulus was higher thanthe tensile modulus indicating that load transfer to thenanotubes in the composite was much higher in com-pression The authors suggested that during load trans-fer to MWNTs only the outer layers are stressed intension whereas all the layers respond in compression

Nanotubeepoxy thin films prepared by Xu et alshowed excellent mechanical properties42 Compared toneat resin thin films for which the elastic modulus was42 GPa a 20 increase in modulus was seen when01 wt- MWNTs were added This was attributed topartial alignment of the MWNTs induced by spincoating Fracture behaviour of the films was investigatedby SEM and showed that pulled-out tubes were oftencovered with polymer suggesting strong interfacial adhe-sion Similar results were reported by Gojny et al59 Thisgroup showed that functionalising arc produced MWNTswith amine groups greatly enhanced bonding with theepoxy matrix As a result the outer shells of the tubesoften remained embedded in the matrix following pull-out as shown in Fig 10

All of the work described so far in this section hasinvolved multiwalled carbon nanotubes Ajayan et al

have described a study of SWNT-containing compo-sites60 Two types of composite were prepared first anSWNTepoxy composite and second a pressed pelletcontaining SWNTs and carbonaceous soot materialformed during nanotube synthesis The SWNTs wereused as-prepared ie generally grouped into bundlesrather than separated into individual tubes The com-posites were loaded axially in tension until failureoccurred and the fracture surfaces were then examinedin detail using SEM In some cases the tube bundleswere found to have been pulled out of the matrix duringthe deformation and fracture of the composites while inother cases the nanotubes were not entirely pulled outbut were stretched between two fracture surfaces Theauthors believed that during pull-out the nanotubeswere sliding axially within the ropes and that the failureobserved at large crack distances was not failure ofindividual tubes but the bundles pulling apart Studiesby micro-Raman spectroscopy supported this hypoth-esis It was found that little shift in the second order A1g

band occurred as a result of the application of axialtension showing that the individual nanotubes were notbeing significantly stretched To take full advantage ofthe high Youngrsquos modulus of SWNTs in polymer com-posites it is clear that load must be transferred effec-tively from the matrix to the nanotubes Ajayan et alsuggested that load transfer could be improved by break-ing bundles down into individual tube fragments beforedispersing in the matrix Alternatively the bundles them-selves could be reinforced by crosslinking the tubes withinbundles (eg by chemical treatments or irradiation)

Single-walled nanotubeepoxy composites have alsobeen prepared by Biercuk et al61 The Vickers hardnessof the polymer was found to increase monotonicallywith addition of SWNTs up to a factor of 35 at 2 wt-nanotube loading Greatly enhanced thermal conducti-vities were also observed Thus samples loaded with1 wt- unpurified SWNT material showed a 70increase in thermal conductivity at 40 K rising to125 at room temperature

The preparation of SWNTpolyelectrolyte compositesby the layer-by-layer method has been described in theprevious section49 The films produced in this way wereshown to have exceptional mechanical properties Thusthe average ultimate tensile strength of the films wasfound to be 220 MPa with some measurements as highas 325 MPa This is several orders of magnitude greater

10 Image of epoxy coated MWNT subject to tensile

stress as result of electron irradiation telescopic

pull-out of outer tube layer has occurred59

9 TEM image of crack in MWNTpolystyrene thin film

induced by thermal stresses34

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 37

than the tensile strength of strong industrial plastics andindicates that the layer-by-layer method has great pro-mise for the production of strong nanotube-containingcomposites

Electrical and optical properties ofcarbon nanotubepolymer compositesThe effect of adding nanotubes on the electrical pro-perties of polymers has been studied by many groupsThese studies fall into a number of distinct categories Insome cases the nanotubes have been used to increase theconductivity of relatively low cost polymers as analternative to currently used fillers such as carbon blackOther studies have involved the incorporation of nano-tubes into conducting polymers such as polyaniline Stillfurther work has been aimed at specific applicationssuch as photovoltaic devices or supercapacitors In manycases improved mechanical properties have been avaluable by-product of the inclusion of nanotubes Theaddition of nanotubes to low cost bulk polymers isconsidered first

Electrostatic dissipationImproving the electrical conductivity of bulk polymers isimportant in a number of applications For example insome aircraft components enhanced conductivity isrequired to provide electrostatic discharge and electro-magnetic radio frequency interference protection Staticelectrical dissipation is also needed in other applicationsincluding computer housings and exterior automotiveparts With these types of application in mind Sandleret al investigated the electrical properties of nanotubeepoxy composites40 Matrix resistivities of about 100 V mwith filler volume fractions as low as 01 vol- wereachieved These figures represented an advance on thebest conductivity values previously obtained with carbonblack in the same epoxy matrix

Lozano and co-workers prepared composites ofcatalytically grown carbon nanofibres in a polypropy-lene matrix5162 The nanofibres were purified and func-tionalised to facilitate bonding with the matrix and highshear processing was used to disperse the fibres in thepolymer matrix Rheological and microscopic analysisshowed that a homogeneous dispersion of nanofibreswas formed and a percolation threshold for electricalconduction of 9ndash18 wt- was observed

These and other published studies suggest that carbonnanotubes have great promise in reducing electrostaticcharging of bulk polymers Nanotubes also have otheradvantages over conventional fillers such as carbonblack and carbon fibres in that they are more amenableto processing and can be more easily dispersed through-out the matrix In addition to the studies available in theopen literature there is undoubtedly a great deal ofcommercial work being carried out on the use of nano-tubes to reduce the electrostatic charging of plastics Infact nanotube containing plastics are already being usedin commercial products63 These include fuel lines inautomobiles where the nanotubes help to dissipate anydangerous charge which may build up Thermoplasticpolymers containing nanotubes are also used in someexterior automobile parts so that they can be earthedduring electrostatic painting

Conducting polymer compositesPolyaniline is a conducting polymer with many attrac-tive characteristics such as processability and environ-mental stability There has been interest in addingnanotubes to this polymer to provide both enhancedconductivity and improved mechanical properties Asmentioned above in the section lsquoPreparation of carbonnanotubepolymer compositesrsquo Cochet et al have pre-pared a MWNTpolyaniline composite by in situ poly-merisation44 The nanotubes were found to have astrong influence on the transport properties of thepolymer Thus the room temperature resistivity of thecomposite was found to be an order of magnitude lowerthan that of pure polyaniline while the low temperatureresistivity was much smaller than that of either poly-aniline or of MWNTs The temperature dependence ofresistivity was also weaker than that of polyanilinealone These observations were explained by assumingthat in situ polymerisation favours charge transfer betweenpolyaniline and MWNTs resulting in an overall materialwhich is more conducting than the starting componentsRaman studies indicated that these charge transferprocesses involved a site-selective interaction betweenthe quinoid ring of the polymer and the MWNTs Othergroups have also demonstrated excellent conductivity inMWNTpolyaniline composites6465 while a group fromDuPont have shown that SWNTpolyaniline compositescan be used as printable conductors for organic elec-tronics devices66

Optoelectronic propertiesSeveral groups have prepared nanotube compositesusing the conjugated polymer poly(p-phenylene viny-lene) (PPV) and its derivatives Polymers of this classhave electroluminescent properties and are widely usedin light-emitting diodes They are also used in photo-voltaic devices In an early study workers from Irelandand the USA described the synthesis of a compositematerial containing arc produced MWNTs in a PPVderivative PmPV67 The composite was prepared usinga solution mixing method and good wetting of the tubesby the polymer was achieved One of the aims of addingnanotubes to the polymer was to increase its electricalconductivity Previous attempts to achieve this by dopinghad resulted in degradation of the optical properties ofthe polymer In the case of nanotubes an increase in theelectrical conductivity of the polymer by up to eightorders of magnitude was seen with no concomitantdegradation of optical properties This was apparentlybecause the tubes acted as nanometric heat sinks pre-venting the build up of large thermal effects whichdamage these conjugated systems It was also shown inthis work that the composite could be used as theemissive layer in an organic light-emitting diode (LED)

As mentioned above several groups have been inter-ested in using nanotubepolymer composites in photo-voltaic devices The photovoltaic effect involves thegeneration of electron and hole pairs and their sub-sequent collection at opposite electrodes In organicmaterials the photon absorption excites an electron intothe conduction band which leads to a generation ofbound electronndashhole pairs or excitons These excitonshave to dissociate into free charges in order to be trans-ported to the electrodes Exciton dissociation can occurat a junction between a polymer and an electron acceptor

Harris Carbon nanotube composites

38 International Materials Reviews 2004 VOL 49 NO 1

molecule This allows the preferential transfer of the elec-trons into the electron acceptor molecule while leavingthe holes to be preferentially transported through thepolymer a process known as photoinduced chargetransfer In 1992 photoinduced charge transfer from aconducting polymer to C60 was discovered68 and thishas led to the fabrication of several efficient photo-voltaic systems using fullerenepolymer combinations69

Photovoltaic devices based on nanotubepolymer com-binations have now been fabricated by several groupsand this work will now be summarised

The earliest attempt to fabricate a nanotubepolymerphotovoltaic device appears to be the 1996 work ofRomero et al70 In this work a layer of MWNTs wasdeposited on Teflon and a PPV derivative was cast ontothis with a thin aluminium layer on top The devicedisplayed clear diode behaviour but only a small photo-current was observed when the junction was irradiatedwith light from a HendashNe laser This may have been dueto the relatively thick PPV active layer Ago et alfabricated a nanotube containing photovoltaic deviceby depositing a layer of PPV onto a glass-supportedMWNT film71 The interaction between MWNTs andPPV was investigated by means of photoluminescence(PL) measurements The photoluminescence efficiencyof the PPV was found to be markedly reduced by thepresence of the MWNT layer Since photoluminescenceis a result of radiative recombination a reduction mayimply improved exciton separation which is what onewishes to achieve in a photovoltaic device The devicefabricated by Ago et al is illustrated in Fig 11a Thisshowed a clear diode characteristic in the dark andupon illumination through the aluminium electrode aphotocurrent was observed A quantum efficiency approxi-mately twice that of the standard indiumndashtin oxidedevice was reported

Kymakis and Amaratunga used a slightly differentapproach to prepare a nanotubeconjugated polymerphotovoltaic device72 In this work composite SWNTpoly(3-octylthiophene) films were drop or spin cast froma solution onto indiumndashtin oxide (ITO) and quartz sub-strates to produce the arrangement shown in Fig 11bDevices were also prepared using the pure P3OT poly-mer A considerable enhancement of the photovoltaic

effect was observed with the P3OTSWNTs blend devicecompared with the pure polymer device

Non-linear optical propertiesAs mentioned above in the section lsquoStructure and pro-perties of carbon nanotubesrsquo it has been demonstratedthat suspensions of both SWNTs and MWNTs havenon-linear optical properties Nanotubepolymer com-posites may therefore have applications in optical devicesTo date however there has been little work on theoptical limiting properties of nanotube containing com-posites In one of the first studies in this area Chen et aldemonstrated third order optical non-linearity in a SWNTpolyimide composite73 Recently OrsquoFlaherty et al pre-pared a composite in which MWNTs were dispersed inthe polymer poly(99-di-n-octylfluorenyl-279-diyl) (PFO)74

Suspensions of the nanotubePFO composite in toluenewere prepared and the optical limiting properties of thesolutions were measured using nanosecond laser pulsesMaximum optical limiting was observed at carbonnanotube loadings in excess of 38 wt- relative to thepolymer

Carbon nanotubeceramic compositesCeramics have high stiffnesses and thermal stabilities butrelatively low breaking strengths Incorporating carbonnanotubes into a ceramic matrix might be expected toproduce a composite with both toughness and high tem-perature stability However achieving a homogeneousdispersion of tubes in an oxide with strong bondingbetween tubes and matrix presents rather more of achallenge than incorporating tubes into a polymer Variousapproaches have been adopted for incorporating nano-tubes into oxides such as alumina and silica and thesewill now be summarised

Peigney and co-workers have been among the pioneersin the synthesis of nanotubeoxide composites75ndash77 Theyhave developed an ingenious technique which involvesthe use of catalysts such as iron-containing a-aluminaPassing a CH4ndashH2 mixture over the catalysts resultedin the in situ growth of nanotubes on iron particlesproducing the precursor for a nanotubeoxide compo-site The resulting powders can then be hot pressed toform the final composite material The mechanicalproperties of these composites were initially disappoint-ing72 Thus the fracture strengths of the nanotubeFendashAl2O3 composites were only marginally higher than thatof Al2O3 and were generally markedly lower than thoseof the carbon-free FendashAl2O3 composites The fracturetoughness values were lower than or similar to that ofAl2O3 However SEM observations of composite frac-tures indicate that the nanotube bundles which are veryflexible could dissipate some fracture energy A similarmethod was used by An and Lim to prepare nanotubeAl2O3 composites in this case using C2H2 as the carbonsource78 Nanotube contents of up to 125 wt- wereachieved and a significant reinforcement effect wasreported

Achieving alignment of nanotubes in a ceramic matrixis much more difficult than in a polymer since ceramicsdo not lend themselves so easily to processing Howevera number of groups have shown that aligned nanotubemetal oxide composites can indeed be produced Peigney

ITO5indiumndashtin oxide PPV5poly(p-phenylene vinylene)

P3OT5poly(3-octylthiophene)

11 a Illustration of photovoltaic device fabricated by Ago

et al71 and b photovoltaic device fabricated by

Kymakis and Amaratunga72

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 39

et al exploited the fact that when their grains are main-tained at sufficiently low sizes (less than about 1 mm)many polycrystalline ceramics become superplasticallowing the use of thermomechanical processing forthe forming of these materials A series of cobalt andiron containing oxide catalysts were prepared77 Theprimary grain size in the powders was smaller than100 nm but in some of them the as produced primarygrains were strongly aggregated and required ball millingin order to reduce aggregate size to less than 1 mmNanotubes were then prepared in situ as describedabove and the powders were then extruded at hightemperatures in a graphite die under vacuum The bestresults were achieved with nanotubes incorporated intoFendashAl2O3 and FeCondashMgAl2O4 matrices In both casesthe nanotubes were not damaged by the processing andalignment of the nanotubes was clearly observed Infact the presence of the nanotubes seemed to facilitatethe superplastic forming by inhibiting grain growth andacting as lubricating agents It was demonstrated thatthe aligned nanotubemetal oxide composites showed ananisotropy of electrical conductivity which could beadjusted by controlling the nanotube content

One of the first studies of a SWNTceramic compositewas described by Zhan et al79 This group used thetechnique of spark plasma sintering (SPS) to prepareAl2O3 composites containing 10 vol- SWNTs Theadvantage of spark plasma sintering is that it allowsceramic powders to be annealed at lower temperaturesand for much shorter times than other sintering pro-cesses leading to the fabrication of fully dense ceramicsor composites with nanocrystalline microstructures undermild conditions Unlike some of the other processes usedto prepare nanotubeoxide composites spark plasmasintering does not damage the nanotubes and the nano-tube bundles could be seen to be located mainly at theboundaries of the Al2O3 grains in good contact with theAl2O3 matrix Before sintering the powders had beenmixed by ball milling which produced a reasonablyhomogeneous dispersion again without damaging thenanotubes The fracture toughnesses of the final com-posites were remarkable ndash about twice that of the fullydensified but unreinforced Al2O3 matrix In contrast toprevious work on nanotubeceramic composites thefracture toughness was found to increase with nanotubedensity This is thought to be due to the formation ofentangled networks of single-walled carbon nanotubeswhich may inhibit crack propagation A subsequentstudy of the nanotubeAl2O3 composites showed thatthey had high electrical conductivities80

Various techniques have been used to prepare carbonnanotubesilica composites Seeger et al described amethod which involved preparing a composite gel ofMWNTs with tetraethoxysilane (TEOS) and then sinter-ing this at 1150uC in argon81 A drawback of this tech-nique was that the sintering process led to a partialcrystallisation of the SiO2 resulting in a very inhomo-geneous matrix An alternative method which avoidsthese problems has been developed82 This involvesusing a Nd YAG laser to rapidly heat a TEOSnano-tube mixture resulting in partial melting of the matrixThis produced an amorphous silica matrix with nocrystallisation

A novel nanotubesilica composite material withpotentially useful optical properties has recently been

described by Han et al83 These workers synthesisedfilms consisting of silica spheres with diameters rangingfrom 200 to 650 nm Molybdenumcobalt catalyst parti-cles were then deposited onto the silica spheres and usedto catalyse the growth of SWNTs Since SWNTs arehighly non-linear and fast-switching materials they canbe incorporated into an optically confining environmentto achieve these characteristics at relatively low levels oflaser intensities

There have been a few studies on preparing nanotubecomposites using oxides other than alumina and silicaVincent et al reported the production of nanotubeTiO2

composites using solndashgel methods84 Possible applica-tions of the composites included optical non-linear waveguides and unidimensional conductive films Solndashgelmethods were also used by Sakamoto and Dunn toprepare SWNTV2O5 composites for use as electrodesin secondary lithium batteries85 Little work has beencarried out so far on incorporating nanotubes into non-oxide ceramics In one of the few published studies Maet al fabricated nanotubesilicon carbide composites bymixing SiC nanoparticles with 10 wt- MWNTs andhot pressing at 2000uC86 They found that the nanotubesplayed a strengthening and toughening role in thecomposite both bend strength and fracture toughnessincreased by about 10 compared to monolithic SiCThere have also been some initial studies on carbonnanotubecarbon fibre composites8788

Carbon nanotubemetal compositesComposite materials containing conventional carbonfibres in a metal matrix such as aluminium or mag-nesium are used in a number of specialist applica-tions8990 Such composites combine low density withhigh strength and modulus making them particularlyattractive to the aerospace industry There is growinginterest in the addition of carbon nanotubes to metalmatrices although only a small number of studies havebeen published to date Kuzumaki et al described thepreparation of a nanotubealuminium composite in 199891

Their method involved mixing a nanotube sample with afine aluminium powder mounting the mixture in a6 mm silver sheath and then drawing and heating thewire at 700uC in a vacuum furnace The result was acomposite wire in which the nanotubes were partiallyaligned along the axial direction The tensile strengths ofthe as prepared composite wires were comparable to thatof pure aluminium but the composite wires retained thisstrength after prolonged annealing at 600uC while thestrength of pure aluminium decreased by about 50after this treatment The same group subsequently pre-pared nanotubetitanium composites which showed alarge increase in hardness and modulus compared topure titanium92

Chen et al prepared nanotubenickelphosphorus com-posite coatings on stainless steel substrates using anelectroless plating method93 The coatings exhibited higherwear resistance and lower friction coefficients than SiCNiP and graphiteNiP composite coatings The samegroup prepared nanotubecopper composites by a powdermetallurgy technique and showed that the presence ofthe nanotubes again resulted in improved wear rates andfriction coefficients These results suggest that nanotube

Harris Carbon nanotube composites

40 International Materials Reviews 2004 VOL 49 NO 1

metal composites may have potential in tribologicalapplications

DiscussionThe subject of carbon nanotube composites has grownrapidly over the past 5 years or so into a vibrant andexciting area of research It is important to recognisehowever that this is still a new subject and the workcarried out so far may have hardly scratched the surfaceof what is possible For example the great majority ofthe studies that have been published to date haveinvolved the use of polymer matrices there is great scopefor more work on incorporating nanotubes into cera-mics metals and other forms of carbon Also most ofthe work up to now has been aimed at exploiting themechanical properties of carbon nanotubes rather thantheir electronic or optical properties although interest inthose is growing

When considering the mechanical performance ofcarbon nanotube composites it is clear that these dependstrongly on whether arc grown tubes or catalyticallyproduced tubes have been used In considering thecharacteristics of the two types of tube it is instructive tocompare the work of Qian et al34 with that of Watts andHsu57 Both incorporated the nanotubes into polymersand then used TEM to characterise the composites andstudy their fracture behaviour Qian et al used cataly-tically produced tubes and these were found to bewell dispersed throughout the polymer matrix with noclustering together of tubes However during compositefracture the tubes were often observed to break sug-gesting that they have relatively poor strength Wattsand Hsu on the other hand employed much moreperfect nanotubes produced by arc evaporation andTEM studies revealed superior mechanical propertieswith no observed breaking However the nanotubestended to cluster together so that during crack growththe tubes could quite easily slide past one another Thuswhile it is clear that the best mechanical performancewill be achieved with arc grown nanotubes it is impor-tant that the tubes be well dispersed throughout thematrix in order fully to exploit their properties

Equally important is achieving good bonding betweennanotubes and matrix Many studies have shown that thisinterfacial bonding can be enhanced by functionalising thenanotubes before embedding them in the matrix Theeffect of this can be seen in Fig 10 from the work ofGojny et al59 where bonding between MWNTs and anepoxy matrix is so strong that the outer shells of the tuberemain embedded in the matrix following pull-out Apossible disadvantage of functionalisation is that it couldweaken the nanotubes by disrupting the perfect cylindricalstructure This might be expected to be a particularproblem with single-walled tubes However functionali-sation of the nanotubes may not be necessary to producebonding with a polymer matrix Liao and Li havemodelled the interfacial characteristics of a nanotubepolystyrene composite assuming no covalent bondingbetween the tubes and the matrix94 It was found thatelectrostatic and van der Waals interactions contributedsignificantly to the interfacial stress transfer A nanotubepull-out simulation suggested that the interfacial shearstress of the nanotubepolystyrene system is about160 MPa significantly higher than for most carbon fibrereinforced polymer composite systems It is also worth

mentioning recent work by Star and co-workers9596

in which it was demonstrated that poly(5-alkoxy-m-phenylenevinylene)-co-[(25-dioctyloxy-p-phenylene)vinylene] (PAmPV) derivatives could be induced towrap around SWNTs forming strong non-covalentbonds

The quality of the nanotubes and the nature of theinterfacial bonding also has a strong influence on theelectrical properties of nanotubepolymer compositesArc grown tubes can have resistivities lower than that ofsingle crystal graphite but as already discussed theirelectronic properties depend on their precise structureAt present we do not have a reliable method of separat-ing metallic and semiconducting tubes so it is notpossible to exploit their full potential Catalytically pro-duced tubes are likely to be poorer conductors thanmetallic arc-grown tubes owing to their relatively dis-ordered structure Nevertheless catalytically grown tubesare sufficiently conducting to be useful in some com-mercial applications such as electrically conductive paintsThe influence of interfacial bonding on electronic pro-perties was demonstrated in the work of Cochet et al44

These workers prepared a MWNTpolyaniline com-posite by in situ polymerisation and found it to haveexcellent electrical transport properties This was attrib-uted to strong interaction between tubes and polymer

As has been shown in this review the possible appli-cations of carbon nanotube composites range widelyfrom electrically conductive paints to optical devicesOnly a relatively small number of these new materialshave been used commercially however and there arestill many problems to overcome before the full potentialof nanotube containing composites can be realised Notleast of these problems is the high cost of good qualitycarbon nanotubes Currently the commercial price ofunpurified arc-grownMWNTs is approximately $10000gwhile purified MWNTs cost around $30000g Cataly-tically grown MWNTs are much cheaper but as empha-sised throughout this review have generally inferiorproperties As far as SWNTs are concerned unpurifiedsamples cost around $5000g while purified SWNTs are$50000g As a comparison the price of gold is cur-rently about $1200g The development of new techni-ques for the low cost synthesis of high quality nanotubesis clearly essential This might be achieved throughimproved catalytic processes and there is currently alarge amount of work being carried out towards thisend Alternatively it may be possible to develop newnon-catalytic techniques for the synthesis of fullerene-related nanotubes In this connection it is interesting tonote that nanotubes can be produced by the high tem-perature heat treatment of conventional carbons such asmicroporous non-graphitising carbons9798 or carbonblack99100 Further work in this area would be of value

Acknowledgement

I thank Milo Shaffer for helpful comments

References1 P J F Harris lsquoCarbon nanotubes and related structures ndash new

materials for the twenty-first centuryrsquo 1999 Cambridge

Cambridge University Press

2 S Iijima Nature 1991 354 56ndash58

3 W Kratschmer L D Lamb K Fostiropoulos and D R

Huffman Nature 1990 347 354ndash358

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 41

4 S Iijima and T Ichihashi Nature 1993 363 603ndash605

5 D S Bethune C H Kiang M S De Vries G Gorman R Savoy

J Vasquez and R Beyers Nature 1993 363 605ndash607

6 H Dai A G Rinzler P Nikolaev A Thess D T Colbert and

R E Smalley Chem Phys Lett 1996 260 471ndash475

7 M M J Treacy T W Ebbesen and J M Gibson Nature 1996

381 678ndash680

8 E W Wong P E Sheehan and C M Lieber Science 1997 277

1971ndash1975

9 M R Falvo G J Clary R M Taylor V Chi F P Brooks

S Washburn and R Superfine Nature 1997 389 582ndash584

10 K Anazawa K Shimotani C Manabe H Watanabe and M

Shimizu Appl Phys Lett 2002 81 739ndash741

11 P Kim L Shi A Majumdar and P L McEuen Phys Rev Lett

2001 87 paper 215502

12 E T Thostenson Z F Ren and T W Chou Compos Sci

Technol 2001 61 1899ndash1912

13 K T Lau and D Hui Composites Part B Engineering 2002 33

263ndash277

14 Y Huh J Y Lee J Cheon Y K Hong J Y Koo T J Lee and

C J Lee J Mater Chem 2003 13 2297ndash2300

15 M Endo T Koyama and Y Hishiyama Jpn J Appl Phys 1976

15 2073ndash2076

16 S C Lyu B C Liu T J Lee Z Y Liu C W Yang C Y Park

and C J Lee Chem Commun 2003 734ndash735

17 M-F Yu O Lourie M J Dyer K Moloni T F Kelly and R S

Ruoff Science 2000 287 637ndash640

18 P Poncharal Z L Wang D Ugarte and W A De Heer Science

1999 283 1513ndash1516

19 J-P Salvetat A J Kulik J-M Bonard G A D Briggs

T Stockli K Metenier S Bonnamy F Beguin N A Burnham

and L Forro Adv Mater 1999 11 161ndash165

20 M-F Yu B S Files S Arepalli and R S Ruoff Phys Rev Lett

2000 84 5552ndash5555

21 M Terrones Int Mater Rev to be published

22 J W Mintmire B I Dunlap and C T White Phys Rev Lett

1992 68 631ndash634

23 R Saito M Fujita G Dresselhaus and M S Dresselhaus Phys

Rev B 1992 46 1804ndash1811

24 N Hamada S Sawada and A Oshiyama Phys Rev Lett 1992

68 1579ndash1581

25 T W Ebbesen H J Lezec H Hiura J W Bennett H F Ghaemi

and T Thio Nature 1996 382 54ndash56

26 J E Riggs D B Walker D L Carroll and Y-P Sun J Phys

Chem B 2000 104 7071ndash7076

27 S C Tsang Y K Chen P J F Harris and M L H Green

Nature 1994 372 159ndash162

28 R M Lago S C Tsang K L Lu Y K Chen and M L H

Green Chem Commun 1995 1355ndash1356

29 Y-P Sun K F Fu Y Lin and W Huang Acc Chem Res 2002

35 1096ndash1104

30 S B Sinnott J Nanosci Nanotechnol 2002 2113ndash123

31 M S P Shaffer X Fan and A H Windle Carbon 1998 36

1603ndash1612

32 M S P Shaffer and A H Windle Adv Mater 1999 11 937ndash941

33 D E Hill Y Lin A M Rao L F Allard and Y-P Sun

Macromolecules 2002 35 9466ndash9471

34 D Qian E C Dickey R Andrews and T Rantell Appl Phys

Lett 2000 76 2868ndash2870

35 R Andrews D Jacques D Qian and T Rantell Acc Chem Res

2002 35 1008ndash1017

36 R Haggenmueller H H Gommans A G Rinzler J E Fischer

and K I Winey Chem Phys Lett 2000 330 219ndash225

37 P Potschke T D Fornes and D R Paul Polymer 2002 43 3247ndash

3255

38 M Sennett E Welsh J B Wright W Z Li J G Wen and Z F

Ren Appl Phys A 2003 76 111ndash113

39 P M Ajayan O Stephan C Colliex and D Trauth Science 1994

265 1212ndash1214

40 J Sandler M S P Shaffer T Prasse W Bauhofer K Schulte

and H Windle Polymer 1999 40 5967ndash5971

41 A Allaoui S Bai H M Cheng and J B Bai Compos Sci

Technol 2002 62 1993ndash1998

42 X Xu M M Thwe C Shearwood and K Liao Appl Phys Lett

2002 81 2833ndash2835

43 D Puglia L Valentini and J M Kenny J Appl Polymer Sci

2003 88 452ndash458

44 M Cochet W K Maser A M Benito M A Callejas M T

Martınez J-M Benoit J Schreiber and O Chauvet Chem

Commun 2001 1450ndash1451

45 M S P Shaffer and K Koziol Chem Commun 2002 2074ndash

2075

46 C Park Z Ounaies K A Watson R E Crooks J Smith S E

Lowther J W Connell E J Siochi J S Harrison and T L St

Clair Chem Phys Lett 2002 364 303ndash308

47 K Jurewicz S Delpeux V Bertagna F Beguin and E

Frackowiak Chem Phys Lett 2001 347 36ndash40

48 M Hughes G Z Chen M S P Shaffer D J Fray and A H

Windle Chem Mater 2002 14 1610ndash1613

49 A A Mamedov N A Kotov M Prato D M Guldi J P

Wicksted and A Hirsch Nature Mater 2002 1 190ndash194

50 J H Rouse and P T Lillehei Nano Lett 2003 3 59ndash62

51 K Lozano and E V Barrera J Appl Polym Sci 2001 79 125ndash

133

52 L Valentini J Biagiotti J M Kenny and S Santucci Compos

Sci Technol 2003 63 1149ndash1153

53 J Sandler P Werner M S P Shaffer V Demchuk V Altstadt

and A H Windle Composites Part A Appl Sci Manuf 2002 33

1033ndash1039

54 R Czerw Z X Guo P M Ajayan Y P Sun and D L Carroll

Nano Lett 2001 1 423ndash427

55 B P Grady F Pompeo R L Shambaugh and D E Resasco

J Phys Chem B 2002 106 5852ndash5858

56 M Cadek J N Coleman V Barron K Hedicke and W J Blau

Appl Phys Lett 2002 81 5123ndash5125

57 P C P Watts and W K Hsu Nanotechnology 2003 14 L7ndashL10

58 L S Schadler S C Giannaris and P M Ajayan Appl Phys Lett

1998 73 3842ndash3844

59 F H Gojny J Nastalczyk Z Roslaniec and K Schulte Chem

Phys Lett 2003 370 820ndash824

60 P M Ajayan L S Schadler C Giannaris and A Rubio Adv

Mater 2000 12 750ndash753

61 M J Biercuk M C Llaguno M Radosavljevic J K Hyun A T

Johnson and J E Fischer Appl Phys Lett 2002 80 2767ndash2769

62 K Lozano J Bonilla-Rios and E V Barrera J Appl Polym Sci

2001 80 1162ndash1172

63 V Jamieson New Sci 15 Mar 2003 30

64 H Zengin W S Zhou J Y Jin W R Czerw D W Smith

L Echegoyen D L Carroll S H Foulger and J Ballato Adv

Mater 2002 14 1480ndash1483

65 J Deng X Ding W Zhang Y Peng J Wang X Long P Li and

A S C Chan Eur Polym J 2002 38 2497ndash2501

66 G B Blanchet C R Fincher and F Gao Appl Phys Lett 2003

82 1290ndash1292

67 S A Curran P M Ajayan W J Blau D L Carroll J N

Coleman A B Dalton A P Davey A Drury B McCarthy

S Maier and A A Strevens Adv Mater 1998 10 1091ndash1093

68 N S Sariciftci L Smilowitz A J Heeger and F Wudl Science

1992 258 1474ndash1476

69 N Camaioni L Garlaschelli A Geri M Maggini G Possamai

and G Ridolfi J Mater Chem 2002 12 2065ndash2070

70 D B Romero M Carrard W De Heer and L Zuppiroli

Adv Mater 1996 8 899ndash902

71 H Ago K Petritsch M S P Shaffer A H Windle and R H

Friend Adv Mater 1999 11 1281ndash1285

72 E Kymakis and G A J Amaratunga Appl Phys Lett 2002 80

112ndash114

73 Y-C Chen N R Raravikar L S Schadler P M Ajayan Y-P

Zhao T-M Lu G-C Wang and X-C Zhang Appl Phys Lett

2002 81 975ndash977

74 S A OrsquoFlaherty R Murphy S V Hold M Cadek J N

Coleman and W J Blau J Phys Chem B 2003 107 958ndash964

75 A Peigney C Laurent F Dobigeon and A Rousset J Mater

Res 1997 12 613ndash615

76 E Flahaut A Peigney C Laurent C Marliere F Chastel and

A Rousset Acta Mater 2000 48 3803ndash3812

77 A Peigney E Flahaut C Laurent F Chastel and A Rousset

Chem Phys Lett 2002 352 20ndash25

78 J W An and D S Lim J Ceram Process Res 2002 3 201ndash204

79 G-D Zhan J D Kuntz J L Wan and A K Mukherjee Nature

Mater 2003 2 38ndash42

80 G-D Zhan J D Kuntz J E Garay and A K Mukherjee Appl

Phys Lett 2003 83 1228ndash1230

81 T Seeger T Kohler T Frauenheim N Grobert M Ruhle

M Terrones and G Seifert Chem Commun 2002 34ndash35

82 T Seeger G De La Fuente W K Maser A M Benito M A

Callejas and M T Martınez Nanotechnology 2003 14 184ndash187

83 H Han S Vijayalakshmi A Lan Z Iqbal H Grebel E Lalanne

and A M Johnson Appl Phys Lett 2003 82 1458ndash1460

Harris Carbon nanotube composites

42 International Materials Reviews 2004 VOL 49 NO 1

84 P Vincent A Brioude C Journet S Rabaste S T Purcell J Le

Brusq and J C Plenet J Non-Cryst Sol 2002 311 130ndash137

85 J S Sakamoto and B Dunn J Electrochem Soc 2002 149 A26ndashA30

86 R Z Ma J Wu B Q Wei J Liang and D H Wu J Mater Sci

1998 33 5243ndash5246

87 R Andrews D Jacques A M Rao T Rantell F Derbyshire

Y Chen J Chen and R C Haddon Appl Phys Lett 1999 75

1329ndash1331

88 E T Thostenson W Z Li D Z Wang Z F Ren and T W

Chou J Appl Phys 2002 91 6034ndash6037

89 D D L Chung lsquoCarbon fiber compositesrsquo chap 7 1994 Boston

Butterworth-Heinemann

90 S Rawal JOM 2001 53 14ndash17

91 T Kuzumaki K Miyazawa H Ichinose and K Ito J Mater

Res 1998 13 2445ndash2449

92 T Kuzumaki O Ujiie H Ichinose and K Ito Adv Eng Mater

2000 2 416ndash418

93 W X Chen J P Tu L Y Wang H Y Gan Z D Xu and X B

Zhang Carbon 2003 41 215ndash222

94 K Liao and S Li Appl Phys Lett 2001 79 4225ndash4227

95 A Star J F Stoddart D Steuerman M Diehl A Boukai E W

Wong X Yang S W Chung H Choi and J R Heath

Angewandte Chemie-Int Ed 2001 40 1721ndash1725

96 A Star Y Liu K Grant L Ridvan J F Stoddart D W

Steuerman M R Diehl A Boukai and J R Heath

Macromolecules 2003 36 553ndash560

97 P J F Harris S C Tsang J B Claridge and M L H Green

J Chem Soc Faraday Trans 1994 90 2799ndash2802

98 A A Setlur S P Doherty J Y Dai and R P H Chang Appl

Phys Lett 2000 76 3008ndash3010

99 S P Doherty D B Buchholz B J Li and R P H Chang

J Mater Res 2003 18 941ndash949

100 D B Buchholz S P Doherty and R P H Chang Carbon 2003

41 1625ndash1634

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 43

Page 6: Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P. J. F. Harris * Carbon nanotubes are molecular-scale tubes of graphitic carbon with

50 wt- could be obtained Coating was carried out byalternate dipping of a glass slide or silicon wafer intodispersions of SWNTs (stabilised by acid treatment) andpolymer solutions The layers were held together by vander Waals forces and by electrostatic attraction betweenthe negatively charged SWNTs and a positively chargedpolyelectrolyte such as branched polyethyleneimine (PEI)Coatings containing up to 50 SWNTPEI bilayers couldbe built up in this way When the procedure was com-plete the multilayer films were heated to 120uC topromote crosslinking The films could then be lifted offthe substrate to obtain uniform free-standing mem-branes A rather similar method was used to synthesisemultilayer composites films of the polyelectrolyte poly-diallyldimethylammonium chloride and SWNTs50

Before proceeding to discuss the properties of carbonnanotubepolymer composites in detail it is worth brieflyconsidering the effect that the inclusion of nanotubescan have on polymer microstructure Although thistopic has not yet been widely studied a number ofgroups have demonstrated that the crystallisation andmorphology of the polymers can be strongly affected bysmall additions of nanotubes For example Lozano andBarrera51 showed that the addition of catalytically growncarbon nanofibres can promote nucleation of polypro-pylene Similar observations were made by Valentiniet al52 The crystallinity of a polymers has also beenshown to increase with the addition of carbon nano-fibres5354 Changes in the crystal structure of polymerscan also be induced by nanotubes55 More work on theseeffects would be welcome

Mechanical properties of carbonnanotubepolymer compositesMuch of the work on the preparation of nanotubepolymer composites has been driven by a desire toexploit the tubesrsquo stiffness and strength Even where theinterest has been focused on other properties the abilityof nanotubes to improve the mechanical characteristicsof a polymer has often been a valuable added benefit Inearly work the measured properties were often dis-appointing but such studies were of value in under-standing the reasons for composite failure and inidentifying the critical issues which need to be addressed

Shaffer and Windle were among the first to carry outa systematic study of the mechanical properties ofnanotubepolymer composites Their technique for pre-paring MWNTPVA composites was described in theprevious section32 The tensile elastic moduli of thecomposite films were assessed in a dynamic mechanicalthermal analyser as a function of nanotube loading andtemperature The stiffness of the composites at roomtemperature was relatively low From the theory deve-loped for short-fibre composites a nanotube elasticmodulus of 150 MPa was obtained from the roomtemperature experimental data This value is well belowthe values reported for isolated nanotubes even defec-tive catalytically grown tubes have measured stiffness ofthe order of 30 GPa19 The low value may have more todo with poor stress transfer than with the weakness ofthe nanotubes themselves Above the glass transitiontemperature of the polymer (y85uC) the nanotubeshad a more significant effect on the properties of thecomposite

Cadek et al achieved rather better results in 2002with composites made from PVA and arc producedMWNTs56 The composites were prepared by solutionmixing and films were formed on glass substrates bydrop casting The presence of 1 wt- nanotubes increasedthe Youngrsquos modulus and hardness of the PVA byfactors of 18 and 16 respectively This improvementover the early results of Shaffer and Windle reflects boththe superior quality of the tubes and stronger interfacialbonding Evidence for the latter was provided by TEMstudies

The mechanical properties of nanotubepolystyrenecomposites were studied by Qian and co-workers3435

Two samples of catalytically produced MWNTs wereused one with an average length of y15 mm the otherwith an average length of y50 mm With the addition of1 wt- nanotubes they achieved a 36 and 42 increasein the elastic stiffness of the polymer for short and longtubes respectively (the neat polymer modulus wasy12 GPa) In both cases a 25 increase in the tensilestrength was observed Theoretical calculations carriedout assuming a nanotube modulus of 450 GPa pro-duced increases of a 48 and 62 in the stiffness of thepolymer for short and long tubes respectively The closeagreement within y10 between the experimental andtheoretically predicted composite moduli indicated thatthe external tensile loads were successfully transmittedto the nanotubes across the tubepolymer interface Athigher nanotube concentrations the changes in themechanical properties of the composites were morepronounced35 For MWNTpolystyrene composites con-taining from 25 to 25 vol- nanotubes Youngrsquosmodulus increased progressively from 19 to 45 GPawith the major increases occurring when the MWNTcontent was at or above about 10 vol- However thedependence of tensile strength on nanotube concentra-tion was more complex At the lower concentrations((10 vol-) the tensile strength decreased from theneat polymer value ofy40 MPa only exceeding it whenthe MWNT content was above 15 vol-

In an attempt to understand the tensile fracturemechanisms of the nanotubepolystyrene compositesQian et al performed deformation studies inside aTEM34 Their elegant method of preparing TEM speci-mens has been described above Focusing the electronbeam onto the thin film resulted in local thermal stresseswhich initiated cracks in the composite The propaga-tion speed of the crack could be controlled by varyingthe beam flux onto the sample These in situ TEMobservations showed that the cracks tended to nucleateat low nanotube density areas then propagate alongweak nanotubepolymer interfaces or relatively low nano-tube density regions The nanotubes became alignedperpendicular to the crack direction and bridged thecrack faces in the wake as shown in Fig 9 When thecrack opening displacement exceeded y800 nm thenanotubes began to break andor pull out of the matrixThe fact that some of the pulled-out tubes do not appearto be coated with polymer suggests that there is room forimprovement in nanotubendashpolymer bonding

Watts and Hsu have also used TEM to study thefracture behaviour of a nanotubepolymer composite57

Their composite consisted of arc grown MWNTsembedded in a diblock copolymer known as MPCndashDEA To prepare specimens of the composite for TEM

Harris Carbon nanotube composites

36 International Materials Reviews 2004 VOL 49 NO 1

the tubes were mixed with the polymer in an acidifiedaqueous solution and drops of this were applied to TEMgrids which were then dried in vacuum As in the workof Qian et al cracks were produced in the material byirradiation with an electron beam but a low acceleratingvoltage (75 kV) was used to avoid massive disruption ofthe thin composite film Individual tubes and bundles oftubes were observed to become stretched as the cracksgrew wider and pull-out was often seen In contrast tothe work of Qian et al however the tubes did notbreak This was probably because high quality arcgrown nanotubes were used rather than catalyticallyproduced tubes

The early work on nanotubeepoxy composites byAjayan et al has been described in the previoussection39 In 1998 Ajayan and co-workers described astudy of the mechanical behaviour of MWNTepoxycomposites in both tension and compression58 It wasfound that the compression modulus was higher thanthe tensile modulus indicating that load transfer to thenanotubes in the composite was much higher in com-pression The authors suggested that during load trans-fer to MWNTs only the outer layers are stressed intension whereas all the layers respond in compression

Nanotubeepoxy thin films prepared by Xu et alshowed excellent mechanical properties42 Compared toneat resin thin films for which the elastic modulus was42 GPa a 20 increase in modulus was seen when01 wt- MWNTs were added This was attributed topartial alignment of the MWNTs induced by spincoating Fracture behaviour of the films was investigatedby SEM and showed that pulled-out tubes were oftencovered with polymer suggesting strong interfacial adhe-sion Similar results were reported by Gojny et al59 Thisgroup showed that functionalising arc produced MWNTswith amine groups greatly enhanced bonding with theepoxy matrix As a result the outer shells of the tubesoften remained embedded in the matrix following pull-out as shown in Fig 10

All of the work described so far in this section hasinvolved multiwalled carbon nanotubes Ajayan et al

have described a study of SWNT-containing compo-sites60 Two types of composite were prepared first anSWNTepoxy composite and second a pressed pelletcontaining SWNTs and carbonaceous soot materialformed during nanotube synthesis The SWNTs wereused as-prepared ie generally grouped into bundlesrather than separated into individual tubes The com-posites were loaded axially in tension until failureoccurred and the fracture surfaces were then examinedin detail using SEM In some cases the tube bundleswere found to have been pulled out of the matrix duringthe deformation and fracture of the composites while inother cases the nanotubes were not entirely pulled outbut were stretched between two fracture surfaces Theauthors believed that during pull-out the nanotubeswere sliding axially within the ropes and that the failureobserved at large crack distances was not failure ofindividual tubes but the bundles pulling apart Studiesby micro-Raman spectroscopy supported this hypoth-esis It was found that little shift in the second order A1g

band occurred as a result of the application of axialtension showing that the individual nanotubes were notbeing significantly stretched To take full advantage ofthe high Youngrsquos modulus of SWNTs in polymer com-posites it is clear that load must be transferred effec-tively from the matrix to the nanotubes Ajayan et alsuggested that load transfer could be improved by break-ing bundles down into individual tube fragments beforedispersing in the matrix Alternatively the bundles them-selves could be reinforced by crosslinking the tubes withinbundles (eg by chemical treatments or irradiation)

Single-walled nanotubeepoxy composites have alsobeen prepared by Biercuk et al61 The Vickers hardnessof the polymer was found to increase monotonicallywith addition of SWNTs up to a factor of 35 at 2 wt-nanotube loading Greatly enhanced thermal conducti-vities were also observed Thus samples loaded with1 wt- unpurified SWNT material showed a 70increase in thermal conductivity at 40 K rising to125 at room temperature

The preparation of SWNTpolyelectrolyte compositesby the layer-by-layer method has been described in theprevious section49 The films produced in this way wereshown to have exceptional mechanical properties Thusthe average ultimate tensile strength of the films wasfound to be 220 MPa with some measurements as highas 325 MPa This is several orders of magnitude greater

10 Image of epoxy coated MWNT subject to tensile

stress as result of electron irradiation telescopic

pull-out of outer tube layer has occurred59

9 TEM image of crack in MWNTpolystyrene thin film

induced by thermal stresses34

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 37

than the tensile strength of strong industrial plastics andindicates that the layer-by-layer method has great pro-mise for the production of strong nanotube-containingcomposites

Electrical and optical properties ofcarbon nanotubepolymer compositesThe effect of adding nanotubes on the electrical pro-perties of polymers has been studied by many groupsThese studies fall into a number of distinct categories Insome cases the nanotubes have been used to increase theconductivity of relatively low cost polymers as analternative to currently used fillers such as carbon blackOther studies have involved the incorporation of nano-tubes into conducting polymers such as polyaniline Stillfurther work has been aimed at specific applicationssuch as photovoltaic devices or supercapacitors In manycases improved mechanical properties have been avaluable by-product of the inclusion of nanotubes Theaddition of nanotubes to low cost bulk polymers isconsidered first

Electrostatic dissipationImproving the electrical conductivity of bulk polymers isimportant in a number of applications For example insome aircraft components enhanced conductivity isrequired to provide electrostatic discharge and electro-magnetic radio frequency interference protection Staticelectrical dissipation is also needed in other applicationsincluding computer housings and exterior automotiveparts With these types of application in mind Sandleret al investigated the electrical properties of nanotubeepoxy composites40 Matrix resistivities of about 100 V mwith filler volume fractions as low as 01 vol- wereachieved These figures represented an advance on thebest conductivity values previously obtained with carbonblack in the same epoxy matrix

Lozano and co-workers prepared composites ofcatalytically grown carbon nanofibres in a polypropy-lene matrix5162 The nanofibres were purified and func-tionalised to facilitate bonding with the matrix and highshear processing was used to disperse the fibres in thepolymer matrix Rheological and microscopic analysisshowed that a homogeneous dispersion of nanofibreswas formed and a percolation threshold for electricalconduction of 9ndash18 wt- was observed

These and other published studies suggest that carbonnanotubes have great promise in reducing electrostaticcharging of bulk polymers Nanotubes also have otheradvantages over conventional fillers such as carbonblack and carbon fibres in that they are more amenableto processing and can be more easily dispersed through-out the matrix In addition to the studies available in theopen literature there is undoubtedly a great deal ofcommercial work being carried out on the use of nano-tubes to reduce the electrostatic charging of plastics Infact nanotube containing plastics are already being usedin commercial products63 These include fuel lines inautomobiles where the nanotubes help to dissipate anydangerous charge which may build up Thermoplasticpolymers containing nanotubes are also used in someexterior automobile parts so that they can be earthedduring electrostatic painting

Conducting polymer compositesPolyaniline is a conducting polymer with many attrac-tive characteristics such as processability and environ-mental stability There has been interest in addingnanotubes to this polymer to provide both enhancedconductivity and improved mechanical properties Asmentioned above in the section lsquoPreparation of carbonnanotubepolymer compositesrsquo Cochet et al have pre-pared a MWNTpolyaniline composite by in situ poly-merisation44 The nanotubes were found to have astrong influence on the transport properties of thepolymer Thus the room temperature resistivity of thecomposite was found to be an order of magnitude lowerthan that of pure polyaniline while the low temperatureresistivity was much smaller than that of either poly-aniline or of MWNTs The temperature dependence ofresistivity was also weaker than that of polyanilinealone These observations were explained by assumingthat in situ polymerisation favours charge transfer betweenpolyaniline and MWNTs resulting in an overall materialwhich is more conducting than the starting componentsRaman studies indicated that these charge transferprocesses involved a site-selective interaction betweenthe quinoid ring of the polymer and the MWNTs Othergroups have also demonstrated excellent conductivity inMWNTpolyaniline composites6465 while a group fromDuPont have shown that SWNTpolyaniline compositescan be used as printable conductors for organic elec-tronics devices66

Optoelectronic propertiesSeveral groups have prepared nanotube compositesusing the conjugated polymer poly(p-phenylene viny-lene) (PPV) and its derivatives Polymers of this classhave electroluminescent properties and are widely usedin light-emitting diodes They are also used in photo-voltaic devices In an early study workers from Irelandand the USA described the synthesis of a compositematerial containing arc produced MWNTs in a PPVderivative PmPV67 The composite was prepared usinga solution mixing method and good wetting of the tubesby the polymer was achieved One of the aims of addingnanotubes to the polymer was to increase its electricalconductivity Previous attempts to achieve this by dopinghad resulted in degradation of the optical properties ofthe polymer In the case of nanotubes an increase in theelectrical conductivity of the polymer by up to eightorders of magnitude was seen with no concomitantdegradation of optical properties This was apparentlybecause the tubes acted as nanometric heat sinks pre-venting the build up of large thermal effects whichdamage these conjugated systems It was also shown inthis work that the composite could be used as theemissive layer in an organic light-emitting diode (LED)

As mentioned above several groups have been inter-ested in using nanotubepolymer composites in photo-voltaic devices The photovoltaic effect involves thegeneration of electron and hole pairs and their sub-sequent collection at opposite electrodes In organicmaterials the photon absorption excites an electron intothe conduction band which leads to a generation ofbound electronndashhole pairs or excitons These excitonshave to dissociate into free charges in order to be trans-ported to the electrodes Exciton dissociation can occurat a junction between a polymer and an electron acceptor

Harris Carbon nanotube composites

38 International Materials Reviews 2004 VOL 49 NO 1

molecule This allows the preferential transfer of the elec-trons into the electron acceptor molecule while leavingthe holes to be preferentially transported through thepolymer a process known as photoinduced chargetransfer In 1992 photoinduced charge transfer from aconducting polymer to C60 was discovered68 and thishas led to the fabrication of several efficient photo-voltaic systems using fullerenepolymer combinations69

Photovoltaic devices based on nanotubepolymer com-binations have now been fabricated by several groupsand this work will now be summarised

The earliest attempt to fabricate a nanotubepolymerphotovoltaic device appears to be the 1996 work ofRomero et al70 In this work a layer of MWNTs wasdeposited on Teflon and a PPV derivative was cast ontothis with a thin aluminium layer on top The devicedisplayed clear diode behaviour but only a small photo-current was observed when the junction was irradiatedwith light from a HendashNe laser This may have been dueto the relatively thick PPV active layer Ago et alfabricated a nanotube containing photovoltaic deviceby depositing a layer of PPV onto a glass-supportedMWNT film71 The interaction between MWNTs andPPV was investigated by means of photoluminescence(PL) measurements The photoluminescence efficiencyof the PPV was found to be markedly reduced by thepresence of the MWNT layer Since photoluminescenceis a result of radiative recombination a reduction mayimply improved exciton separation which is what onewishes to achieve in a photovoltaic device The devicefabricated by Ago et al is illustrated in Fig 11a Thisshowed a clear diode characteristic in the dark andupon illumination through the aluminium electrode aphotocurrent was observed A quantum efficiency approxi-mately twice that of the standard indiumndashtin oxidedevice was reported

Kymakis and Amaratunga used a slightly differentapproach to prepare a nanotubeconjugated polymerphotovoltaic device72 In this work composite SWNTpoly(3-octylthiophene) films were drop or spin cast froma solution onto indiumndashtin oxide (ITO) and quartz sub-strates to produce the arrangement shown in Fig 11bDevices were also prepared using the pure P3OT poly-mer A considerable enhancement of the photovoltaic

effect was observed with the P3OTSWNTs blend devicecompared with the pure polymer device

Non-linear optical propertiesAs mentioned above in the section lsquoStructure and pro-perties of carbon nanotubesrsquo it has been demonstratedthat suspensions of both SWNTs and MWNTs havenon-linear optical properties Nanotubepolymer com-posites may therefore have applications in optical devicesTo date however there has been little work on theoptical limiting properties of nanotube containing com-posites In one of the first studies in this area Chen et aldemonstrated third order optical non-linearity in a SWNTpolyimide composite73 Recently OrsquoFlaherty et al pre-pared a composite in which MWNTs were dispersed inthe polymer poly(99-di-n-octylfluorenyl-279-diyl) (PFO)74

Suspensions of the nanotubePFO composite in toluenewere prepared and the optical limiting properties of thesolutions were measured using nanosecond laser pulsesMaximum optical limiting was observed at carbonnanotube loadings in excess of 38 wt- relative to thepolymer

Carbon nanotubeceramic compositesCeramics have high stiffnesses and thermal stabilities butrelatively low breaking strengths Incorporating carbonnanotubes into a ceramic matrix might be expected toproduce a composite with both toughness and high tem-perature stability However achieving a homogeneousdispersion of tubes in an oxide with strong bondingbetween tubes and matrix presents rather more of achallenge than incorporating tubes into a polymer Variousapproaches have been adopted for incorporating nano-tubes into oxides such as alumina and silica and thesewill now be summarised

Peigney and co-workers have been among the pioneersin the synthesis of nanotubeoxide composites75ndash77 Theyhave developed an ingenious technique which involvesthe use of catalysts such as iron-containing a-aluminaPassing a CH4ndashH2 mixture over the catalysts resultedin the in situ growth of nanotubes on iron particlesproducing the precursor for a nanotubeoxide compo-site The resulting powders can then be hot pressed toform the final composite material The mechanicalproperties of these composites were initially disappoint-ing72 Thus the fracture strengths of the nanotubeFendashAl2O3 composites were only marginally higher than thatof Al2O3 and were generally markedly lower than thoseof the carbon-free FendashAl2O3 composites The fracturetoughness values were lower than or similar to that ofAl2O3 However SEM observations of composite frac-tures indicate that the nanotube bundles which are veryflexible could dissipate some fracture energy A similarmethod was used by An and Lim to prepare nanotubeAl2O3 composites in this case using C2H2 as the carbonsource78 Nanotube contents of up to 125 wt- wereachieved and a significant reinforcement effect wasreported

Achieving alignment of nanotubes in a ceramic matrixis much more difficult than in a polymer since ceramicsdo not lend themselves so easily to processing Howevera number of groups have shown that aligned nanotubemetal oxide composites can indeed be produced Peigney

ITO5indiumndashtin oxide PPV5poly(p-phenylene vinylene)

P3OT5poly(3-octylthiophene)

11 a Illustration of photovoltaic device fabricated by Ago

et al71 and b photovoltaic device fabricated by

Kymakis and Amaratunga72

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 39

et al exploited the fact that when their grains are main-tained at sufficiently low sizes (less than about 1 mm)many polycrystalline ceramics become superplasticallowing the use of thermomechanical processing forthe forming of these materials A series of cobalt andiron containing oxide catalysts were prepared77 Theprimary grain size in the powders was smaller than100 nm but in some of them the as produced primarygrains were strongly aggregated and required ball millingin order to reduce aggregate size to less than 1 mmNanotubes were then prepared in situ as describedabove and the powders were then extruded at hightemperatures in a graphite die under vacuum The bestresults were achieved with nanotubes incorporated intoFendashAl2O3 and FeCondashMgAl2O4 matrices In both casesthe nanotubes were not damaged by the processing andalignment of the nanotubes was clearly observed Infact the presence of the nanotubes seemed to facilitatethe superplastic forming by inhibiting grain growth andacting as lubricating agents It was demonstrated thatthe aligned nanotubemetal oxide composites showed ananisotropy of electrical conductivity which could beadjusted by controlling the nanotube content

One of the first studies of a SWNTceramic compositewas described by Zhan et al79 This group used thetechnique of spark plasma sintering (SPS) to prepareAl2O3 composites containing 10 vol- SWNTs Theadvantage of spark plasma sintering is that it allowsceramic powders to be annealed at lower temperaturesand for much shorter times than other sintering pro-cesses leading to the fabrication of fully dense ceramicsor composites with nanocrystalline microstructures undermild conditions Unlike some of the other processes usedto prepare nanotubeoxide composites spark plasmasintering does not damage the nanotubes and the nano-tube bundles could be seen to be located mainly at theboundaries of the Al2O3 grains in good contact with theAl2O3 matrix Before sintering the powders had beenmixed by ball milling which produced a reasonablyhomogeneous dispersion again without damaging thenanotubes The fracture toughnesses of the final com-posites were remarkable ndash about twice that of the fullydensified but unreinforced Al2O3 matrix In contrast toprevious work on nanotubeceramic composites thefracture toughness was found to increase with nanotubedensity This is thought to be due to the formation ofentangled networks of single-walled carbon nanotubeswhich may inhibit crack propagation A subsequentstudy of the nanotubeAl2O3 composites showed thatthey had high electrical conductivities80

Various techniques have been used to prepare carbonnanotubesilica composites Seeger et al described amethod which involved preparing a composite gel ofMWNTs with tetraethoxysilane (TEOS) and then sinter-ing this at 1150uC in argon81 A drawback of this tech-nique was that the sintering process led to a partialcrystallisation of the SiO2 resulting in a very inhomo-geneous matrix An alternative method which avoidsthese problems has been developed82 This involvesusing a Nd YAG laser to rapidly heat a TEOSnano-tube mixture resulting in partial melting of the matrixThis produced an amorphous silica matrix with nocrystallisation

A novel nanotubesilica composite material withpotentially useful optical properties has recently been

described by Han et al83 These workers synthesisedfilms consisting of silica spheres with diameters rangingfrom 200 to 650 nm Molybdenumcobalt catalyst parti-cles were then deposited onto the silica spheres and usedto catalyse the growth of SWNTs Since SWNTs arehighly non-linear and fast-switching materials they canbe incorporated into an optically confining environmentto achieve these characteristics at relatively low levels oflaser intensities

There have been a few studies on preparing nanotubecomposites using oxides other than alumina and silicaVincent et al reported the production of nanotubeTiO2

composites using solndashgel methods84 Possible applica-tions of the composites included optical non-linear waveguides and unidimensional conductive films Solndashgelmethods were also used by Sakamoto and Dunn toprepare SWNTV2O5 composites for use as electrodesin secondary lithium batteries85 Little work has beencarried out so far on incorporating nanotubes into non-oxide ceramics In one of the few published studies Maet al fabricated nanotubesilicon carbide composites bymixing SiC nanoparticles with 10 wt- MWNTs andhot pressing at 2000uC86 They found that the nanotubesplayed a strengthening and toughening role in thecomposite both bend strength and fracture toughnessincreased by about 10 compared to monolithic SiCThere have also been some initial studies on carbonnanotubecarbon fibre composites8788

Carbon nanotubemetal compositesComposite materials containing conventional carbonfibres in a metal matrix such as aluminium or mag-nesium are used in a number of specialist applica-tions8990 Such composites combine low density withhigh strength and modulus making them particularlyattractive to the aerospace industry There is growinginterest in the addition of carbon nanotubes to metalmatrices although only a small number of studies havebeen published to date Kuzumaki et al described thepreparation of a nanotubealuminium composite in 199891

Their method involved mixing a nanotube sample with afine aluminium powder mounting the mixture in a6 mm silver sheath and then drawing and heating thewire at 700uC in a vacuum furnace The result was acomposite wire in which the nanotubes were partiallyaligned along the axial direction The tensile strengths ofthe as prepared composite wires were comparable to thatof pure aluminium but the composite wires retained thisstrength after prolonged annealing at 600uC while thestrength of pure aluminium decreased by about 50after this treatment The same group subsequently pre-pared nanotubetitanium composites which showed alarge increase in hardness and modulus compared topure titanium92

Chen et al prepared nanotubenickelphosphorus com-posite coatings on stainless steel substrates using anelectroless plating method93 The coatings exhibited higherwear resistance and lower friction coefficients than SiCNiP and graphiteNiP composite coatings The samegroup prepared nanotubecopper composites by a powdermetallurgy technique and showed that the presence ofthe nanotubes again resulted in improved wear rates andfriction coefficients These results suggest that nanotube

Harris Carbon nanotube composites

40 International Materials Reviews 2004 VOL 49 NO 1

metal composites may have potential in tribologicalapplications

DiscussionThe subject of carbon nanotube composites has grownrapidly over the past 5 years or so into a vibrant andexciting area of research It is important to recognisehowever that this is still a new subject and the workcarried out so far may have hardly scratched the surfaceof what is possible For example the great majority ofthe studies that have been published to date haveinvolved the use of polymer matrices there is great scopefor more work on incorporating nanotubes into cera-mics metals and other forms of carbon Also most ofthe work up to now has been aimed at exploiting themechanical properties of carbon nanotubes rather thantheir electronic or optical properties although interest inthose is growing

When considering the mechanical performance ofcarbon nanotube composites it is clear that these dependstrongly on whether arc grown tubes or catalyticallyproduced tubes have been used In considering thecharacteristics of the two types of tube it is instructive tocompare the work of Qian et al34 with that of Watts andHsu57 Both incorporated the nanotubes into polymersand then used TEM to characterise the composites andstudy their fracture behaviour Qian et al used cataly-tically produced tubes and these were found to bewell dispersed throughout the polymer matrix with noclustering together of tubes However during compositefracture the tubes were often observed to break sug-gesting that they have relatively poor strength Wattsand Hsu on the other hand employed much moreperfect nanotubes produced by arc evaporation andTEM studies revealed superior mechanical propertieswith no observed breaking However the nanotubestended to cluster together so that during crack growththe tubes could quite easily slide past one another Thuswhile it is clear that the best mechanical performancewill be achieved with arc grown nanotubes it is impor-tant that the tubes be well dispersed throughout thematrix in order fully to exploit their properties

Equally important is achieving good bonding betweennanotubes and matrix Many studies have shown that thisinterfacial bonding can be enhanced by functionalising thenanotubes before embedding them in the matrix Theeffect of this can be seen in Fig 10 from the work ofGojny et al59 where bonding between MWNTs and anepoxy matrix is so strong that the outer shells of the tuberemain embedded in the matrix following pull-out Apossible disadvantage of functionalisation is that it couldweaken the nanotubes by disrupting the perfect cylindricalstructure This might be expected to be a particularproblem with single-walled tubes However functionali-sation of the nanotubes may not be necessary to producebonding with a polymer matrix Liao and Li havemodelled the interfacial characteristics of a nanotubepolystyrene composite assuming no covalent bondingbetween the tubes and the matrix94 It was found thatelectrostatic and van der Waals interactions contributedsignificantly to the interfacial stress transfer A nanotubepull-out simulation suggested that the interfacial shearstress of the nanotubepolystyrene system is about160 MPa significantly higher than for most carbon fibrereinforced polymer composite systems It is also worth

mentioning recent work by Star and co-workers9596

in which it was demonstrated that poly(5-alkoxy-m-phenylenevinylene)-co-[(25-dioctyloxy-p-phenylene)vinylene] (PAmPV) derivatives could be induced towrap around SWNTs forming strong non-covalentbonds

The quality of the nanotubes and the nature of theinterfacial bonding also has a strong influence on theelectrical properties of nanotubepolymer compositesArc grown tubes can have resistivities lower than that ofsingle crystal graphite but as already discussed theirelectronic properties depend on their precise structureAt present we do not have a reliable method of separat-ing metallic and semiconducting tubes so it is notpossible to exploit their full potential Catalytically pro-duced tubes are likely to be poorer conductors thanmetallic arc-grown tubes owing to their relatively dis-ordered structure Nevertheless catalytically grown tubesare sufficiently conducting to be useful in some com-mercial applications such as electrically conductive paintsThe influence of interfacial bonding on electronic pro-perties was demonstrated in the work of Cochet et al44

These workers prepared a MWNTpolyaniline com-posite by in situ polymerisation and found it to haveexcellent electrical transport properties This was attrib-uted to strong interaction between tubes and polymer

As has been shown in this review the possible appli-cations of carbon nanotube composites range widelyfrom electrically conductive paints to optical devicesOnly a relatively small number of these new materialshave been used commercially however and there arestill many problems to overcome before the full potentialof nanotube containing composites can be realised Notleast of these problems is the high cost of good qualitycarbon nanotubes Currently the commercial price ofunpurified arc-grownMWNTs is approximately $10000gwhile purified MWNTs cost around $30000g Cataly-tically grown MWNTs are much cheaper but as empha-sised throughout this review have generally inferiorproperties As far as SWNTs are concerned unpurifiedsamples cost around $5000g while purified SWNTs are$50000g As a comparison the price of gold is cur-rently about $1200g The development of new techni-ques for the low cost synthesis of high quality nanotubesis clearly essential This might be achieved throughimproved catalytic processes and there is currently alarge amount of work being carried out towards thisend Alternatively it may be possible to develop newnon-catalytic techniques for the synthesis of fullerene-related nanotubes In this connection it is interesting tonote that nanotubes can be produced by the high tem-perature heat treatment of conventional carbons such asmicroporous non-graphitising carbons9798 or carbonblack99100 Further work in this area would be of value

Acknowledgement

I thank Milo Shaffer for helpful comments

References1 P J F Harris lsquoCarbon nanotubes and related structures ndash new

materials for the twenty-first centuryrsquo 1999 Cambridge

Cambridge University Press

2 S Iijima Nature 1991 354 56ndash58

3 W Kratschmer L D Lamb K Fostiropoulos and D R

Huffman Nature 1990 347 354ndash358

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 41

4 S Iijima and T Ichihashi Nature 1993 363 603ndash605

5 D S Bethune C H Kiang M S De Vries G Gorman R Savoy

J Vasquez and R Beyers Nature 1993 363 605ndash607

6 H Dai A G Rinzler P Nikolaev A Thess D T Colbert and

R E Smalley Chem Phys Lett 1996 260 471ndash475

7 M M J Treacy T W Ebbesen and J M Gibson Nature 1996

381 678ndash680

8 E W Wong P E Sheehan and C M Lieber Science 1997 277

1971ndash1975

9 M R Falvo G J Clary R M Taylor V Chi F P Brooks

S Washburn and R Superfine Nature 1997 389 582ndash584

10 K Anazawa K Shimotani C Manabe H Watanabe and M

Shimizu Appl Phys Lett 2002 81 739ndash741

11 P Kim L Shi A Majumdar and P L McEuen Phys Rev Lett

2001 87 paper 215502

12 E T Thostenson Z F Ren and T W Chou Compos Sci

Technol 2001 61 1899ndash1912

13 K T Lau and D Hui Composites Part B Engineering 2002 33

263ndash277

14 Y Huh J Y Lee J Cheon Y K Hong J Y Koo T J Lee and

C J Lee J Mater Chem 2003 13 2297ndash2300

15 M Endo T Koyama and Y Hishiyama Jpn J Appl Phys 1976

15 2073ndash2076

16 S C Lyu B C Liu T J Lee Z Y Liu C W Yang C Y Park

and C J Lee Chem Commun 2003 734ndash735

17 M-F Yu O Lourie M J Dyer K Moloni T F Kelly and R S

Ruoff Science 2000 287 637ndash640

18 P Poncharal Z L Wang D Ugarte and W A De Heer Science

1999 283 1513ndash1516

19 J-P Salvetat A J Kulik J-M Bonard G A D Briggs

T Stockli K Metenier S Bonnamy F Beguin N A Burnham

and L Forro Adv Mater 1999 11 161ndash165

20 M-F Yu B S Files S Arepalli and R S Ruoff Phys Rev Lett

2000 84 5552ndash5555

21 M Terrones Int Mater Rev to be published

22 J W Mintmire B I Dunlap and C T White Phys Rev Lett

1992 68 631ndash634

23 R Saito M Fujita G Dresselhaus and M S Dresselhaus Phys

Rev B 1992 46 1804ndash1811

24 N Hamada S Sawada and A Oshiyama Phys Rev Lett 1992

68 1579ndash1581

25 T W Ebbesen H J Lezec H Hiura J W Bennett H F Ghaemi

and T Thio Nature 1996 382 54ndash56

26 J E Riggs D B Walker D L Carroll and Y-P Sun J Phys

Chem B 2000 104 7071ndash7076

27 S C Tsang Y K Chen P J F Harris and M L H Green

Nature 1994 372 159ndash162

28 R M Lago S C Tsang K L Lu Y K Chen and M L H

Green Chem Commun 1995 1355ndash1356

29 Y-P Sun K F Fu Y Lin and W Huang Acc Chem Res 2002

35 1096ndash1104

30 S B Sinnott J Nanosci Nanotechnol 2002 2113ndash123

31 M S P Shaffer X Fan and A H Windle Carbon 1998 36

1603ndash1612

32 M S P Shaffer and A H Windle Adv Mater 1999 11 937ndash941

33 D E Hill Y Lin A M Rao L F Allard and Y-P Sun

Macromolecules 2002 35 9466ndash9471

34 D Qian E C Dickey R Andrews and T Rantell Appl Phys

Lett 2000 76 2868ndash2870

35 R Andrews D Jacques D Qian and T Rantell Acc Chem Res

2002 35 1008ndash1017

36 R Haggenmueller H H Gommans A G Rinzler J E Fischer

and K I Winey Chem Phys Lett 2000 330 219ndash225

37 P Potschke T D Fornes and D R Paul Polymer 2002 43 3247ndash

3255

38 M Sennett E Welsh J B Wright W Z Li J G Wen and Z F

Ren Appl Phys A 2003 76 111ndash113

39 P M Ajayan O Stephan C Colliex and D Trauth Science 1994

265 1212ndash1214

40 J Sandler M S P Shaffer T Prasse W Bauhofer K Schulte

and H Windle Polymer 1999 40 5967ndash5971

41 A Allaoui S Bai H M Cheng and J B Bai Compos Sci

Technol 2002 62 1993ndash1998

42 X Xu M M Thwe C Shearwood and K Liao Appl Phys Lett

2002 81 2833ndash2835

43 D Puglia L Valentini and J M Kenny J Appl Polymer Sci

2003 88 452ndash458

44 M Cochet W K Maser A M Benito M A Callejas M T

Martınez J-M Benoit J Schreiber and O Chauvet Chem

Commun 2001 1450ndash1451

45 M S P Shaffer and K Koziol Chem Commun 2002 2074ndash

2075

46 C Park Z Ounaies K A Watson R E Crooks J Smith S E

Lowther J W Connell E J Siochi J S Harrison and T L St

Clair Chem Phys Lett 2002 364 303ndash308

47 K Jurewicz S Delpeux V Bertagna F Beguin and E

Frackowiak Chem Phys Lett 2001 347 36ndash40

48 M Hughes G Z Chen M S P Shaffer D J Fray and A H

Windle Chem Mater 2002 14 1610ndash1613

49 A A Mamedov N A Kotov M Prato D M Guldi J P

Wicksted and A Hirsch Nature Mater 2002 1 190ndash194

50 J H Rouse and P T Lillehei Nano Lett 2003 3 59ndash62

51 K Lozano and E V Barrera J Appl Polym Sci 2001 79 125ndash

133

52 L Valentini J Biagiotti J M Kenny and S Santucci Compos

Sci Technol 2003 63 1149ndash1153

53 J Sandler P Werner M S P Shaffer V Demchuk V Altstadt

and A H Windle Composites Part A Appl Sci Manuf 2002 33

1033ndash1039

54 R Czerw Z X Guo P M Ajayan Y P Sun and D L Carroll

Nano Lett 2001 1 423ndash427

55 B P Grady F Pompeo R L Shambaugh and D E Resasco

J Phys Chem B 2002 106 5852ndash5858

56 M Cadek J N Coleman V Barron K Hedicke and W J Blau

Appl Phys Lett 2002 81 5123ndash5125

57 P C P Watts and W K Hsu Nanotechnology 2003 14 L7ndashL10

58 L S Schadler S C Giannaris and P M Ajayan Appl Phys Lett

1998 73 3842ndash3844

59 F H Gojny J Nastalczyk Z Roslaniec and K Schulte Chem

Phys Lett 2003 370 820ndash824

60 P M Ajayan L S Schadler C Giannaris and A Rubio Adv

Mater 2000 12 750ndash753

61 M J Biercuk M C Llaguno M Radosavljevic J K Hyun A T

Johnson and J E Fischer Appl Phys Lett 2002 80 2767ndash2769

62 K Lozano J Bonilla-Rios and E V Barrera J Appl Polym Sci

2001 80 1162ndash1172

63 V Jamieson New Sci 15 Mar 2003 30

64 H Zengin W S Zhou J Y Jin W R Czerw D W Smith

L Echegoyen D L Carroll S H Foulger and J Ballato Adv

Mater 2002 14 1480ndash1483

65 J Deng X Ding W Zhang Y Peng J Wang X Long P Li and

A S C Chan Eur Polym J 2002 38 2497ndash2501

66 G B Blanchet C R Fincher and F Gao Appl Phys Lett 2003

82 1290ndash1292

67 S A Curran P M Ajayan W J Blau D L Carroll J N

Coleman A B Dalton A P Davey A Drury B McCarthy

S Maier and A A Strevens Adv Mater 1998 10 1091ndash1093

68 N S Sariciftci L Smilowitz A J Heeger and F Wudl Science

1992 258 1474ndash1476

69 N Camaioni L Garlaschelli A Geri M Maggini G Possamai

and G Ridolfi J Mater Chem 2002 12 2065ndash2070

70 D B Romero M Carrard W De Heer and L Zuppiroli

Adv Mater 1996 8 899ndash902

71 H Ago K Petritsch M S P Shaffer A H Windle and R H

Friend Adv Mater 1999 11 1281ndash1285

72 E Kymakis and G A J Amaratunga Appl Phys Lett 2002 80

112ndash114

73 Y-C Chen N R Raravikar L S Schadler P M Ajayan Y-P

Zhao T-M Lu G-C Wang and X-C Zhang Appl Phys Lett

2002 81 975ndash977

74 S A OrsquoFlaherty R Murphy S V Hold M Cadek J N

Coleman and W J Blau J Phys Chem B 2003 107 958ndash964

75 A Peigney C Laurent F Dobigeon and A Rousset J Mater

Res 1997 12 613ndash615

76 E Flahaut A Peigney C Laurent C Marliere F Chastel and

A Rousset Acta Mater 2000 48 3803ndash3812

77 A Peigney E Flahaut C Laurent F Chastel and A Rousset

Chem Phys Lett 2002 352 20ndash25

78 J W An and D S Lim J Ceram Process Res 2002 3 201ndash204

79 G-D Zhan J D Kuntz J L Wan and A K Mukherjee Nature

Mater 2003 2 38ndash42

80 G-D Zhan J D Kuntz J E Garay and A K Mukherjee Appl

Phys Lett 2003 83 1228ndash1230

81 T Seeger T Kohler T Frauenheim N Grobert M Ruhle

M Terrones and G Seifert Chem Commun 2002 34ndash35

82 T Seeger G De La Fuente W K Maser A M Benito M A

Callejas and M T Martınez Nanotechnology 2003 14 184ndash187

83 H Han S Vijayalakshmi A Lan Z Iqbal H Grebel E Lalanne

and A M Johnson Appl Phys Lett 2003 82 1458ndash1460

Harris Carbon nanotube composites

42 International Materials Reviews 2004 VOL 49 NO 1

84 P Vincent A Brioude C Journet S Rabaste S T Purcell J Le

Brusq and J C Plenet J Non-Cryst Sol 2002 311 130ndash137

85 J S Sakamoto and B Dunn J Electrochem Soc 2002 149 A26ndashA30

86 R Z Ma J Wu B Q Wei J Liang and D H Wu J Mater Sci

1998 33 5243ndash5246

87 R Andrews D Jacques A M Rao T Rantell F Derbyshire

Y Chen J Chen and R C Haddon Appl Phys Lett 1999 75

1329ndash1331

88 E T Thostenson W Z Li D Z Wang Z F Ren and T W

Chou J Appl Phys 2002 91 6034ndash6037

89 D D L Chung lsquoCarbon fiber compositesrsquo chap 7 1994 Boston

Butterworth-Heinemann

90 S Rawal JOM 2001 53 14ndash17

91 T Kuzumaki K Miyazawa H Ichinose and K Ito J Mater

Res 1998 13 2445ndash2449

92 T Kuzumaki O Ujiie H Ichinose and K Ito Adv Eng Mater

2000 2 416ndash418

93 W X Chen J P Tu L Y Wang H Y Gan Z D Xu and X B

Zhang Carbon 2003 41 215ndash222

94 K Liao and S Li Appl Phys Lett 2001 79 4225ndash4227

95 A Star J F Stoddart D Steuerman M Diehl A Boukai E W

Wong X Yang S W Chung H Choi and J R Heath

Angewandte Chemie-Int Ed 2001 40 1721ndash1725

96 A Star Y Liu K Grant L Ridvan J F Stoddart D W

Steuerman M R Diehl A Boukai and J R Heath

Macromolecules 2003 36 553ndash560

97 P J F Harris S C Tsang J B Claridge and M L H Green

J Chem Soc Faraday Trans 1994 90 2799ndash2802

98 A A Setlur S P Doherty J Y Dai and R P H Chang Appl

Phys Lett 2000 76 3008ndash3010

99 S P Doherty D B Buchholz B J Li and R P H Chang

J Mater Res 2003 18 941ndash949

100 D B Buchholz S P Doherty and R P H Chang Carbon 2003

41 1625ndash1634

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 43

Page 7: Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P. J. F. Harris * Carbon nanotubes are molecular-scale tubes of graphitic carbon with

the tubes were mixed with the polymer in an acidifiedaqueous solution and drops of this were applied to TEMgrids which were then dried in vacuum As in the workof Qian et al cracks were produced in the material byirradiation with an electron beam but a low acceleratingvoltage (75 kV) was used to avoid massive disruption ofthe thin composite film Individual tubes and bundles oftubes were observed to become stretched as the cracksgrew wider and pull-out was often seen In contrast tothe work of Qian et al however the tubes did notbreak This was probably because high quality arcgrown nanotubes were used rather than catalyticallyproduced tubes

The early work on nanotubeepoxy composites byAjayan et al has been described in the previoussection39 In 1998 Ajayan and co-workers described astudy of the mechanical behaviour of MWNTepoxycomposites in both tension and compression58 It wasfound that the compression modulus was higher thanthe tensile modulus indicating that load transfer to thenanotubes in the composite was much higher in com-pression The authors suggested that during load trans-fer to MWNTs only the outer layers are stressed intension whereas all the layers respond in compression

Nanotubeepoxy thin films prepared by Xu et alshowed excellent mechanical properties42 Compared toneat resin thin films for which the elastic modulus was42 GPa a 20 increase in modulus was seen when01 wt- MWNTs were added This was attributed topartial alignment of the MWNTs induced by spincoating Fracture behaviour of the films was investigatedby SEM and showed that pulled-out tubes were oftencovered with polymer suggesting strong interfacial adhe-sion Similar results were reported by Gojny et al59 Thisgroup showed that functionalising arc produced MWNTswith amine groups greatly enhanced bonding with theepoxy matrix As a result the outer shells of the tubesoften remained embedded in the matrix following pull-out as shown in Fig 10

All of the work described so far in this section hasinvolved multiwalled carbon nanotubes Ajayan et al

have described a study of SWNT-containing compo-sites60 Two types of composite were prepared first anSWNTepoxy composite and second a pressed pelletcontaining SWNTs and carbonaceous soot materialformed during nanotube synthesis The SWNTs wereused as-prepared ie generally grouped into bundlesrather than separated into individual tubes The com-posites were loaded axially in tension until failureoccurred and the fracture surfaces were then examinedin detail using SEM In some cases the tube bundleswere found to have been pulled out of the matrix duringthe deformation and fracture of the composites while inother cases the nanotubes were not entirely pulled outbut were stretched between two fracture surfaces Theauthors believed that during pull-out the nanotubeswere sliding axially within the ropes and that the failureobserved at large crack distances was not failure ofindividual tubes but the bundles pulling apart Studiesby micro-Raman spectroscopy supported this hypoth-esis It was found that little shift in the second order A1g

band occurred as a result of the application of axialtension showing that the individual nanotubes were notbeing significantly stretched To take full advantage ofthe high Youngrsquos modulus of SWNTs in polymer com-posites it is clear that load must be transferred effec-tively from the matrix to the nanotubes Ajayan et alsuggested that load transfer could be improved by break-ing bundles down into individual tube fragments beforedispersing in the matrix Alternatively the bundles them-selves could be reinforced by crosslinking the tubes withinbundles (eg by chemical treatments or irradiation)

Single-walled nanotubeepoxy composites have alsobeen prepared by Biercuk et al61 The Vickers hardnessof the polymer was found to increase monotonicallywith addition of SWNTs up to a factor of 35 at 2 wt-nanotube loading Greatly enhanced thermal conducti-vities were also observed Thus samples loaded with1 wt- unpurified SWNT material showed a 70increase in thermal conductivity at 40 K rising to125 at room temperature

The preparation of SWNTpolyelectrolyte compositesby the layer-by-layer method has been described in theprevious section49 The films produced in this way wereshown to have exceptional mechanical properties Thusthe average ultimate tensile strength of the films wasfound to be 220 MPa with some measurements as highas 325 MPa This is several orders of magnitude greater

10 Image of epoxy coated MWNT subject to tensile

stress as result of electron irradiation telescopic

pull-out of outer tube layer has occurred59

9 TEM image of crack in MWNTpolystyrene thin film

induced by thermal stresses34

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 37

than the tensile strength of strong industrial plastics andindicates that the layer-by-layer method has great pro-mise for the production of strong nanotube-containingcomposites

Electrical and optical properties ofcarbon nanotubepolymer compositesThe effect of adding nanotubes on the electrical pro-perties of polymers has been studied by many groupsThese studies fall into a number of distinct categories Insome cases the nanotubes have been used to increase theconductivity of relatively low cost polymers as analternative to currently used fillers such as carbon blackOther studies have involved the incorporation of nano-tubes into conducting polymers such as polyaniline Stillfurther work has been aimed at specific applicationssuch as photovoltaic devices or supercapacitors In manycases improved mechanical properties have been avaluable by-product of the inclusion of nanotubes Theaddition of nanotubes to low cost bulk polymers isconsidered first

Electrostatic dissipationImproving the electrical conductivity of bulk polymers isimportant in a number of applications For example insome aircraft components enhanced conductivity isrequired to provide electrostatic discharge and electro-magnetic radio frequency interference protection Staticelectrical dissipation is also needed in other applicationsincluding computer housings and exterior automotiveparts With these types of application in mind Sandleret al investigated the electrical properties of nanotubeepoxy composites40 Matrix resistivities of about 100 V mwith filler volume fractions as low as 01 vol- wereachieved These figures represented an advance on thebest conductivity values previously obtained with carbonblack in the same epoxy matrix

Lozano and co-workers prepared composites ofcatalytically grown carbon nanofibres in a polypropy-lene matrix5162 The nanofibres were purified and func-tionalised to facilitate bonding with the matrix and highshear processing was used to disperse the fibres in thepolymer matrix Rheological and microscopic analysisshowed that a homogeneous dispersion of nanofibreswas formed and a percolation threshold for electricalconduction of 9ndash18 wt- was observed

These and other published studies suggest that carbonnanotubes have great promise in reducing electrostaticcharging of bulk polymers Nanotubes also have otheradvantages over conventional fillers such as carbonblack and carbon fibres in that they are more amenableto processing and can be more easily dispersed through-out the matrix In addition to the studies available in theopen literature there is undoubtedly a great deal ofcommercial work being carried out on the use of nano-tubes to reduce the electrostatic charging of plastics Infact nanotube containing plastics are already being usedin commercial products63 These include fuel lines inautomobiles where the nanotubes help to dissipate anydangerous charge which may build up Thermoplasticpolymers containing nanotubes are also used in someexterior automobile parts so that they can be earthedduring electrostatic painting

Conducting polymer compositesPolyaniline is a conducting polymer with many attrac-tive characteristics such as processability and environ-mental stability There has been interest in addingnanotubes to this polymer to provide both enhancedconductivity and improved mechanical properties Asmentioned above in the section lsquoPreparation of carbonnanotubepolymer compositesrsquo Cochet et al have pre-pared a MWNTpolyaniline composite by in situ poly-merisation44 The nanotubes were found to have astrong influence on the transport properties of thepolymer Thus the room temperature resistivity of thecomposite was found to be an order of magnitude lowerthan that of pure polyaniline while the low temperatureresistivity was much smaller than that of either poly-aniline or of MWNTs The temperature dependence ofresistivity was also weaker than that of polyanilinealone These observations were explained by assumingthat in situ polymerisation favours charge transfer betweenpolyaniline and MWNTs resulting in an overall materialwhich is more conducting than the starting componentsRaman studies indicated that these charge transferprocesses involved a site-selective interaction betweenthe quinoid ring of the polymer and the MWNTs Othergroups have also demonstrated excellent conductivity inMWNTpolyaniline composites6465 while a group fromDuPont have shown that SWNTpolyaniline compositescan be used as printable conductors for organic elec-tronics devices66

Optoelectronic propertiesSeveral groups have prepared nanotube compositesusing the conjugated polymer poly(p-phenylene viny-lene) (PPV) and its derivatives Polymers of this classhave electroluminescent properties and are widely usedin light-emitting diodes They are also used in photo-voltaic devices In an early study workers from Irelandand the USA described the synthesis of a compositematerial containing arc produced MWNTs in a PPVderivative PmPV67 The composite was prepared usinga solution mixing method and good wetting of the tubesby the polymer was achieved One of the aims of addingnanotubes to the polymer was to increase its electricalconductivity Previous attempts to achieve this by dopinghad resulted in degradation of the optical properties ofthe polymer In the case of nanotubes an increase in theelectrical conductivity of the polymer by up to eightorders of magnitude was seen with no concomitantdegradation of optical properties This was apparentlybecause the tubes acted as nanometric heat sinks pre-venting the build up of large thermal effects whichdamage these conjugated systems It was also shown inthis work that the composite could be used as theemissive layer in an organic light-emitting diode (LED)

As mentioned above several groups have been inter-ested in using nanotubepolymer composites in photo-voltaic devices The photovoltaic effect involves thegeneration of electron and hole pairs and their sub-sequent collection at opposite electrodes In organicmaterials the photon absorption excites an electron intothe conduction band which leads to a generation ofbound electronndashhole pairs or excitons These excitonshave to dissociate into free charges in order to be trans-ported to the electrodes Exciton dissociation can occurat a junction between a polymer and an electron acceptor

Harris Carbon nanotube composites

38 International Materials Reviews 2004 VOL 49 NO 1

molecule This allows the preferential transfer of the elec-trons into the electron acceptor molecule while leavingthe holes to be preferentially transported through thepolymer a process known as photoinduced chargetransfer In 1992 photoinduced charge transfer from aconducting polymer to C60 was discovered68 and thishas led to the fabrication of several efficient photo-voltaic systems using fullerenepolymer combinations69

Photovoltaic devices based on nanotubepolymer com-binations have now been fabricated by several groupsand this work will now be summarised

The earliest attempt to fabricate a nanotubepolymerphotovoltaic device appears to be the 1996 work ofRomero et al70 In this work a layer of MWNTs wasdeposited on Teflon and a PPV derivative was cast ontothis with a thin aluminium layer on top The devicedisplayed clear diode behaviour but only a small photo-current was observed when the junction was irradiatedwith light from a HendashNe laser This may have been dueto the relatively thick PPV active layer Ago et alfabricated a nanotube containing photovoltaic deviceby depositing a layer of PPV onto a glass-supportedMWNT film71 The interaction between MWNTs andPPV was investigated by means of photoluminescence(PL) measurements The photoluminescence efficiencyof the PPV was found to be markedly reduced by thepresence of the MWNT layer Since photoluminescenceis a result of radiative recombination a reduction mayimply improved exciton separation which is what onewishes to achieve in a photovoltaic device The devicefabricated by Ago et al is illustrated in Fig 11a Thisshowed a clear diode characteristic in the dark andupon illumination through the aluminium electrode aphotocurrent was observed A quantum efficiency approxi-mately twice that of the standard indiumndashtin oxidedevice was reported

Kymakis and Amaratunga used a slightly differentapproach to prepare a nanotubeconjugated polymerphotovoltaic device72 In this work composite SWNTpoly(3-octylthiophene) films were drop or spin cast froma solution onto indiumndashtin oxide (ITO) and quartz sub-strates to produce the arrangement shown in Fig 11bDevices were also prepared using the pure P3OT poly-mer A considerable enhancement of the photovoltaic

effect was observed with the P3OTSWNTs blend devicecompared with the pure polymer device

Non-linear optical propertiesAs mentioned above in the section lsquoStructure and pro-perties of carbon nanotubesrsquo it has been demonstratedthat suspensions of both SWNTs and MWNTs havenon-linear optical properties Nanotubepolymer com-posites may therefore have applications in optical devicesTo date however there has been little work on theoptical limiting properties of nanotube containing com-posites In one of the first studies in this area Chen et aldemonstrated third order optical non-linearity in a SWNTpolyimide composite73 Recently OrsquoFlaherty et al pre-pared a composite in which MWNTs were dispersed inthe polymer poly(99-di-n-octylfluorenyl-279-diyl) (PFO)74

Suspensions of the nanotubePFO composite in toluenewere prepared and the optical limiting properties of thesolutions were measured using nanosecond laser pulsesMaximum optical limiting was observed at carbonnanotube loadings in excess of 38 wt- relative to thepolymer

Carbon nanotubeceramic compositesCeramics have high stiffnesses and thermal stabilities butrelatively low breaking strengths Incorporating carbonnanotubes into a ceramic matrix might be expected toproduce a composite with both toughness and high tem-perature stability However achieving a homogeneousdispersion of tubes in an oxide with strong bondingbetween tubes and matrix presents rather more of achallenge than incorporating tubes into a polymer Variousapproaches have been adopted for incorporating nano-tubes into oxides such as alumina and silica and thesewill now be summarised

Peigney and co-workers have been among the pioneersin the synthesis of nanotubeoxide composites75ndash77 Theyhave developed an ingenious technique which involvesthe use of catalysts such as iron-containing a-aluminaPassing a CH4ndashH2 mixture over the catalysts resultedin the in situ growth of nanotubes on iron particlesproducing the precursor for a nanotubeoxide compo-site The resulting powders can then be hot pressed toform the final composite material The mechanicalproperties of these composites were initially disappoint-ing72 Thus the fracture strengths of the nanotubeFendashAl2O3 composites were only marginally higher than thatof Al2O3 and were generally markedly lower than thoseof the carbon-free FendashAl2O3 composites The fracturetoughness values were lower than or similar to that ofAl2O3 However SEM observations of composite frac-tures indicate that the nanotube bundles which are veryflexible could dissipate some fracture energy A similarmethod was used by An and Lim to prepare nanotubeAl2O3 composites in this case using C2H2 as the carbonsource78 Nanotube contents of up to 125 wt- wereachieved and a significant reinforcement effect wasreported

Achieving alignment of nanotubes in a ceramic matrixis much more difficult than in a polymer since ceramicsdo not lend themselves so easily to processing Howevera number of groups have shown that aligned nanotubemetal oxide composites can indeed be produced Peigney

ITO5indiumndashtin oxide PPV5poly(p-phenylene vinylene)

P3OT5poly(3-octylthiophene)

11 a Illustration of photovoltaic device fabricated by Ago

et al71 and b photovoltaic device fabricated by

Kymakis and Amaratunga72

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 39

et al exploited the fact that when their grains are main-tained at sufficiently low sizes (less than about 1 mm)many polycrystalline ceramics become superplasticallowing the use of thermomechanical processing forthe forming of these materials A series of cobalt andiron containing oxide catalysts were prepared77 Theprimary grain size in the powders was smaller than100 nm but in some of them the as produced primarygrains were strongly aggregated and required ball millingin order to reduce aggregate size to less than 1 mmNanotubes were then prepared in situ as describedabove and the powders were then extruded at hightemperatures in a graphite die under vacuum The bestresults were achieved with nanotubes incorporated intoFendashAl2O3 and FeCondashMgAl2O4 matrices In both casesthe nanotubes were not damaged by the processing andalignment of the nanotubes was clearly observed Infact the presence of the nanotubes seemed to facilitatethe superplastic forming by inhibiting grain growth andacting as lubricating agents It was demonstrated thatthe aligned nanotubemetal oxide composites showed ananisotropy of electrical conductivity which could beadjusted by controlling the nanotube content

One of the first studies of a SWNTceramic compositewas described by Zhan et al79 This group used thetechnique of spark plasma sintering (SPS) to prepareAl2O3 composites containing 10 vol- SWNTs Theadvantage of spark plasma sintering is that it allowsceramic powders to be annealed at lower temperaturesand for much shorter times than other sintering pro-cesses leading to the fabrication of fully dense ceramicsor composites with nanocrystalline microstructures undermild conditions Unlike some of the other processes usedto prepare nanotubeoxide composites spark plasmasintering does not damage the nanotubes and the nano-tube bundles could be seen to be located mainly at theboundaries of the Al2O3 grains in good contact with theAl2O3 matrix Before sintering the powders had beenmixed by ball milling which produced a reasonablyhomogeneous dispersion again without damaging thenanotubes The fracture toughnesses of the final com-posites were remarkable ndash about twice that of the fullydensified but unreinforced Al2O3 matrix In contrast toprevious work on nanotubeceramic composites thefracture toughness was found to increase with nanotubedensity This is thought to be due to the formation ofentangled networks of single-walled carbon nanotubeswhich may inhibit crack propagation A subsequentstudy of the nanotubeAl2O3 composites showed thatthey had high electrical conductivities80

Various techniques have been used to prepare carbonnanotubesilica composites Seeger et al described amethod which involved preparing a composite gel ofMWNTs with tetraethoxysilane (TEOS) and then sinter-ing this at 1150uC in argon81 A drawback of this tech-nique was that the sintering process led to a partialcrystallisation of the SiO2 resulting in a very inhomo-geneous matrix An alternative method which avoidsthese problems has been developed82 This involvesusing a Nd YAG laser to rapidly heat a TEOSnano-tube mixture resulting in partial melting of the matrixThis produced an amorphous silica matrix with nocrystallisation

A novel nanotubesilica composite material withpotentially useful optical properties has recently been

described by Han et al83 These workers synthesisedfilms consisting of silica spheres with diameters rangingfrom 200 to 650 nm Molybdenumcobalt catalyst parti-cles were then deposited onto the silica spheres and usedto catalyse the growth of SWNTs Since SWNTs arehighly non-linear and fast-switching materials they canbe incorporated into an optically confining environmentto achieve these characteristics at relatively low levels oflaser intensities

There have been a few studies on preparing nanotubecomposites using oxides other than alumina and silicaVincent et al reported the production of nanotubeTiO2

composites using solndashgel methods84 Possible applica-tions of the composites included optical non-linear waveguides and unidimensional conductive films Solndashgelmethods were also used by Sakamoto and Dunn toprepare SWNTV2O5 composites for use as electrodesin secondary lithium batteries85 Little work has beencarried out so far on incorporating nanotubes into non-oxide ceramics In one of the few published studies Maet al fabricated nanotubesilicon carbide composites bymixing SiC nanoparticles with 10 wt- MWNTs andhot pressing at 2000uC86 They found that the nanotubesplayed a strengthening and toughening role in thecomposite both bend strength and fracture toughnessincreased by about 10 compared to monolithic SiCThere have also been some initial studies on carbonnanotubecarbon fibre composites8788

Carbon nanotubemetal compositesComposite materials containing conventional carbonfibres in a metal matrix such as aluminium or mag-nesium are used in a number of specialist applica-tions8990 Such composites combine low density withhigh strength and modulus making them particularlyattractive to the aerospace industry There is growinginterest in the addition of carbon nanotubes to metalmatrices although only a small number of studies havebeen published to date Kuzumaki et al described thepreparation of a nanotubealuminium composite in 199891

Their method involved mixing a nanotube sample with afine aluminium powder mounting the mixture in a6 mm silver sheath and then drawing and heating thewire at 700uC in a vacuum furnace The result was acomposite wire in which the nanotubes were partiallyaligned along the axial direction The tensile strengths ofthe as prepared composite wires were comparable to thatof pure aluminium but the composite wires retained thisstrength after prolonged annealing at 600uC while thestrength of pure aluminium decreased by about 50after this treatment The same group subsequently pre-pared nanotubetitanium composites which showed alarge increase in hardness and modulus compared topure titanium92

Chen et al prepared nanotubenickelphosphorus com-posite coatings on stainless steel substrates using anelectroless plating method93 The coatings exhibited higherwear resistance and lower friction coefficients than SiCNiP and graphiteNiP composite coatings The samegroup prepared nanotubecopper composites by a powdermetallurgy technique and showed that the presence ofthe nanotubes again resulted in improved wear rates andfriction coefficients These results suggest that nanotube

Harris Carbon nanotube composites

40 International Materials Reviews 2004 VOL 49 NO 1

metal composites may have potential in tribologicalapplications

DiscussionThe subject of carbon nanotube composites has grownrapidly over the past 5 years or so into a vibrant andexciting area of research It is important to recognisehowever that this is still a new subject and the workcarried out so far may have hardly scratched the surfaceof what is possible For example the great majority ofthe studies that have been published to date haveinvolved the use of polymer matrices there is great scopefor more work on incorporating nanotubes into cera-mics metals and other forms of carbon Also most ofthe work up to now has been aimed at exploiting themechanical properties of carbon nanotubes rather thantheir electronic or optical properties although interest inthose is growing

When considering the mechanical performance ofcarbon nanotube composites it is clear that these dependstrongly on whether arc grown tubes or catalyticallyproduced tubes have been used In considering thecharacteristics of the two types of tube it is instructive tocompare the work of Qian et al34 with that of Watts andHsu57 Both incorporated the nanotubes into polymersand then used TEM to characterise the composites andstudy their fracture behaviour Qian et al used cataly-tically produced tubes and these were found to bewell dispersed throughout the polymer matrix with noclustering together of tubes However during compositefracture the tubes were often observed to break sug-gesting that they have relatively poor strength Wattsand Hsu on the other hand employed much moreperfect nanotubes produced by arc evaporation andTEM studies revealed superior mechanical propertieswith no observed breaking However the nanotubestended to cluster together so that during crack growththe tubes could quite easily slide past one another Thuswhile it is clear that the best mechanical performancewill be achieved with arc grown nanotubes it is impor-tant that the tubes be well dispersed throughout thematrix in order fully to exploit their properties

Equally important is achieving good bonding betweennanotubes and matrix Many studies have shown that thisinterfacial bonding can be enhanced by functionalising thenanotubes before embedding them in the matrix Theeffect of this can be seen in Fig 10 from the work ofGojny et al59 where bonding between MWNTs and anepoxy matrix is so strong that the outer shells of the tuberemain embedded in the matrix following pull-out Apossible disadvantage of functionalisation is that it couldweaken the nanotubes by disrupting the perfect cylindricalstructure This might be expected to be a particularproblem with single-walled tubes However functionali-sation of the nanotubes may not be necessary to producebonding with a polymer matrix Liao and Li havemodelled the interfacial characteristics of a nanotubepolystyrene composite assuming no covalent bondingbetween the tubes and the matrix94 It was found thatelectrostatic and van der Waals interactions contributedsignificantly to the interfacial stress transfer A nanotubepull-out simulation suggested that the interfacial shearstress of the nanotubepolystyrene system is about160 MPa significantly higher than for most carbon fibrereinforced polymer composite systems It is also worth

mentioning recent work by Star and co-workers9596

in which it was demonstrated that poly(5-alkoxy-m-phenylenevinylene)-co-[(25-dioctyloxy-p-phenylene)vinylene] (PAmPV) derivatives could be induced towrap around SWNTs forming strong non-covalentbonds

The quality of the nanotubes and the nature of theinterfacial bonding also has a strong influence on theelectrical properties of nanotubepolymer compositesArc grown tubes can have resistivities lower than that ofsingle crystal graphite but as already discussed theirelectronic properties depend on their precise structureAt present we do not have a reliable method of separat-ing metallic and semiconducting tubes so it is notpossible to exploit their full potential Catalytically pro-duced tubes are likely to be poorer conductors thanmetallic arc-grown tubes owing to their relatively dis-ordered structure Nevertheless catalytically grown tubesare sufficiently conducting to be useful in some com-mercial applications such as electrically conductive paintsThe influence of interfacial bonding on electronic pro-perties was demonstrated in the work of Cochet et al44

These workers prepared a MWNTpolyaniline com-posite by in situ polymerisation and found it to haveexcellent electrical transport properties This was attrib-uted to strong interaction between tubes and polymer

As has been shown in this review the possible appli-cations of carbon nanotube composites range widelyfrom electrically conductive paints to optical devicesOnly a relatively small number of these new materialshave been used commercially however and there arestill many problems to overcome before the full potentialof nanotube containing composites can be realised Notleast of these problems is the high cost of good qualitycarbon nanotubes Currently the commercial price ofunpurified arc-grownMWNTs is approximately $10000gwhile purified MWNTs cost around $30000g Cataly-tically grown MWNTs are much cheaper but as empha-sised throughout this review have generally inferiorproperties As far as SWNTs are concerned unpurifiedsamples cost around $5000g while purified SWNTs are$50000g As a comparison the price of gold is cur-rently about $1200g The development of new techni-ques for the low cost synthesis of high quality nanotubesis clearly essential This might be achieved throughimproved catalytic processes and there is currently alarge amount of work being carried out towards thisend Alternatively it may be possible to develop newnon-catalytic techniques for the synthesis of fullerene-related nanotubes In this connection it is interesting tonote that nanotubes can be produced by the high tem-perature heat treatment of conventional carbons such asmicroporous non-graphitising carbons9798 or carbonblack99100 Further work in this area would be of value

Acknowledgement

I thank Milo Shaffer for helpful comments

References1 P J F Harris lsquoCarbon nanotubes and related structures ndash new

materials for the twenty-first centuryrsquo 1999 Cambridge

Cambridge University Press

2 S Iijima Nature 1991 354 56ndash58

3 W Kratschmer L D Lamb K Fostiropoulos and D R

Huffman Nature 1990 347 354ndash358

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 41

4 S Iijima and T Ichihashi Nature 1993 363 603ndash605

5 D S Bethune C H Kiang M S De Vries G Gorman R Savoy

J Vasquez and R Beyers Nature 1993 363 605ndash607

6 H Dai A G Rinzler P Nikolaev A Thess D T Colbert and

R E Smalley Chem Phys Lett 1996 260 471ndash475

7 M M J Treacy T W Ebbesen and J M Gibson Nature 1996

381 678ndash680

8 E W Wong P E Sheehan and C M Lieber Science 1997 277

1971ndash1975

9 M R Falvo G J Clary R M Taylor V Chi F P Brooks

S Washburn and R Superfine Nature 1997 389 582ndash584

10 K Anazawa K Shimotani C Manabe H Watanabe and M

Shimizu Appl Phys Lett 2002 81 739ndash741

11 P Kim L Shi A Majumdar and P L McEuen Phys Rev Lett

2001 87 paper 215502

12 E T Thostenson Z F Ren and T W Chou Compos Sci

Technol 2001 61 1899ndash1912

13 K T Lau and D Hui Composites Part B Engineering 2002 33

263ndash277

14 Y Huh J Y Lee J Cheon Y K Hong J Y Koo T J Lee and

C J Lee J Mater Chem 2003 13 2297ndash2300

15 M Endo T Koyama and Y Hishiyama Jpn J Appl Phys 1976

15 2073ndash2076

16 S C Lyu B C Liu T J Lee Z Y Liu C W Yang C Y Park

and C J Lee Chem Commun 2003 734ndash735

17 M-F Yu O Lourie M J Dyer K Moloni T F Kelly and R S

Ruoff Science 2000 287 637ndash640

18 P Poncharal Z L Wang D Ugarte and W A De Heer Science

1999 283 1513ndash1516

19 J-P Salvetat A J Kulik J-M Bonard G A D Briggs

T Stockli K Metenier S Bonnamy F Beguin N A Burnham

and L Forro Adv Mater 1999 11 161ndash165

20 M-F Yu B S Files S Arepalli and R S Ruoff Phys Rev Lett

2000 84 5552ndash5555

21 M Terrones Int Mater Rev to be published

22 J W Mintmire B I Dunlap and C T White Phys Rev Lett

1992 68 631ndash634

23 R Saito M Fujita G Dresselhaus and M S Dresselhaus Phys

Rev B 1992 46 1804ndash1811

24 N Hamada S Sawada and A Oshiyama Phys Rev Lett 1992

68 1579ndash1581

25 T W Ebbesen H J Lezec H Hiura J W Bennett H F Ghaemi

and T Thio Nature 1996 382 54ndash56

26 J E Riggs D B Walker D L Carroll and Y-P Sun J Phys

Chem B 2000 104 7071ndash7076

27 S C Tsang Y K Chen P J F Harris and M L H Green

Nature 1994 372 159ndash162

28 R M Lago S C Tsang K L Lu Y K Chen and M L H

Green Chem Commun 1995 1355ndash1356

29 Y-P Sun K F Fu Y Lin and W Huang Acc Chem Res 2002

35 1096ndash1104

30 S B Sinnott J Nanosci Nanotechnol 2002 2113ndash123

31 M S P Shaffer X Fan and A H Windle Carbon 1998 36

1603ndash1612

32 M S P Shaffer and A H Windle Adv Mater 1999 11 937ndash941

33 D E Hill Y Lin A M Rao L F Allard and Y-P Sun

Macromolecules 2002 35 9466ndash9471

34 D Qian E C Dickey R Andrews and T Rantell Appl Phys

Lett 2000 76 2868ndash2870

35 R Andrews D Jacques D Qian and T Rantell Acc Chem Res

2002 35 1008ndash1017

36 R Haggenmueller H H Gommans A G Rinzler J E Fischer

and K I Winey Chem Phys Lett 2000 330 219ndash225

37 P Potschke T D Fornes and D R Paul Polymer 2002 43 3247ndash

3255

38 M Sennett E Welsh J B Wright W Z Li J G Wen and Z F

Ren Appl Phys A 2003 76 111ndash113

39 P M Ajayan O Stephan C Colliex and D Trauth Science 1994

265 1212ndash1214

40 J Sandler M S P Shaffer T Prasse W Bauhofer K Schulte

and H Windle Polymer 1999 40 5967ndash5971

41 A Allaoui S Bai H M Cheng and J B Bai Compos Sci

Technol 2002 62 1993ndash1998

42 X Xu M M Thwe C Shearwood and K Liao Appl Phys Lett

2002 81 2833ndash2835

43 D Puglia L Valentini and J M Kenny J Appl Polymer Sci

2003 88 452ndash458

44 M Cochet W K Maser A M Benito M A Callejas M T

Martınez J-M Benoit J Schreiber and O Chauvet Chem

Commun 2001 1450ndash1451

45 M S P Shaffer and K Koziol Chem Commun 2002 2074ndash

2075

46 C Park Z Ounaies K A Watson R E Crooks J Smith S E

Lowther J W Connell E J Siochi J S Harrison and T L St

Clair Chem Phys Lett 2002 364 303ndash308

47 K Jurewicz S Delpeux V Bertagna F Beguin and E

Frackowiak Chem Phys Lett 2001 347 36ndash40

48 M Hughes G Z Chen M S P Shaffer D J Fray and A H

Windle Chem Mater 2002 14 1610ndash1613

49 A A Mamedov N A Kotov M Prato D M Guldi J P

Wicksted and A Hirsch Nature Mater 2002 1 190ndash194

50 J H Rouse and P T Lillehei Nano Lett 2003 3 59ndash62

51 K Lozano and E V Barrera J Appl Polym Sci 2001 79 125ndash

133

52 L Valentini J Biagiotti J M Kenny and S Santucci Compos

Sci Technol 2003 63 1149ndash1153

53 J Sandler P Werner M S P Shaffer V Demchuk V Altstadt

and A H Windle Composites Part A Appl Sci Manuf 2002 33

1033ndash1039

54 R Czerw Z X Guo P M Ajayan Y P Sun and D L Carroll

Nano Lett 2001 1 423ndash427

55 B P Grady F Pompeo R L Shambaugh and D E Resasco

J Phys Chem B 2002 106 5852ndash5858

56 M Cadek J N Coleman V Barron K Hedicke and W J Blau

Appl Phys Lett 2002 81 5123ndash5125

57 P C P Watts and W K Hsu Nanotechnology 2003 14 L7ndashL10

58 L S Schadler S C Giannaris and P M Ajayan Appl Phys Lett

1998 73 3842ndash3844

59 F H Gojny J Nastalczyk Z Roslaniec and K Schulte Chem

Phys Lett 2003 370 820ndash824

60 P M Ajayan L S Schadler C Giannaris and A Rubio Adv

Mater 2000 12 750ndash753

61 M J Biercuk M C Llaguno M Radosavljevic J K Hyun A T

Johnson and J E Fischer Appl Phys Lett 2002 80 2767ndash2769

62 K Lozano J Bonilla-Rios and E V Barrera J Appl Polym Sci

2001 80 1162ndash1172

63 V Jamieson New Sci 15 Mar 2003 30

64 H Zengin W S Zhou J Y Jin W R Czerw D W Smith

L Echegoyen D L Carroll S H Foulger and J Ballato Adv

Mater 2002 14 1480ndash1483

65 J Deng X Ding W Zhang Y Peng J Wang X Long P Li and

A S C Chan Eur Polym J 2002 38 2497ndash2501

66 G B Blanchet C R Fincher and F Gao Appl Phys Lett 2003

82 1290ndash1292

67 S A Curran P M Ajayan W J Blau D L Carroll J N

Coleman A B Dalton A P Davey A Drury B McCarthy

S Maier and A A Strevens Adv Mater 1998 10 1091ndash1093

68 N S Sariciftci L Smilowitz A J Heeger and F Wudl Science

1992 258 1474ndash1476

69 N Camaioni L Garlaschelli A Geri M Maggini G Possamai

and G Ridolfi J Mater Chem 2002 12 2065ndash2070

70 D B Romero M Carrard W De Heer and L Zuppiroli

Adv Mater 1996 8 899ndash902

71 H Ago K Petritsch M S P Shaffer A H Windle and R H

Friend Adv Mater 1999 11 1281ndash1285

72 E Kymakis and G A J Amaratunga Appl Phys Lett 2002 80

112ndash114

73 Y-C Chen N R Raravikar L S Schadler P M Ajayan Y-P

Zhao T-M Lu G-C Wang and X-C Zhang Appl Phys Lett

2002 81 975ndash977

74 S A OrsquoFlaherty R Murphy S V Hold M Cadek J N

Coleman and W J Blau J Phys Chem B 2003 107 958ndash964

75 A Peigney C Laurent F Dobigeon and A Rousset J Mater

Res 1997 12 613ndash615

76 E Flahaut A Peigney C Laurent C Marliere F Chastel and

A Rousset Acta Mater 2000 48 3803ndash3812

77 A Peigney E Flahaut C Laurent F Chastel and A Rousset

Chem Phys Lett 2002 352 20ndash25

78 J W An and D S Lim J Ceram Process Res 2002 3 201ndash204

79 G-D Zhan J D Kuntz J L Wan and A K Mukherjee Nature

Mater 2003 2 38ndash42

80 G-D Zhan J D Kuntz J E Garay and A K Mukherjee Appl

Phys Lett 2003 83 1228ndash1230

81 T Seeger T Kohler T Frauenheim N Grobert M Ruhle

M Terrones and G Seifert Chem Commun 2002 34ndash35

82 T Seeger G De La Fuente W K Maser A M Benito M A

Callejas and M T Martınez Nanotechnology 2003 14 184ndash187

83 H Han S Vijayalakshmi A Lan Z Iqbal H Grebel E Lalanne

and A M Johnson Appl Phys Lett 2003 82 1458ndash1460

Harris Carbon nanotube composites

42 International Materials Reviews 2004 VOL 49 NO 1

84 P Vincent A Brioude C Journet S Rabaste S T Purcell J Le

Brusq and J C Plenet J Non-Cryst Sol 2002 311 130ndash137

85 J S Sakamoto and B Dunn J Electrochem Soc 2002 149 A26ndashA30

86 R Z Ma J Wu B Q Wei J Liang and D H Wu J Mater Sci

1998 33 5243ndash5246

87 R Andrews D Jacques A M Rao T Rantell F Derbyshire

Y Chen J Chen and R C Haddon Appl Phys Lett 1999 75

1329ndash1331

88 E T Thostenson W Z Li D Z Wang Z F Ren and T W

Chou J Appl Phys 2002 91 6034ndash6037

89 D D L Chung lsquoCarbon fiber compositesrsquo chap 7 1994 Boston

Butterworth-Heinemann

90 S Rawal JOM 2001 53 14ndash17

91 T Kuzumaki K Miyazawa H Ichinose and K Ito J Mater

Res 1998 13 2445ndash2449

92 T Kuzumaki O Ujiie H Ichinose and K Ito Adv Eng Mater

2000 2 416ndash418

93 W X Chen J P Tu L Y Wang H Y Gan Z D Xu and X B

Zhang Carbon 2003 41 215ndash222

94 K Liao and S Li Appl Phys Lett 2001 79 4225ndash4227

95 A Star J F Stoddart D Steuerman M Diehl A Boukai E W

Wong X Yang S W Chung H Choi and J R Heath

Angewandte Chemie-Int Ed 2001 40 1721ndash1725

96 A Star Y Liu K Grant L Ridvan J F Stoddart D W

Steuerman M R Diehl A Boukai and J R Heath

Macromolecules 2003 36 553ndash560

97 P J F Harris S C Tsang J B Claridge and M L H Green

J Chem Soc Faraday Trans 1994 90 2799ndash2802

98 A A Setlur S P Doherty J Y Dai and R P H Chang Appl

Phys Lett 2000 76 3008ndash3010

99 S P Doherty D B Buchholz B J Li and R P H Chang

J Mater Res 2003 18 941ndash949

100 D B Buchholz S P Doherty and R P H Chang Carbon 2003

41 1625ndash1634

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 43

Page 8: Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P. J. F. Harris * Carbon nanotubes are molecular-scale tubes of graphitic carbon with

than the tensile strength of strong industrial plastics andindicates that the layer-by-layer method has great pro-mise for the production of strong nanotube-containingcomposites

Electrical and optical properties ofcarbon nanotubepolymer compositesThe effect of adding nanotubes on the electrical pro-perties of polymers has been studied by many groupsThese studies fall into a number of distinct categories Insome cases the nanotubes have been used to increase theconductivity of relatively low cost polymers as analternative to currently used fillers such as carbon blackOther studies have involved the incorporation of nano-tubes into conducting polymers such as polyaniline Stillfurther work has been aimed at specific applicationssuch as photovoltaic devices or supercapacitors In manycases improved mechanical properties have been avaluable by-product of the inclusion of nanotubes Theaddition of nanotubes to low cost bulk polymers isconsidered first

Electrostatic dissipationImproving the electrical conductivity of bulk polymers isimportant in a number of applications For example insome aircraft components enhanced conductivity isrequired to provide electrostatic discharge and electro-magnetic radio frequency interference protection Staticelectrical dissipation is also needed in other applicationsincluding computer housings and exterior automotiveparts With these types of application in mind Sandleret al investigated the electrical properties of nanotubeepoxy composites40 Matrix resistivities of about 100 V mwith filler volume fractions as low as 01 vol- wereachieved These figures represented an advance on thebest conductivity values previously obtained with carbonblack in the same epoxy matrix

Lozano and co-workers prepared composites ofcatalytically grown carbon nanofibres in a polypropy-lene matrix5162 The nanofibres were purified and func-tionalised to facilitate bonding with the matrix and highshear processing was used to disperse the fibres in thepolymer matrix Rheological and microscopic analysisshowed that a homogeneous dispersion of nanofibreswas formed and a percolation threshold for electricalconduction of 9ndash18 wt- was observed

These and other published studies suggest that carbonnanotubes have great promise in reducing electrostaticcharging of bulk polymers Nanotubes also have otheradvantages over conventional fillers such as carbonblack and carbon fibres in that they are more amenableto processing and can be more easily dispersed through-out the matrix In addition to the studies available in theopen literature there is undoubtedly a great deal ofcommercial work being carried out on the use of nano-tubes to reduce the electrostatic charging of plastics Infact nanotube containing plastics are already being usedin commercial products63 These include fuel lines inautomobiles where the nanotubes help to dissipate anydangerous charge which may build up Thermoplasticpolymers containing nanotubes are also used in someexterior automobile parts so that they can be earthedduring electrostatic painting

Conducting polymer compositesPolyaniline is a conducting polymer with many attrac-tive characteristics such as processability and environ-mental stability There has been interest in addingnanotubes to this polymer to provide both enhancedconductivity and improved mechanical properties Asmentioned above in the section lsquoPreparation of carbonnanotubepolymer compositesrsquo Cochet et al have pre-pared a MWNTpolyaniline composite by in situ poly-merisation44 The nanotubes were found to have astrong influence on the transport properties of thepolymer Thus the room temperature resistivity of thecomposite was found to be an order of magnitude lowerthan that of pure polyaniline while the low temperatureresistivity was much smaller than that of either poly-aniline or of MWNTs The temperature dependence ofresistivity was also weaker than that of polyanilinealone These observations were explained by assumingthat in situ polymerisation favours charge transfer betweenpolyaniline and MWNTs resulting in an overall materialwhich is more conducting than the starting componentsRaman studies indicated that these charge transferprocesses involved a site-selective interaction betweenthe quinoid ring of the polymer and the MWNTs Othergroups have also demonstrated excellent conductivity inMWNTpolyaniline composites6465 while a group fromDuPont have shown that SWNTpolyaniline compositescan be used as printable conductors for organic elec-tronics devices66

Optoelectronic propertiesSeveral groups have prepared nanotube compositesusing the conjugated polymer poly(p-phenylene viny-lene) (PPV) and its derivatives Polymers of this classhave electroluminescent properties and are widely usedin light-emitting diodes They are also used in photo-voltaic devices In an early study workers from Irelandand the USA described the synthesis of a compositematerial containing arc produced MWNTs in a PPVderivative PmPV67 The composite was prepared usinga solution mixing method and good wetting of the tubesby the polymer was achieved One of the aims of addingnanotubes to the polymer was to increase its electricalconductivity Previous attempts to achieve this by dopinghad resulted in degradation of the optical properties ofthe polymer In the case of nanotubes an increase in theelectrical conductivity of the polymer by up to eightorders of magnitude was seen with no concomitantdegradation of optical properties This was apparentlybecause the tubes acted as nanometric heat sinks pre-venting the build up of large thermal effects whichdamage these conjugated systems It was also shown inthis work that the composite could be used as theemissive layer in an organic light-emitting diode (LED)

As mentioned above several groups have been inter-ested in using nanotubepolymer composites in photo-voltaic devices The photovoltaic effect involves thegeneration of electron and hole pairs and their sub-sequent collection at opposite electrodes In organicmaterials the photon absorption excites an electron intothe conduction band which leads to a generation ofbound electronndashhole pairs or excitons These excitonshave to dissociate into free charges in order to be trans-ported to the electrodes Exciton dissociation can occurat a junction between a polymer and an electron acceptor

Harris Carbon nanotube composites

38 International Materials Reviews 2004 VOL 49 NO 1

molecule This allows the preferential transfer of the elec-trons into the electron acceptor molecule while leavingthe holes to be preferentially transported through thepolymer a process known as photoinduced chargetransfer In 1992 photoinduced charge transfer from aconducting polymer to C60 was discovered68 and thishas led to the fabrication of several efficient photo-voltaic systems using fullerenepolymer combinations69

Photovoltaic devices based on nanotubepolymer com-binations have now been fabricated by several groupsand this work will now be summarised

The earliest attempt to fabricate a nanotubepolymerphotovoltaic device appears to be the 1996 work ofRomero et al70 In this work a layer of MWNTs wasdeposited on Teflon and a PPV derivative was cast ontothis with a thin aluminium layer on top The devicedisplayed clear diode behaviour but only a small photo-current was observed when the junction was irradiatedwith light from a HendashNe laser This may have been dueto the relatively thick PPV active layer Ago et alfabricated a nanotube containing photovoltaic deviceby depositing a layer of PPV onto a glass-supportedMWNT film71 The interaction between MWNTs andPPV was investigated by means of photoluminescence(PL) measurements The photoluminescence efficiencyof the PPV was found to be markedly reduced by thepresence of the MWNT layer Since photoluminescenceis a result of radiative recombination a reduction mayimply improved exciton separation which is what onewishes to achieve in a photovoltaic device The devicefabricated by Ago et al is illustrated in Fig 11a Thisshowed a clear diode characteristic in the dark andupon illumination through the aluminium electrode aphotocurrent was observed A quantum efficiency approxi-mately twice that of the standard indiumndashtin oxidedevice was reported

Kymakis and Amaratunga used a slightly differentapproach to prepare a nanotubeconjugated polymerphotovoltaic device72 In this work composite SWNTpoly(3-octylthiophene) films were drop or spin cast froma solution onto indiumndashtin oxide (ITO) and quartz sub-strates to produce the arrangement shown in Fig 11bDevices were also prepared using the pure P3OT poly-mer A considerable enhancement of the photovoltaic

effect was observed with the P3OTSWNTs blend devicecompared with the pure polymer device

Non-linear optical propertiesAs mentioned above in the section lsquoStructure and pro-perties of carbon nanotubesrsquo it has been demonstratedthat suspensions of both SWNTs and MWNTs havenon-linear optical properties Nanotubepolymer com-posites may therefore have applications in optical devicesTo date however there has been little work on theoptical limiting properties of nanotube containing com-posites In one of the first studies in this area Chen et aldemonstrated third order optical non-linearity in a SWNTpolyimide composite73 Recently OrsquoFlaherty et al pre-pared a composite in which MWNTs were dispersed inthe polymer poly(99-di-n-octylfluorenyl-279-diyl) (PFO)74

Suspensions of the nanotubePFO composite in toluenewere prepared and the optical limiting properties of thesolutions were measured using nanosecond laser pulsesMaximum optical limiting was observed at carbonnanotube loadings in excess of 38 wt- relative to thepolymer

Carbon nanotubeceramic compositesCeramics have high stiffnesses and thermal stabilities butrelatively low breaking strengths Incorporating carbonnanotubes into a ceramic matrix might be expected toproduce a composite with both toughness and high tem-perature stability However achieving a homogeneousdispersion of tubes in an oxide with strong bondingbetween tubes and matrix presents rather more of achallenge than incorporating tubes into a polymer Variousapproaches have been adopted for incorporating nano-tubes into oxides such as alumina and silica and thesewill now be summarised

Peigney and co-workers have been among the pioneersin the synthesis of nanotubeoxide composites75ndash77 Theyhave developed an ingenious technique which involvesthe use of catalysts such as iron-containing a-aluminaPassing a CH4ndashH2 mixture over the catalysts resultedin the in situ growth of nanotubes on iron particlesproducing the precursor for a nanotubeoxide compo-site The resulting powders can then be hot pressed toform the final composite material The mechanicalproperties of these composites were initially disappoint-ing72 Thus the fracture strengths of the nanotubeFendashAl2O3 composites were only marginally higher than thatof Al2O3 and were generally markedly lower than thoseof the carbon-free FendashAl2O3 composites The fracturetoughness values were lower than or similar to that ofAl2O3 However SEM observations of composite frac-tures indicate that the nanotube bundles which are veryflexible could dissipate some fracture energy A similarmethod was used by An and Lim to prepare nanotubeAl2O3 composites in this case using C2H2 as the carbonsource78 Nanotube contents of up to 125 wt- wereachieved and a significant reinforcement effect wasreported

Achieving alignment of nanotubes in a ceramic matrixis much more difficult than in a polymer since ceramicsdo not lend themselves so easily to processing Howevera number of groups have shown that aligned nanotubemetal oxide composites can indeed be produced Peigney

ITO5indiumndashtin oxide PPV5poly(p-phenylene vinylene)

P3OT5poly(3-octylthiophene)

11 a Illustration of photovoltaic device fabricated by Ago

et al71 and b photovoltaic device fabricated by

Kymakis and Amaratunga72

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 39

et al exploited the fact that when their grains are main-tained at sufficiently low sizes (less than about 1 mm)many polycrystalline ceramics become superplasticallowing the use of thermomechanical processing forthe forming of these materials A series of cobalt andiron containing oxide catalysts were prepared77 Theprimary grain size in the powders was smaller than100 nm but in some of them the as produced primarygrains were strongly aggregated and required ball millingin order to reduce aggregate size to less than 1 mmNanotubes were then prepared in situ as describedabove and the powders were then extruded at hightemperatures in a graphite die under vacuum The bestresults were achieved with nanotubes incorporated intoFendashAl2O3 and FeCondashMgAl2O4 matrices In both casesthe nanotubes were not damaged by the processing andalignment of the nanotubes was clearly observed Infact the presence of the nanotubes seemed to facilitatethe superplastic forming by inhibiting grain growth andacting as lubricating agents It was demonstrated thatthe aligned nanotubemetal oxide composites showed ananisotropy of electrical conductivity which could beadjusted by controlling the nanotube content

One of the first studies of a SWNTceramic compositewas described by Zhan et al79 This group used thetechnique of spark plasma sintering (SPS) to prepareAl2O3 composites containing 10 vol- SWNTs Theadvantage of spark plasma sintering is that it allowsceramic powders to be annealed at lower temperaturesand for much shorter times than other sintering pro-cesses leading to the fabrication of fully dense ceramicsor composites with nanocrystalline microstructures undermild conditions Unlike some of the other processes usedto prepare nanotubeoxide composites spark plasmasintering does not damage the nanotubes and the nano-tube bundles could be seen to be located mainly at theboundaries of the Al2O3 grains in good contact with theAl2O3 matrix Before sintering the powders had beenmixed by ball milling which produced a reasonablyhomogeneous dispersion again without damaging thenanotubes The fracture toughnesses of the final com-posites were remarkable ndash about twice that of the fullydensified but unreinforced Al2O3 matrix In contrast toprevious work on nanotubeceramic composites thefracture toughness was found to increase with nanotubedensity This is thought to be due to the formation ofentangled networks of single-walled carbon nanotubeswhich may inhibit crack propagation A subsequentstudy of the nanotubeAl2O3 composites showed thatthey had high electrical conductivities80

Various techniques have been used to prepare carbonnanotubesilica composites Seeger et al described amethod which involved preparing a composite gel ofMWNTs with tetraethoxysilane (TEOS) and then sinter-ing this at 1150uC in argon81 A drawback of this tech-nique was that the sintering process led to a partialcrystallisation of the SiO2 resulting in a very inhomo-geneous matrix An alternative method which avoidsthese problems has been developed82 This involvesusing a Nd YAG laser to rapidly heat a TEOSnano-tube mixture resulting in partial melting of the matrixThis produced an amorphous silica matrix with nocrystallisation

A novel nanotubesilica composite material withpotentially useful optical properties has recently been

described by Han et al83 These workers synthesisedfilms consisting of silica spheres with diameters rangingfrom 200 to 650 nm Molybdenumcobalt catalyst parti-cles were then deposited onto the silica spheres and usedto catalyse the growth of SWNTs Since SWNTs arehighly non-linear and fast-switching materials they canbe incorporated into an optically confining environmentto achieve these characteristics at relatively low levels oflaser intensities

There have been a few studies on preparing nanotubecomposites using oxides other than alumina and silicaVincent et al reported the production of nanotubeTiO2

composites using solndashgel methods84 Possible applica-tions of the composites included optical non-linear waveguides and unidimensional conductive films Solndashgelmethods were also used by Sakamoto and Dunn toprepare SWNTV2O5 composites for use as electrodesin secondary lithium batteries85 Little work has beencarried out so far on incorporating nanotubes into non-oxide ceramics In one of the few published studies Maet al fabricated nanotubesilicon carbide composites bymixing SiC nanoparticles with 10 wt- MWNTs andhot pressing at 2000uC86 They found that the nanotubesplayed a strengthening and toughening role in thecomposite both bend strength and fracture toughnessincreased by about 10 compared to monolithic SiCThere have also been some initial studies on carbonnanotubecarbon fibre composites8788

Carbon nanotubemetal compositesComposite materials containing conventional carbonfibres in a metal matrix such as aluminium or mag-nesium are used in a number of specialist applica-tions8990 Such composites combine low density withhigh strength and modulus making them particularlyattractive to the aerospace industry There is growinginterest in the addition of carbon nanotubes to metalmatrices although only a small number of studies havebeen published to date Kuzumaki et al described thepreparation of a nanotubealuminium composite in 199891

Their method involved mixing a nanotube sample with afine aluminium powder mounting the mixture in a6 mm silver sheath and then drawing and heating thewire at 700uC in a vacuum furnace The result was acomposite wire in which the nanotubes were partiallyaligned along the axial direction The tensile strengths ofthe as prepared composite wires were comparable to thatof pure aluminium but the composite wires retained thisstrength after prolonged annealing at 600uC while thestrength of pure aluminium decreased by about 50after this treatment The same group subsequently pre-pared nanotubetitanium composites which showed alarge increase in hardness and modulus compared topure titanium92

Chen et al prepared nanotubenickelphosphorus com-posite coatings on stainless steel substrates using anelectroless plating method93 The coatings exhibited higherwear resistance and lower friction coefficients than SiCNiP and graphiteNiP composite coatings The samegroup prepared nanotubecopper composites by a powdermetallurgy technique and showed that the presence ofthe nanotubes again resulted in improved wear rates andfriction coefficients These results suggest that nanotube

Harris Carbon nanotube composites

40 International Materials Reviews 2004 VOL 49 NO 1

metal composites may have potential in tribologicalapplications

DiscussionThe subject of carbon nanotube composites has grownrapidly over the past 5 years or so into a vibrant andexciting area of research It is important to recognisehowever that this is still a new subject and the workcarried out so far may have hardly scratched the surfaceof what is possible For example the great majority ofthe studies that have been published to date haveinvolved the use of polymer matrices there is great scopefor more work on incorporating nanotubes into cera-mics metals and other forms of carbon Also most ofthe work up to now has been aimed at exploiting themechanical properties of carbon nanotubes rather thantheir electronic or optical properties although interest inthose is growing

When considering the mechanical performance ofcarbon nanotube composites it is clear that these dependstrongly on whether arc grown tubes or catalyticallyproduced tubes have been used In considering thecharacteristics of the two types of tube it is instructive tocompare the work of Qian et al34 with that of Watts andHsu57 Both incorporated the nanotubes into polymersand then used TEM to characterise the composites andstudy their fracture behaviour Qian et al used cataly-tically produced tubes and these were found to bewell dispersed throughout the polymer matrix with noclustering together of tubes However during compositefracture the tubes were often observed to break sug-gesting that they have relatively poor strength Wattsand Hsu on the other hand employed much moreperfect nanotubes produced by arc evaporation andTEM studies revealed superior mechanical propertieswith no observed breaking However the nanotubestended to cluster together so that during crack growththe tubes could quite easily slide past one another Thuswhile it is clear that the best mechanical performancewill be achieved with arc grown nanotubes it is impor-tant that the tubes be well dispersed throughout thematrix in order fully to exploit their properties

Equally important is achieving good bonding betweennanotubes and matrix Many studies have shown that thisinterfacial bonding can be enhanced by functionalising thenanotubes before embedding them in the matrix Theeffect of this can be seen in Fig 10 from the work ofGojny et al59 where bonding between MWNTs and anepoxy matrix is so strong that the outer shells of the tuberemain embedded in the matrix following pull-out Apossible disadvantage of functionalisation is that it couldweaken the nanotubes by disrupting the perfect cylindricalstructure This might be expected to be a particularproblem with single-walled tubes However functionali-sation of the nanotubes may not be necessary to producebonding with a polymer matrix Liao and Li havemodelled the interfacial characteristics of a nanotubepolystyrene composite assuming no covalent bondingbetween the tubes and the matrix94 It was found thatelectrostatic and van der Waals interactions contributedsignificantly to the interfacial stress transfer A nanotubepull-out simulation suggested that the interfacial shearstress of the nanotubepolystyrene system is about160 MPa significantly higher than for most carbon fibrereinforced polymer composite systems It is also worth

mentioning recent work by Star and co-workers9596

in which it was demonstrated that poly(5-alkoxy-m-phenylenevinylene)-co-[(25-dioctyloxy-p-phenylene)vinylene] (PAmPV) derivatives could be induced towrap around SWNTs forming strong non-covalentbonds

The quality of the nanotubes and the nature of theinterfacial bonding also has a strong influence on theelectrical properties of nanotubepolymer compositesArc grown tubes can have resistivities lower than that ofsingle crystal graphite but as already discussed theirelectronic properties depend on their precise structureAt present we do not have a reliable method of separat-ing metallic and semiconducting tubes so it is notpossible to exploit their full potential Catalytically pro-duced tubes are likely to be poorer conductors thanmetallic arc-grown tubes owing to their relatively dis-ordered structure Nevertheless catalytically grown tubesare sufficiently conducting to be useful in some com-mercial applications such as electrically conductive paintsThe influence of interfacial bonding on electronic pro-perties was demonstrated in the work of Cochet et al44

These workers prepared a MWNTpolyaniline com-posite by in situ polymerisation and found it to haveexcellent electrical transport properties This was attrib-uted to strong interaction between tubes and polymer

As has been shown in this review the possible appli-cations of carbon nanotube composites range widelyfrom electrically conductive paints to optical devicesOnly a relatively small number of these new materialshave been used commercially however and there arestill many problems to overcome before the full potentialof nanotube containing composites can be realised Notleast of these problems is the high cost of good qualitycarbon nanotubes Currently the commercial price ofunpurified arc-grownMWNTs is approximately $10000gwhile purified MWNTs cost around $30000g Cataly-tically grown MWNTs are much cheaper but as empha-sised throughout this review have generally inferiorproperties As far as SWNTs are concerned unpurifiedsamples cost around $5000g while purified SWNTs are$50000g As a comparison the price of gold is cur-rently about $1200g The development of new techni-ques for the low cost synthesis of high quality nanotubesis clearly essential This might be achieved throughimproved catalytic processes and there is currently alarge amount of work being carried out towards thisend Alternatively it may be possible to develop newnon-catalytic techniques for the synthesis of fullerene-related nanotubes In this connection it is interesting tonote that nanotubes can be produced by the high tem-perature heat treatment of conventional carbons such asmicroporous non-graphitising carbons9798 or carbonblack99100 Further work in this area would be of value

Acknowledgement

I thank Milo Shaffer for helpful comments

References1 P J F Harris lsquoCarbon nanotubes and related structures ndash new

materials for the twenty-first centuryrsquo 1999 Cambridge

Cambridge University Press

2 S Iijima Nature 1991 354 56ndash58

3 W Kratschmer L D Lamb K Fostiropoulos and D R

Huffman Nature 1990 347 354ndash358

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 41

4 S Iijima and T Ichihashi Nature 1993 363 603ndash605

5 D S Bethune C H Kiang M S De Vries G Gorman R Savoy

J Vasquez and R Beyers Nature 1993 363 605ndash607

6 H Dai A G Rinzler P Nikolaev A Thess D T Colbert and

R E Smalley Chem Phys Lett 1996 260 471ndash475

7 M M J Treacy T W Ebbesen and J M Gibson Nature 1996

381 678ndash680

8 E W Wong P E Sheehan and C M Lieber Science 1997 277

1971ndash1975

9 M R Falvo G J Clary R M Taylor V Chi F P Brooks

S Washburn and R Superfine Nature 1997 389 582ndash584

10 K Anazawa K Shimotani C Manabe H Watanabe and M

Shimizu Appl Phys Lett 2002 81 739ndash741

11 P Kim L Shi A Majumdar and P L McEuen Phys Rev Lett

2001 87 paper 215502

12 E T Thostenson Z F Ren and T W Chou Compos Sci

Technol 2001 61 1899ndash1912

13 K T Lau and D Hui Composites Part B Engineering 2002 33

263ndash277

14 Y Huh J Y Lee J Cheon Y K Hong J Y Koo T J Lee and

C J Lee J Mater Chem 2003 13 2297ndash2300

15 M Endo T Koyama and Y Hishiyama Jpn J Appl Phys 1976

15 2073ndash2076

16 S C Lyu B C Liu T J Lee Z Y Liu C W Yang C Y Park

and C J Lee Chem Commun 2003 734ndash735

17 M-F Yu O Lourie M J Dyer K Moloni T F Kelly and R S

Ruoff Science 2000 287 637ndash640

18 P Poncharal Z L Wang D Ugarte and W A De Heer Science

1999 283 1513ndash1516

19 J-P Salvetat A J Kulik J-M Bonard G A D Briggs

T Stockli K Metenier S Bonnamy F Beguin N A Burnham

and L Forro Adv Mater 1999 11 161ndash165

20 M-F Yu B S Files S Arepalli and R S Ruoff Phys Rev Lett

2000 84 5552ndash5555

21 M Terrones Int Mater Rev to be published

22 J W Mintmire B I Dunlap and C T White Phys Rev Lett

1992 68 631ndash634

23 R Saito M Fujita G Dresselhaus and M S Dresselhaus Phys

Rev B 1992 46 1804ndash1811

24 N Hamada S Sawada and A Oshiyama Phys Rev Lett 1992

68 1579ndash1581

25 T W Ebbesen H J Lezec H Hiura J W Bennett H F Ghaemi

and T Thio Nature 1996 382 54ndash56

26 J E Riggs D B Walker D L Carroll and Y-P Sun J Phys

Chem B 2000 104 7071ndash7076

27 S C Tsang Y K Chen P J F Harris and M L H Green

Nature 1994 372 159ndash162

28 R M Lago S C Tsang K L Lu Y K Chen and M L H

Green Chem Commun 1995 1355ndash1356

29 Y-P Sun K F Fu Y Lin and W Huang Acc Chem Res 2002

35 1096ndash1104

30 S B Sinnott J Nanosci Nanotechnol 2002 2113ndash123

31 M S P Shaffer X Fan and A H Windle Carbon 1998 36

1603ndash1612

32 M S P Shaffer and A H Windle Adv Mater 1999 11 937ndash941

33 D E Hill Y Lin A M Rao L F Allard and Y-P Sun

Macromolecules 2002 35 9466ndash9471

34 D Qian E C Dickey R Andrews and T Rantell Appl Phys

Lett 2000 76 2868ndash2870

35 R Andrews D Jacques D Qian and T Rantell Acc Chem Res

2002 35 1008ndash1017

36 R Haggenmueller H H Gommans A G Rinzler J E Fischer

and K I Winey Chem Phys Lett 2000 330 219ndash225

37 P Potschke T D Fornes and D R Paul Polymer 2002 43 3247ndash

3255

38 M Sennett E Welsh J B Wright W Z Li J G Wen and Z F

Ren Appl Phys A 2003 76 111ndash113

39 P M Ajayan O Stephan C Colliex and D Trauth Science 1994

265 1212ndash1214

40 J Sandler M S P Shaffer T Prasse W Bauhofer K Schulte

and H Windle Polymer 1999 40 5967ndash5971

41 A Allaoui S Bai H M Cheng and J B Bai Compos Sci

Technol 2002 62 1993ndash1998

42 X Xu M M Thwe C Shearwood and K Liao Appl Phys Lett

2002 81 2833ndash2835

43 D Puglia L Valentini and J M Kenny J Appl Polymer Sci

2003 88 452ndash458

44 M Cochet W K Maser A M Benito M A Callejas M T

Martınez J-M Benoit J Schreiber and O Chauvet Chem

Commun 2001 1450ndash1451

45 M S P Shaffer and K Koziol Chem Commun 2002 2074ndash

2075

46 C Park Z Ounaies K A Watson R E Crooks J Smith S E

Lowther J W Connell E J Siochi J S Harrison and T L St

Clair Chem Phys Lett 2002 364 303ndash308

47 K Jurewicz S Delpeux V Bertagna F Beguin and E

Frackowiak Chem Phys Lett 2001 347 36ndash40

48 M Hughes G Z Chen M S P Shaffer D J Fray and A H

Windle Chem Mater 2002 14 1610ndash1613

49 A A Mamedov N A Kotov M Prato D M Guldi J P

Wicksted and A Hirsch Nature Mater 2002 1 190ndash194

50 J H Rouse and P T Lillehei Nano Lett 2003 3 59ndash62

51 K Lozano and E V Barrera J Appl Polym Sci 2001 79 125ndash

133

52 L Valentini J Biagiotti J M Kenny and S Santucci Compos

Sci Technol 2003 63 1149ndash1153

53 J Sandler P Werner M S P Shaffer V Demchuk V Altstadt

and A H Windle Composites Part A Appl Sci Manuf 2002 33

1033ndash1039

54 R Czerw Z X Guo P M Ajayan Y P Sun and D L Carroll

Nano Lett 2001 1 423ndash427

55 B P Grady F Pompeo R L Shambaugh and D E Resasco

J Phys Chem B 2002 106 5852ndash5858

56 M Cadek J N Coleman V Barron K Hedicke and W J Blau

Appl Phys Lett 2002 81 5123ndash5125

57 P C P Watts and W K Hsu Nanotechnology 2003 14 L7ndashL10

58 L S Schadler S C Giannaris and P M Ajayan Appl Phys Lett

1998 73 3842ndash3844

59 F H Gojny J Nastalczyk Z Roslaniec and K Schulte Chem

Phys Lett 2003 370 820ndash824

60 P M Ajayan L S Schadler C Giannaris and A Rubio Adv

Mater 2000 12 750ndash753

61 M J Biercuk M C Llaguno M Radosavljevic J K Hyun A T

Johnson and J E Fischer Appl Phys Lett 2002 80 2767ndash2769

62 K Lozano J Bonilla-Rios and E V Barrera J Appl Polym Sci

2001 80 1162ndash1172

63 V Jamieson New Sci 15 Mar 2003 30

64 H Zengin W S Zhou J Y Jin W R Czerw D W Smith

L Echegoyen D L Carroll S H Foulger and J Ballato Adv

Mater 2002 14 1480ndash1483

65 J Deng X Ding W Zhang Y Peng J Wang X Long P Li and

A S C Chan Eur Polym J 2002 38 2497ndash2501

66 G B Blanchet C R Fincher and F Gao Appl Phys Lett 2003

82 1290ndash1292

67 S A Curran P M Ajayan W J Blau D L Carroll J N

Coleman A B Dalton A P Davey A Drury B McCarthy

S Maier and A A Strevens Adv Mater 1998 10 1091ndash1093

68 N S Sariciftci L Smilowitz A J Heeger and F Wudl Science

1992 258 1474ndash1476

69 N Camaioni L Garlaschelli A Geri M Maggini G Possamai

and G Ridolfi J Mater Chem 2002 12 2065ndash2070

70 D B Romero M Carrard W De Heer and L Zuppiroli

Adv Mater 1996 8 899ndash902

71 H Ago K Petritsch M S P Shaffer A H Windle and R H

Friend Adv Mater 1999 11 1281ndash1285

72 E Kymakis and G A J Amaratunga Appl Phys Lett 2002 80

112ndash114

73 Y-C Chen N R Raravikar L S Schadler P M Ajayan Y-P

Zhao T-M Lu G-C Wang and X-C Zhang Appl Phys Lett

2002 81 975ndash977

74 S A OrsquoFlaherty R Murphy S V Hold M Cadek J N

Coleman and W J Blau J Phys Chem B 2003 107 958ndash964

75 A Peigney C Laurent F Dobigeon and A Rousset J Mater

Res 1997 12 613ndash615

76 E Flahaut A Peigney C Laurent C Marliere F Chastel and

A Rousset Acta Mater 2000 48 3803ndash3812

77 A Peigney E Flahaut C Laurent F Chastel and A Rousset

Chem Phys Lett 2002 352 20ndash25

78 J W An and D S Lim J Ceram Process Res 2002 3 201ndash204

79 G-D Zhan J D Kuntz J L Wan and A K Mukherjee Nature

Mater 2003 2 38ndash42

80 G-D Zhan J D Kuntz J E Garay and A K Mukherjee Appl

Phys Lett 2003 83 1228ndash1230

81 T Seeger T Kohler T Frauenheim N Grobert M Ruhle

M Terrones and G Seifert Chem Commun 2002 34ndash35

82 T Seeger G De La Fuente W K Maser A M Benito M A

Callejas and M T Martınez Nanotechnology 2003 14 184ndash187

83 H Han S Vijayalakshmi A Lan Z Iqbal H Grebel E Lalanne

and A M Johnson Appl Phys Lett 2003 82 1458ndash1460

Harris Carbon nanotube composites

42 International Materials Reviews 2004 VOL 49 NO 1

84 P Vincent A Brioude C Journet S Rabaste S T Purcell J Le

Brusq and J C Plenet J Non-Cryst Sol 2002 311 130ndash137

85 J S Sakamoto and B Dunn J Electrochem Soc 2002 149 A26ndashA30

86 R Z Ma J Wu B Q Wei J Liang and D H Wu J Mater Sci

1998 33 5243ndash5246

87 R Andrews D Jacques A M Rao T Rantell F Derbyshire

Y Chen J Chen and R C Haddon Appl Phys Lett 1999 75

1329ndash1331

88 E T Thostenson W Z Li D Z Wang Z F Ren and T W

Chou J Appl Phys 2002 91 6034ndash6037

89 D D L Chung lsquoCarbon fiber compositesrsquo chap 7 1994 Boston

Butterworth-Heinemann

90 S Rawal JOM 2001 53 14ndash17

91 T Kuzumaki K Miyazawa H Ichinose and K Ito J Mater

Res 1998 13 2445ndash2449

92 T Kuzumaki O Ujiie H Ichinose and K Ito Adv Eng Mater

2000 2 416ndash418

93 W X Chen J P Tu L Y Wang H Y Gan Z D Xu and X B

Zhang Carbon 2003 41 215ndash222

94 K Liao and S Li Appl Phys Lett 2001 79 4225ndash4227

95 A Star J F Stoddart D Steuerman M Diehl A Boukai E W

Wong X Yang S W Chung H Choi and J R Heath

Angewandte Chemie-Int Ed 2001 40 1721ndash1725

96 A Star Y Liu K Grant L Ridvan J F Stoddart D W

Steuerman M R Diehl A Boukai and J R Heath

Macromolecules 2003 36 553ndash560

97 P J F Harris S C Tsang J B Claridge and M L H Green

J Chem Soc Faraday Trans 1994 90 2799ndash2802

98 A A Setlur S P Doherty J Y Dai and R P H Chang Appl

Phys Lett 2000 76 3008ndash3010

99 S P Doherty D B Buchholz B J Li and R P H Chang

J Mater Res 2003 18 941ndash949

100 D B Buchholz S P Doherty and R P H Chang Carbon 2003

41 1625ndash1634

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 43

Page 9: Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P. J. F. Harris * Carbon nanotubes are molecular-scale tubes of graphitic carbon with

molecule This allows the preferential transfer of the elec-trons into the electron acceptor molecule while leavingthe holes to be preferentially transported through thepolymer a process known as photoinduced chargetransfer In 1992 photoinduced charge transfer from aconducting polymer to C60 was discovered68 and thishas led to the fabrication of several efficient photo-voltaic systems using fullerenepolymer combinations69

Photovoltaic devices based on nanotubepolymer com-binations have now been fabricated by several groupsand this work will now be summarised

The earliest attempt to fabricate a nanotubepolymerphotovoltaic device appears to be the 1996 work ofRomero et al70 In this work a layer of MWNTs wasdeposited on Teflon and a PPV derivative was cast ontothis with a thin aluminium layer on top The devicedisplayed clear diode behaviour but only a small photo-current was observed when the junction was irradiatedwith light from a HendashNe laser This may have been dueto the relatively thick PPV active layer Ago et alfabricated a nanotube containing photovoltaic deviceby depositing a layer of PPV onto a glass-supportedMWNT film71 The interaction between MWNTs andPPV was investigated by means of photoluminescence(PL) measurements The photoluminescence efficiencyof the PPV was found to be markedly reduced by thepresence of the MWNT layer Since photoluminescenceis a result of radiative recombination a reduction mayimply improved exciton separation which is what onewishes to achieve in a photovoltaic device The devicefabricated by Ago et al is illustrated in Fig 11a Thisshowed a clear diode characteristic in the dark andupon illumination through the aluminium electrode aphotocurrent was observed A quantum efficiency approxi-mately twice that of the standard indiumndashtin oxidedevice was reported

Kymakis and Amaratunga used a slightly differentapproach to prepare a nanotubeconjugated polymerphotovoltaic device72 In this work composite SWNTpoly(3-octylthiophene) films were drop or spin cast froma solution onto indiumndashtin oxide (ITO) and quartz sub-strates to produce the arrangement shown in Fig 11bDevices were also prepared using the pure P3OT poly-mer A considerable enhancement of the photovoltaic

effect was observed with the P3OTSWNTs blend devicecompared with the pure polymer device

Non-linear optical propertiesAs mentioned above in the section lsquoStructure and pro-perties of carbon nanotubesrsquo it has been demonstratedthat suspensions of both SWNTs and MWNTs havenon-linear optical properties Nanotubepolymer com-posites may therefore have applications in optical devicesTo date however there has been little work on theoptical limiting properties of nanotube containing com-posites In one of the first studies in this area Chen et aldemonstrated third order optical non-linearity in a SWNTpolyimide composite73 Recently OrsquoFlaherty et al pre-pared a composite in which MWNTs were dispersed inthe polymer poly(99-di-n-octylfluorenyl-279-diyl) (PFO)74

Suspensions of the nanotubePFO composite in toluenewere prepared and the optical limiting properties of thesolutions were measured using nanosecond laser pulsesMaximum optical limiting was observed at carbonnanotube loadings in excess of 38 wt- relative to thepolymer

Carbon nanotubeceramic compositesCeramics have high stiffnesses and thermal stabilities butrelatively low breaking strengths Incorporating carbonnanotubes into a ceramic matrix might be expected toproduce a composite with both toughness and high tem-perature stability However achieving a homogeneousdispersion of tubes in an oxide with strong bondingbetween tubes and matrix presents rather more of achallenge than incorporating tubes into a polymer Variousapproaches have been adopted for incorporating nano-tubes into oxides such as alumina and silica and thesewill now be summarised

Peigney and co-workers have been among the pioneersin the synthesis of nanotubeoxide composites75ndash77 Theyhave developed an ingenious technique which involvesthe use of catalysts such as iron-containing a-aluminaPassing a CH4ndashH2 mixture over the catalysts resultedin the in situ growth of nanotubes on iron particlesproducing the precursor for a nanotubeoxide compo-site The resulting powders can then be hot pressed toform the final composite material The mechanicalproperties of these composites were initially disappoint-ing72 Thus the fracture strengths of the nanotubeFendashAl2O3 composites were only marginally higher than thatof Al2O3 and were generally markedly lower than thoseof the carbon-free FendashAl2O3 composites The fracturetoughness values were lower than or similar to that ofAl2O3 However SEM observations of composite frac-tures indicate that the nanotube bundles which are veryflexible could dissipate some fracture energy A similarmethod was used by An and Lim to prepare nanotubeAl2O3 composites in this case using C2H2 as the carbonsource78 Nanotube contents of up to 125 wt- wereachieved and a significant reinforcement effect wasreported

Achieving alignment of nanotubes in a ceramic matrixis much more difficult than in a polymer since ceramicsdo not lend themselves so easily to processing Howevera number of groups have shown that aligned nanotubemetal oxide composites can indeed be produced Peigney

ITO5indiumndashtin oxide PPV5poly(p-phenylene vinylene)

P3OT5poly(3-octylthiophene)

11 a Illustration of photovoltaic device fabricated by Ago

et al71 and b photovoltaic device fabricated by

Kymakis and Amaratunga72

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 39

et al exploited the fact that when their grains are main-tained at sufficiently low sizes (less than about 1 mm)many polycrystalline ceramics become superplasticallowing the use of thermomechanical processing forthe forming of these materials A series of cobalt andiron containing oxide catalysts were prepared77 Theprimary grain size in the powders was smaller than100 nm but in some of them the as produced primarygrains were strongly aggregated and required ball millingin order to reduce aggregate size to less than 1 mmNanotubes were then prepared in situ as describedabove and the powders were then extruded at hightemperatures in a graphite die under vacuum The bestresults were achieved with nanotubes incorporated intoFendashAl2O3 and FeCondashMgAl2O4 matrices In both casesthe nanotubes were not damaged by the processing andalignment of the nanotubes was clearly observed Infact the presence of the nanotubes seemed to facilitatethe superplastic forming by inhibiting grain growth andacting as lubricating agents It was demonstrated thatthe aligned nanotubemetal oxide composites showed ananisotropy of electrical conductivity which could beadjusted by controlling the nanotube content

One of the first studies of a SWNTceramic compositewas described by Zhan et al79 This group used thetechnique of spark plasma sintering (SPS) to prepareAl2O3 composites containing 10 vol- SWNTs Theadvantage of spark plasma sintering is that it allowsceramic powders to be annealed at lower temperaturesand for much shorter times than other sintering pro-cesses leading to the fabrication of fully dense ceramicsor composites with nanocrystalline microstructures undermild conditions Unlike some of the other processes usedto prepare nanotubeoxide composites spark plasmasintering does not damage the nanotubes and the nano-tube bundles could be seen to be located mainly at theboundaries of the Al2O3 grains in good contact with theAl2O3 matrix Before sintering the powders had beenmixed by ball milling which produced a reasonablyhomogeneous dispersion again without damaging thenanotubes The fracture toughnesses of the final com-posites were remarkable ndash about twice that of the fullydensified but unreinforced Al2O3 matrix In contrast toprevious work on nanotubeceramic composites thefracture toughness was found to increase with nanotubedensity This is thought to be due to the formation ofentangled networks of single-walled carbon nanotubeswhich may inhibit crack propagation A subsequentstudy of the nanotubeAl2O3 composites showed thatthey had high electrical conductivities80

Various techniques have been used to prepare carbonnanotubesilica composites Seeger et al described amethod which involved preparing a composite gel ofMWNTs with tetraethoxysilane (TEOS) and then sinter-ing this at 1150uC in argon81 A drawback of this tech-nique was that the sintering process led to a partialcrystallisation of the SiO2 resulting in a very inhomo-geneous matrix An alternative method which avoidsthese problems has been developed82 This involvesusing a Nd YAG laser to rapidly heat a TEOSnano-tube mixture resulting in partial melting of the matrixThis produced an amorphous silica matrix with nocrystallisation

A novel nanotubesilica composite material withpotentially useful optical properties has recently been

described by Han et al83 These workers synthesisedfilms consisting of silica spheres with diameters rangingfrom 200 to 650 nm Molybdenumcobalt catalyst parti-cles were then deposited onto the silica spheres and usedto catalyse the growth of SWNTs Since SWNTs arehighly non-linear and fast-switching materials they canbe incorporated into an optically confining environmentto achieve these characteristics at relatively low levels oflaser intensities

There have been a few studies on preparing nanotubecomposites using oxides other than alumina and silicaVincent et al reported the production of nanotubeTiO2

composites using solndashgel methods84 Possible applica-tions of the composites included optical non-linear waveguides and unidimensional conductive films Solndashgelmethods were also used by Sakamoto and Dunn toprepare SWNTV2O5 composites for use as electrodesin secondary lithium batteries85 Little work has beencarried out so far on incorporating nanotubes into non-oxide ceramics In one of the few published studies Maet al fabricated nanotubesilicon carbide composites bymixing SiC nanoparticles with 10 wt- MWNTs andhot pressing at 2000uC86 They found that the nanotubesplayed a strengthening and toughening role in thecomposite both bend strength and fracture toughnessincreased by about 10 compared to monolithic SiCThere have also been some initial studies on carbonnanotubecarbon fibre composites8788

Carbon nanotubemetal compositesComposite materials containing conventional carbonfibres in a metal matrix such as aluminium or mag-nesium are used in a number of specialist applica-tions8990 Such composites combine low density withhigh strength and modulus making them particularlyattractive to the aerospace industry There is growinginterest in the addition of carbon nanotubes to metalmatrices although only a small number of studies havebeen published to date Kuzumaki et al described thepreparation of a nanotubealuminium composite in 199891

Their method involved mixing a nanotube sample with afine aluminium powder mounting the mixture in a6 mm silver sheath and then drawing and heating thewire at 700uC in a vacuum furnace The result was acomposite wire in which the nanotubes were partiallyaligned along the axial direction The tensile strengths ofthe as prepared composite wires were comparable to thatof pure aluminium but the composite wires retained thisstrength after prolonged annealing at 600uC while thestrength of pure aluminium decreased by about 50after this treatment The same group subsequently pre-pared nanotubetitanium composites which showed alarge increase in hardness and modulus compared topure titanium92

Chen et al prepared nanotubenickelphosphorus com-posite coatings on stainless steel substrates using anelectroless plating method93 The coatings exhibited higherwear resistance and lower friction coefficients than SiCNiP and graphiteNiP composite coatings The samegroup prepared nanotubecopper composites by a powdermetallurgy technique and showed that the presence ofthe nanotubes again resulted in improved wear rates andfriction coefficients These results suggest that nanotube

Harris Carbon nanotube composites

40 International Materials Reviews 2004 VOL 49 NO 1

metal composites may have potential in tribologicalapplications

DiscussionThe subject of carbon nanotube composites has grownrapidly over the past 5 years or so into a vibrant andexciting area of research It is important to recognisehowever that this is still a new subject and the workcarried out so far may have hardly scratched the surfaceof what is possible For example the great majority ofthe studies that have been published to date haveinvolved the use of polymer matrices there is great scopefor more work on incorporating nanotubes into cera-mics metals and other forms of carbon Also most ofthe work up to now has been aimed at exploiting themechanical properties of carbon nanotubes rather thantheir electronic or optical properties although interest inthose is growing

When considering the mechanical performance ofcarbon nanotube composites it is clear that these dependstrongly on whether arc grown tubes or catalyticallyproduced tubes have been used In considering thecharacteristics of the two types of tube it is instructive tocompare the work of Qian et al34 with that of Watts andHsu57 Both incorporated the nanotubes into polymersand then used TEM to characterise the composites andstudy their fracture behaviour Qian et al used cataly-tically produced tubes and these were found to bewell dispersed throughout the polymer matrix with noclustering together of tubes However during compositefracture the tubes were often observed to break sug-gesting that they have relatively poor strength Wattsand Hsu on the other hand employed much moreperfect nanotubes produced by arc evaporation andTEM studies revealed superior mechanical propertieswith no observed breaking However the nanotubestended to cluster together so that during crack growththe tubes could quite easily slide past one another Thuswhile it is clear that the best mechanical performancewill be achieved with arc grown nanotubes it is impor-tant that the tubes be well dispersed throughout thematrix in order fully to exploit their properties

Equally important is achieving good bonding betweennanotubes and matrix Many studies have shown that thisinterfacial bonding can be enhanced by functionalising thenanotubes before embedding them in the matrix Theeffect of this can be seen in Fig 10 from the work ofGojny et al59 where bonding between MWNTs and anepoxy matrix is so strong that the outer shells of the tuberemain embedded in the matrix following pull-out Apossible disadvantage of functionalisation is that it couldweaken the nanotubes by disrupting the perfect cylindricalstructure This might be expected to be a particularproblem with single-walled tubes However functionali-sation of the nanotubes may not be necessary to producebonding with a polymer matrix Liao and Li havemodelled the interfacial characteristics of a nanotubepolystyrene composite assuming no covalent bondingbetween the tubes and the matrix94 It was found thatelectrostatic and van der Waals interactions contributedsignificantly to the interfacial stress transfer A nanotubepull-out simulation suggested that the interfacial shearstress of the nanotubepolystyrene system is about160 MPa significantly higher than for most carbon fibrereinforced polymer composite systems It is also worth

mentioning recent work by Star and co-workers9596

in which it was demonstrated that poly(5-alkoxy-m-phenylenevinylene)-co-[(25-dioctyloxy-p-phenylene)vinylene] (PAmPV) derivatives could be induced towrap around SWNTs forming strong non-covalentbonds

The quality of the nanotubes and the nature of theinterfacial bonding also has a strong influence on theelectrical properties of nanotubepolymer compositesArc grown tubes can have resistivities lower than that ofsingle crystal graphite but as already discussed theirelectronic properties depend on their precise structureAt present we do not have a reliable method of separat-ing metallic and semiconducting tubes so it is notpossible to exploit their full potential Catalytically pro-duced tubes are likely to be poorer conductors thanmetallic arc-grown tubes owing to their relatively dis-ordered structure Nevertheless catalytically grown tubesare sufficiently conducting to be useful in some com-mercial applications such as electrically conductive paintsThe influence of interfacial bonding on electronic pro-perties was demonstrated in the work of Cochet et al44

These workers prepared a MWNTpolyaniline com-posite by in situ polymerisation and found it to haveexcellent electrical transport properties This was attrib-uted to strong interaction between tubes and polymer

As has been shown in this review the possible appli-cations of carbon nanotube composites range widelyfrom electrically conductive paints to optical devicesOnly a relatively small number of these new materialshave been used commercially however and there arestill many problems to overcome before the full potentialof nanotube containing composites can be realised Notleast of these problems is the high cost of good qualitycarbon nanotubes Currently the commercial price ofunpurified arc-grownMWNTs is approximately $10000gwhile purified MWNTs cost around $30000g Cataly-tically grown MWNTs are much cheaper but as empha-sised throughout this review have generally inferiorproperties As far as SWNTs are concerned unpurifiedsamples cost around $5000g while purified SWNTs are$50000g As a comparison the price of gold is cur-rently about $1200g The development of new techni-ques for the low cost synthesis of high quality nanotubesis clearly essential This might be achieved throughimproved catalytic processes and there is currently alarge amount of work being carried out towards thisend Alternatively it may be possible to develop newnon-catalytic techniques for the synthesis of fullerene-related nanotubes In this connection it is interesting tonote that nanotubes can be produced by the high tem-perature heat treatment of conventional carbons such asmicroporous non-graphitising carbons9798 or carbonblack99100 Further work in this area would be of value

Acknowledgement

I thank Milo Shaffer for helpful comments

References1 P J F Harris lsquoCarbon nanotubes and related structures ndash new

materials for the twenty-first centuryrsquo 1999 Cambridge

Cambridge University Press

2 S Iijima Nature 1991 354 56ndash58

3 W Kratschmer L D Lamb K Fostiropoulos and D R

Huffman Nature 1990 347 354ndash358

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 41

4 S Iijima and T Ichihashi Nature 1993 363 603ndash605

5 D S Bethune C H Kiang M S De Vries G Gorman R Savoy

J Vasquez and R Beyers Nature 1993 363 605ndash607

6 H Dai A G Rinzler P Nikolaev A Thess D T Colbert and

R E Smalley Chem Phys Lett 1996 260 471ndash475

7 M M J Treacy T W Ebbesen and J M Gibson Nature 1996

381 678ndash680

8 E W Wong P E Sheehan and C M Lieber Science 1997 277

1971ndash1975

9 M R Falvo G J Clary R M Taylor V Chi F P Brooks

S Washburn and R Superfine Nature 1997 389 582ndash584

10 K Anazawa K Shimotani C Manabe H Watanabe and M

Shimizu Appl Phys Lett 2002 81 739ndash741

11 P Kim L Shi A Majumdar and P L McEuen Phys Rev Lett

2001 87 paper 215502

12 E T Thostenson Z F Ren and T W Chou Compos Sci

Technol 2001 61 1899ndash1912

13 K T Lau and D Hui Composites Part B Engineering 2002 33

263ndash277

14 Y Huh J Y Lee J Cheon Y K Hong J Y Koo T J Lee and

C J Lee J Mater Chem 2003 13 2297ndash2300

15 M Endo T Koyama and Y Hishiyama Jpn J Appl Phys 1976

15 2073ndash2076

16 S C Lyu B C Liu T J Lee Z Y Liu C W Yang C Y Park

and C J Lee Chem Commun 2003 734ndash735

17 M-F Yu O Lourie M J Dyer K Moloni T F Kelly and R S

Ruoff Science 2000 287 637ndash640

18 P Poncharal Z L Wang D Ugarte and W A De Heer Science

1999 283 1513ndash1516

19 J-P Salvetat A J Kulik J-M Bonard G A D Briggs

T Stockli K Metenier S Bonnamy F Beguin N A Burnham

and L Forro Adv Mater 1999 11 161ndash165

20 M-F Yu B S Files S Arepalli and R S Ruoff Phys Rev Lett

2000 84 5552ndash5555

21 M Terrones Int Mater Rev to be published

22 J W Mintmire B I Dunlap and C T White Phys Rev Lett

1992 68 631ndash634

23 R Saito M Fujita G Dresselhaus and M S Dresselhaus Phys

Rev B 1992 46 1804ndash1811

24 N Hamada S Sawada and A Oshiyama Phys Rev Lett 1992

68 1579ndash1581

25 T W Ebbesen H J Lezec H Hiura J W Bennett H F Ghaemi

and T Thio Nature 1996 382 54ndash56

26 J E Riggs D B Walker D L Carroll and Y-P Sun J Phys

Chem B 2000 104 7071ndash7076

27 S C Tsang Y K Chen P J F Harris and M L H Green

Nature 1994 372 159ndash162

28 R M Lago S C Tsang K L Lu Y K Chen and M L H

Green Chem Commun 1995 1355ndash1356

29 Y-P Sun K F Fu Y Lin and W Huang Acc Chem Res 2002

35 1096ndash1104

30 S B Sinnott J Nanosci Nanotechnol 2002 2113ndash123

31 M S P Shaffer X Fan and A H Windle Carbon 1998 36

1603ndash1612

32 M S P Shaffer and A H Windle Adv Mater 1999 11 937ndash941

33 D E Hill Y Lin A M Rao L F Allard and Y-P Sun

Macromolecules 2002 35 9466ndash9471

34 D Qian E C Dickey R Andrews and T Rantell Appl Phys

Lett 2000 76 2868ndash2870

35 R Andrews D Jacques D Qian and T Rantell Acc Chem Res

2002 35 1008ndash1017

36 R Haggenmueller H H Gommans A G Rinzler J E Fischer

and K I Winey Chem Phys Lett 2000 330 219ndash225

37 P Potschke T D Fornes and D R Paul Polymer 2002 43 3247ndash

3255

38 M Sennett E Welsh J B Wright W Z Li J G Wen and Z F

Ren Appl Phys A 2003 76 111ndash113

39 P M Ajayan O Stephan C Colliex and D Trauth Science 1994

265 1212ndash1214

40 J Sandler M S P Shaffer T Prasse W Bauhofer K Schulte

and H Windle Polymer 1999 40 5967ndash5971

41 A Allaoui S Bai H M Cheng and J B Bai Compos Sci

Technol 2002 62 1993ndash1998

42 X Xu M M Thwe C Shearwood and K Liao Appl Phys Lett

2002 81 2833ndash2835

43 D Puglia L Valentini and J M Kenny J Appl Polymer Sci

2003 88 452ndash458

44 M Cochet W K Maser A M Benito M A Callejas M T

Martınez J-M Benoit J Schreiber and O Chauvet Chem

Commun 2001 1450ndash1451

45 M S P Shaffer and K Koziol Chem Commun 2002 2074ndash

2075

46 C Park Z Ounaies K A Watson R E Crooks J Smith S E

Lowther J W Connell E J Siochi J S Harrison and T L St

Clair Chem Phys Lett 2002 364 303ndash308

47 K Jurewicz S Delpeux V Bertagna F Beguin and E

Frackowiak Chem Phys Lett 2001 347 36ndash40

48 M Hughes G Z Chen M S P Shaffer D J Fray and A H

Windle Chem Mater 2002 14 1610ndash1613

49 A A Mamedov N A Kotov M Prato D M Guldi J P

Wicksted and A Hirsch Nature Mater 2002 1 190ndash194

50 J H Rouse and P T Lillehei Nano Lett 2003 3 59ndash62

51 K Lozano and E V Barrera J Appl Polym Sci 2001 79 125ndash

133

52 L Valentini J Biagiotti J M Kenny and S Santucci Compos

Sci Technol 2003 63 1149ndash1153

53 J Sandler P Werner M S P Shaffer V Demchuk V Altstadt

and A H Windle Composites Part A Appl Sci Manuf 2002 33

1033ndash1039

54 R Czerw Z X Guo P M Ajayan Y P Sun and D L Carroll

Nano Lett 2001 1 423ndash427

55 B P Grady F Pompeo R L Shambaugh and D E Resasco

J Phys Chem B 2002 106 5852ndash5858

56 M Cadek J N Coleman V Barron K Hedicke and W J Blau

Appl Phys Lett 2002 81 5123ndash5125

57 P C P Watts and W K Hsu Nanotechnology 2003 14 L7ndashL10

58 L S Schadler S C Giannaris and P M Ajayan Appl Phys Lett

1998 73 3842ndash3844

59 F H Gojny J Nastalczyk Z Roslaniec and K Schulte Chem

Phys Lett 2003 370 820ndash824

60 P M Ajayan L S Schadler C Giannaris and A Rubio Adv

Mater 2000 12 750ndash753

61 M J Biercuk M C Llaguno M Radosavljevic J K Hyun A T

Johnson and J E Fischer Appl Phys Lett 2002 80 2767ndash2769

62 K Lozano J Bonilla-Rios and E V Barrera J Appl Polym Sci

2001 80 1162ndash1172

63 V Jamieson New Sci 15 Mar 2003 30

64 H Zengin W S Zhou J Y Jin W R Czerw D W Smith

L Echegoyen D L Carroll S H Foulger and J Ballato Adv

Mater 2002 14 1480ndash1483

65 J Deng X Ding W Zhang Y Peng J Wang X Long P Li and

A S C Chan Eur Polym J 2002 38 2497ndash2501

66 G B Blanchet C R Fincher and F Gao Appl Phys Lett 2003

82 1290ndash1292

67 S A Curran P M Ajayan W J Blau D L Carroll J N

Coleman A B Dalton A P Davey A Drury B McCarthy

S Maier and A A Strevens Adv Mater 1998 10 1091ndash1093

68 N S Sariciftci L Smilowitz A J Heeger and F Wudl Science

1992 258 1474ndash1476

69 N Camaioni L Garlaschelli A Geri M Maggini G Possamai

and G Ridolfi J Mater Chem 2002 12 2065ndash2070

70 D B Romero M Carrard W De Heer and L Zuppiroli

Adv Mater 1996 8 899ndash902

71 H Ago K Petritsch M S P Shaffer A H Windle and R H

Friend Adv Mater 1999 11 1281ndash1285

72 E Kymakis and G A J Amaratunga Appl Phys Lett 2002 80

112ndash114

73 Y-C Chen N R Raravikar L S Schadler P M Ajayan Y-P

Zhao T-M Lu G-C Wang and X-C Zhang Appl Phys Lett

2002 81 975ndash977

74 S A OrsquoFlaherty R Murphy S V Hold M Cadek J N

Coleman and W J Blau J Phys Chem B 2003 107 958ndash964

75 A Peigney C Laurent F Dobigeon and A Rousset J Mater

Res 1997 12 613ndash615

76 E Flahaut A Peigney C Laurent C Marliere F Chastel and

A Rousset Acta Mater 2000 48 3803ndash3812

77 A Peigney E Flahaut C Laurent F Chastel and A Rousset

Chem Phys Lett 2002 352 20ndash25

78 J W An and D S Lim J Ceram Process Res 2002 3 201ndash204

79 G-D Zhan J D Kuntz J L Wan and A K Mukherjee Nature

Mater 2003 2 38ndash42

80 G-D Zhan J D Kuntz J E Garay and A K Mukherjee Appl

Phys Lett 2003 83 1228ndash1230

81 T Seeger T Kohler T Frauenheim N Grobert M Ruhle

M Terrones and G Seifert Chem Commun 2002 34ndash35

82 T Seeger G De La Fuente W K Maser A M Benito M A

Callejas and M T Martınez Nanotechnology 2003 14 184ndash187

83 H Han S Vijayalakshmi A Lan Z Iqbal H Grebel E Lalanne

and A M Johnson Appl Phys Lett 2003 82 1458ndash1460

Harris Carbon nanotube composites

42 International Materials Reviews 2004 VOL 49 NO 1

84 P Vincent A Brioude C Journet S Rabaste S T Purcell J Le

Brusq and J C Plenet J Non-Cryst Sol 2002 311 130ndash137

85 J S Sakamoto and B Dunn J Electrochem Soc 2002 149 A26ndashA30

86 R Z Ma J Wu B Q Wei J Liang and D H Wu J Mater Sci

1998 33 5243ndash5246

87 R Andrews D Jacques A M Rao T Rantell F Derbyshire

Y Chen J Chen and R C Haddon Appl Phys Lett 1999 75

1329ndash1331

88 E T Thostenson W Z Li D Z Wang Z F Ren and T W

Chou J Appl Phys 2002 91 6034ndash6037

89 D D L Chung lsquoCarbon fiber compositesrsquo chap 7 1994 Boston

Butterworth-Heinemann

90 S Rawal JOM 2001 53 14ndash17

91 T Kuzumaki K Miyazawa H Ichinose and K Ito J Mater

Res 1998 13 2445ndash2449

92 T Kuzumaki O Ujiie H Ichinose and K Ito Adv Eng Mater

2000 2 416ndash418

93 W X Chen J P Tu L Y Wang H Y Gan Z D Xu and X B

Zhang Carbon 2003 41 215ndash222

94 K Liao and S Li Appl Phys Lett 2001 79 4225ndash4227

95 A Star J F Stoddart D Steuerman M Diehl A Boukai E W

Wong X Yang S W Chung H Choi and J R Heath

Angewandte Chemie-Int Ed 2001 40 1721ndash1725

96 A Star Y Liu K Grant L Ridvan J F Stoddart D W

Steuerman M R Diehl A Boukai and J R Heath

Macromolecules 2003 36 553ndash560

97 P J F Harris S C Tsang J B Claridge and M L H Green

J Chem Soc Faraday Trans 1994 90 2799ndash2802

98 A A Setlur S P Doherty J Y Dai and R P H Chang Appl

Phys Lett 2000 76 3008ndash3010

99 S P Doherty D B Buchholz B J Li and R P H Chang

J Mater Res 2003 18 941ndash949

100 D B Buchholz S P Doherty and R P H Chang Carbon 2003

41 1625ndash1634

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 43

Page 10: Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P. J. F. Harris * Carbon nanotubes are molecular-scale tubes of graphitic carbon with

et al exploited the fact that when their grains are main-tained at sufficiently low sizes (less than about 1 mm)many polycrystalline ceramics become superplasticallowing the use of thermomechanical processing forthe forming of these materials A series of cobalt andiron containing oxide catalysts were prepared77 Theprimary grain size in the powders was smaller than100 nm but in some of them the as produced primarygrains were strongly aggregated and required ball millingin order to reduce aggregate size to less than 1 mmNanotubes were then prepared in situ as describedabove and the powders were then extruded at hightemperatures in a graphite die under vacuum The bestresults were achieved with nanotubes incorporated intoFendashAl2O3 and FeCondashMgAl2O4 matrices In both casesthe nanotubes were not damaged by the processing andalignment of the nanotubes was clearly observed Infact the presence of the nanotubes seemed to facilitatethe superplastic forming by inhibiting grain growth andacting as lubricating agents It was demonstrated thatthe aligned nanotubemetal oxide composites showed ananisotropy of electrical conductivity which could beadjusted by controlling the nanotube content

One of the first studies of a SWNTceramic compositewas described by Zhan et al79 This group used thetechnique of spark plasma sintering (SPS) to prepareAl2O3 composites containing 10 vol- SWNTs Theadvantage of spark plasma sintering is that it allowsceramic powders to be annealed at lower temperaturesand for much shorter times than other sintering pro-cesses leading to the fabrication of fully dense ceramicsor composites with nanocrystalline microstructures undermild conditions Unlike some of the other processes usedto prepare nanotubeoxide composites spark plasmasintering does not damage the nanotubes and the nano-tube bundles could be seen to be located mainly at theboundaries of the Al2O3 grains in good contact with theAl2O3 matrix Before sintering the powders had beenmixed by ball milling which produced a reasonablyhomogeneous dispersion again without damaging thenanotubes The fracture toughnesses of the final com-posites were remarkable ndash about twice that of the fullydensified but unreinforced Al2O3 matrix In contrast toprevious work on nanotubeceramic composites thefracture toughness was found to increase with nanotubedensity This is thought to be due to the formation ofentangled networks of single-walled carbon nanotubeswhich may inhibit crack propagation A subsequentstudy of the nanotubeAl2O3 composites showed thatthey had high electrical conductivities80

Various techniques have been used to prepare carbonnanotubesilica composites Seeger et al described amethod which involved preparing a composite gel ofMWNTs with tetraethoxysilane (TEOS) and then sinter-ing this at 1150uC in argon81 A drawback of this tech-nique was that the sintering process led to a partialcrystallisation of the SiO2 resulting in a very inhomo-geneous matrix An alternative method which avoidsthese problems has been developed82 This involvesusing a Nd YAG laser to rapidly heat a TEOSnano-tube mixture resulting in partial melting of the matrixThis produced an amorphous silica matrix with nocrystallisation

A novel nanotubesilica composite material withpotentially useful optical properties has recently been

described by Han et al83 These workers synthesisedfilms consisting of silica spheres with diameters rangingfrom 200 to 650 nm Molybdenumcobalt catalyst parti-cles were then deposited onto the silica spheres and usedto catalyse the growth of SWNTs Since SWNTs arehighly non-linear and fast-switching materials they canbe incorporated into an optically confining environmentto achieve these characteristics at relatively low levels oflaser intensities

There have been a few studies on preparing nanotubecomposites using oxides other than alumina and silicaVincent et al reported the production of nanotubeTiO2

composites using solndashgel methods84 Possible applica-tions of the composites included optical non-linear waveguides and unidimensional conductive films Solndashgelmethods were also used by Sakamoto and Dunn toprepare SWNTV2O5 composites for use as electrodesin secondary lithium batteries85 Little work has beencarried out so far on incorporating nanotubes into non-oxide ceramics In one of the few published studies Maet al fabricated nanotubesilicon carbide composites bymixing SiC nanoparticles with 10 wt- MWNTs andhot pressing at 2000uC86 They found that the nanotubesplayed a strengthening and toughening role in thecomposite both bend strength and fracture toughnessincreased by about 10 compared to monolithic SiCThere have also been some initial studies on carbonnanotubecarbon fibre composites8788

Carbon nanotubemetal compositesComposite materials containing conventional carbonfibres in a metal matrix such as aluminium or mag-nesium are used in a number of specialist applica-tions8990 Such composites combine low density withhigh strength and modulus making them particularlyattractive to the aerospace industry There is growinginterest in the addition of carbon nanotubes to metalmatrices although only a small number of studies havebeen published to date Kuzumaki et al described thepreparation of a nanotubealuminium composite in 199891

Their method involved mixing a nanotube sample with afine aluminium powder mounting the mixture in a6 mm silver sheath and then drawing and heating thewire at 700uC in a vacuum furnace The result was acomposite wire in which the nanotubes were partiallyaligned along the axial direction The tensile strengths ofthe as prepared composite wires were comparable to thatof pure aluminium but the composite wires retained thisstrength after prolonged annealing at 600uC while thestrength of pure aluminium decreased by about 50after this treatment The same group subsequently pre-pared nanotubetitanium composites which showed alarge increase in hardness and modulus compared topure titanium92

Chen et al prepared nanotubenickelphosphorus com-posite coatings on stainless steel substrates using anelectroless plating method93 The coatings exhibited higherwear resistance and lower friction coefficients than SiCNiP and graphiteNiP composite coatings The samegroup prepared nanotubecopper composites by a powdermetallurgy technique and showed that the presence ofthe nanotubes again resulted in improved wear rates andfriction coefficients These results suggest that nanotube

Harris Carbon nanotube composites

40 International Materials Reviews 2004 VOL 49 NO 1

metal composites may have potential in tribologicalapplications

DiscussionThe subject of carbon nanotube composites has grownrapidly over the past 5 years or so into a vibrant andexciting area of research It is important to recognisehowever that this is still a new subject and the workcarried out so far may have hardly scratched the surfaceof what is possible For example the great majority ofthe studies that have been published to date haveinvolved the use of polymer matrices there is great scopefor more work on incorporating nanotubes into cera-mics metals and other forms of carbon Also most ofthe work up to now has been aimed at exploiting themechanical properties of carbon nanotubes rather thantheir electronic or optical properties although interest inthose is growing

When considering the mechanical performance ofcarbon nanotube composites it is clear that these dependstrongly on whether arc grown tubes or catalyticallyproduced tubes have been used In considering thecharacteristics of the two types of tube it is instructive tocompare the work of Qian et al34 with that of Watts andHsu57 Both incorporated the nanotubes into polymersand then used TEM to characterise the composites andstudy their fracture behaviour Qian et al used cataly-tically produced tubes and these were found to bewell dispersed throughout the polymer matrix with noclustering together of tubes However during compositefracture the tubes were often observed to break sug-gesting that they have relatively poor strength Wattsand Hsu on the other hand employed much moreperfect nanotubes produced by arc evaporation andTEM studies revealed superior mechanical propertieswith no observed breaking However the nanotubestended to cluster together so that during crack growththe tubes could quite easily slide past one another Thuswhile it is clear that the best mechanical performancewill be achieved with arc grown nanotubes it is impor-tant that the tubes be well dispersed throughout thematrix in order fully to exploit their properties

Equally important is achieving good bonding betweennanotubes and matrix Many studies have shown that thisinterfacial bonding can be enhanced by functionalising thenanotubes before embedding them in the matrix Theeffect of this can be seen in Fig 10 from the work ofGojny et al59 where bonding between MWNTs and anepoxy matrix is so strong that the outer shells of the tuberemain embedded in the matrix following pull-out Apossible disadvantage of functionalisation is that it couldweaken the nanotubes by disrupting the perfect cylindricalstructure This might be expected to be a particularproblem with single-walled tubes However functionali-sation of the nanotubes may not be necessary to producebonding with a polymer matrix Liao and Li havemodelled the interfacial characteristics of a nanotubepolystyrene composite assuming no covalent bondingbetween the tubes and the matrix94 It was found thatelectrostatic and van der Waals interactions contributedsignificantly to the interfacial stress transfer A nanotubepull-out simulation suggested that the interfacial shearstress of the nanotubepolystyrene system is about160 MPa significantly higher than for most carbon fibrereinforced polymer composite systems It is also worth

mentioning recent work by Star and co-workers9596

in which it was demonstrated that poly(5-alkoxy-m-phenylenevinylene)-co-[(25-dioctyloxy-p-phenylene)vinylene] (PAmPV) derivatives could be induced towrap around SWNTs forming strong non-covalentbonds

The quality of the nanotubes and the nature of theinterfacial bonding also has a strong influence on theelectrical properties of nanotubepolymer compositesArc grown tubes can have resistivities lower than that ofsingle crystal graphite but as already discussed theirelectronic properties depend on their precise structureAt present we do not have a reliable method of separat-ing metallic and semiconducting tubes so it is notpossible to exploit their full potential Catalytically pro-duced tubes are likely to be poorer conductors thanmetallic arc-grown tubes owing to their relatively dis-ordered structure Nevertheless catalytically grown tubesare sufficiently conducting to be useful in some com-mercial applications such as electrically conductive paintsThe influence of interfacial bonding on electronic pro-perties was demonstrated in the work of Cochet et al44

These workers prepared a MWNTpolyaniline com-posite by in situ polymerisation and found it to haveexcellent electrical transport properties This was attrib-uted to strong interaction between tubes and polymer

As has been shown in this review the possible appli-cations of carbon nanotube composites range widelyfrom electrically conductive paints to optical devicesOnly a relatively small number of these new materialshave been used commercially however and there arestill many problems to overcome before the full potentialof nanotube containing composites can be realised Notleast of these problems is the high cost of good qualitycarbon nanotubes Currently the commercial price ofunpurified arc-grownMWNTs is approximately $10000gwhile purified MWNTs cost around $30000g Cataly-tically grown MWNTs are much cheaper but as empha-sised throughout this review have generally inferiorproperties As far as SWNTs are concerned unpurifiedsamples cost around $5000g while purified SWNTs are$50000g As a comparison the price of gold is cur-rently about $1200g The development of new techni-ques for the low cost synthesis of high quality nanotubesis clearly essential This might be achieved throughimproved catalytic processes and there is currently alarge amount of work being carried out towards thisend Alternatively it may be possible to develop newnon-catalytic techniques for the synthesis of fullerene-related nanotubes In this connection it is interesting tonote that nanotubes can be produced by the high tem-perature heat treatment of conventional carbons such asmicroporous non-graphitising carbons9798 or carbonblack99100 Further work in this area would be of value

Acknowledgement

I thank Milo Shaffer for helpful comments

References1 P J F Harris lsquoCarbon nanotubes and related structures ndash new

materials for the twenty-first centuryrsquo 1999 Cambridge

Cambridge University Press

2 S Iijima Nature 1991 354 56ndash58

3 W Kratschmer L D Lamb K Fostiropoulos and D R

Huffman Nature 1990 347 354ndash358

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 41

4 S Iijima and T Ichihashi Nature 1993 363 603ndash605

5 D S Bethune C H Kiang M S De Vries G Gorman R Savoy

J Vasquez and R Beyers Nature 1993 363 605ndash607

6 H Dai A G Rinzler P Nikolaev A Thess D T Colbert and

R E Smalley Chem Phys Lett 1996 260 471ndash475

7 M M J Treacy T W Ebbesen and J M Gibson Nature 1996

381 678ndash680

8 E W Wong P E Sheehan and C M Lieber Science 1997 277

1971ndash1975

9 M R Falvo G J Clary R M Taylor V Chi F P Brooks

S Washburn and R Superfine Nature 1997 389 582ndash584

10 K Anazawa K Shimotani C Manabe H Watanabe and M

Shimizu Appl Phys Lett 2002 81 739ndash741

11 P Kim L Shi A Majumdar and P L McEuen Phys Rev Lett

2001 87 paper 215502

12 E T Thostenson Z F Ren and T W Chou Compos Sci

Technol 2001 61 1899ndash1912

13 K T Lau and D Hui Composites Part B Engineering 2002 33

263ndash277

14 Y Huh J Y Lee J Cheon Y K Hong J Y Koo T J Lee and

C J Lee J Mater Chem 2003 13 2297ndash2300

15 M Endo T Koyama and Y Hishiyama Jpn J Appl Phys 1976

15 2073ndash2076

16 S C Lyu B C Liu T J Lee Z Y Liu C W Yang C Y Park

and C J Lee Chem Commun 2003 734ndash735

17 M-F Yu O Lourie M J Dyer K Moloni T F Kelly and R S

Ruoff Science 2000 287 637ndash640

18 P Poncharal Z L Wang D Ugarte and W A De Heer Science

1999 283 1513ndash1516

19 J-P Salvetat A J Kulik J-M Bonard G A D Briggs

T Stockli K Metenier S Bonnamy F Beguin N A Burnham

and L Forro Adv Mater 1999 11 161ndash165

20 M-F Yu B S Files S Arepalli and R S Ruoff Phys Rev Lett

2000 84 5552ndash5555

21 M Terrones Int Mater Rev to be published

22 J W Mintmire B I Dunlap and C T White Phys Rev Lett

1992 68 631ndash634

23 R Saito M Fujita G Dresselhaus and M S Dresselhaus Phys

Rev B 1992 46 1804ndash1811

24 N Hamada S Sawada and A Oshiyama Phys Rev Lett 1992

68 1579ndash1581

25 T W Ebbesen H J Lezec H Hiura J W Bennett H F Ghaemi

and T Thio Nature 1996 382 54ndash56

26 J E Riggs D B Walker D L Carroll and Y-P Sun J Phys

Chem B 2000 104 7071ndash7076

27 S C Tsang Y K Chen P J F Harris and M L H Green

Nature 1994 372 159ndash162

28 R M Lago S C Tsang K L Lu Y K Chen and M L H

Green Chem Commun 1995 1355ndash1356

29 Y-P Sun K F Fu Y Lin and W Huang Acc Chem Res 2002

35 1096ndash1104

30 S B Sinnott J Nanosci Nanotechnol 2002 2113ndash123

31 M S P Shaffer X Fan and A H Windle Carbon 1998 36

1603ndash1612

32 M S P Shaffer and A H Windle Adv Mater 1999 11 937ndash941

33 D E Hill Y Lin A M Rao L F Allard and Y-P Sun

Macromolecules 2002 35 9466ndash9471

34 D Qian E C Dickey R Andrews and T Rantell Appl Phys

Lett 2000 76 2868ndash2870

35 R Andrews D Jacques D Qian and T Rantell Acc Chem Res

2002 35 1008ndash1017

36 R Haggenmueller H H Gommans A G Rinzler J E Fischer

and K I Winey Chem Phys Lett 2000 330 219ndash225

37 P Potschke T D Fornes and D R Paul Polymer 2002 43 3247ndash

3255

38 M Sennett E Welsh J B Wright W Z Li J G Wen and Z F

Ren Appl Phys A 2003 76 111ndash113

39 P M Ajayan O Stephan C Colliex and D Trauth Science 1994

265 1212ndash1214

40 J Sandler M S P Shaffer T Prasse W Bauhofer K Schulte

and H Windle Polymer 1999 40 5967ndash5971

41 A Allaoui S Bai H M Cheng and J B Bai Compos Sci

Technol 2002 62 1993ndash1998

42 X Xu M M Thwe C Shearwood and K Liao Appl Phys Lett

2002 81 2833ndash2835

43 D Puglia L Valentini and J M Kenny J Appl Polymer Sci

2003 88 452ndash458

44 M Cochet W K Maser A M Benito M A Callejas M T

Martınez J-M Benoit J Schreiber and O Chauvet Chem

Commun 2001 1450ndash1451

45 M S P Shaffer and K Koziol Chem Commun 2002 2074ndash

2075

46 C Park Z Ounaies K A Watson R E Crooks J Smith S E

Lowther J W Connell E J Siochi J S Harrison and T L St

Clair Chem Phys Lett 2002 364 303ndash308

47 K Jurewicz S Delpeux V Bertagna F Beguin and E

Frackowiak Chem Phys Lett 2001 347 36ndash40

48 M Hughes G Z Chen M S P Shaffer D J Fray and A H

Windle Chem Mater 2002 14 1610ndash1613

49 A A Mamedov N A Kotov M Prato D M Guldi J P

Wicksted and A Hirsch Nature Mater 2002 1 190ndash194

50 J H Rouse and P T Lillehei Nano Lett 2003 3 59ndash62

51 K Lozano and E V Barrera J Appl Polym Sci 2001 79 125ndash

133

52 L Valentini J Biagiotti J M Kenny and S Santucci Compos

Sci Technol 2003 63 1149ndash1153

53 J Sandler P Werner M S P Shaffer V Demchuk V Altstadt

and A H Windle Composites Part A Appl Sci Manuf 2002 33

1033ndash1039

54 R Czerw Z X Guo P M Ajayan Y P Sun and D L Carroll

Nano Lett 2001 1 423ndash427

55 B P Grady F Pompeo R L Shambaugh and D E Resasco

J Phys Chem B 2002 106 5852ndash5858

56 M Cadek J N Coleman V Barron K Hedicke and W J Blau

Appl Phys Lett 2002 81 5123ndash5125

57 P C P Watts and W K Hsu Nanotechnology 2003 14 L7ndashL10

58 L S Schadler S C Giannaris and P M Ajayan Appl Phys Lett

1998 73 3842ndash3844

59 F H Gojny J Nastalczyk Z Roslaniec and K Schulte Chem

Phys Lett 2003 370 820ndash824

60 P M Ajayan L S Schadler C Giannaris and A Rubio Adv

Mater 2000 12 750ndash753

61 M J Biercuk M C Llaguno M Radosavljevic J K Hyun A T

Johnson and J E Fischer Appl Phys Lett 2002 80 2767ndash2769

62 K Lozano J Bonilla-Rios and E V Barrera J Appl Polym Sci

2001 80 1162ndash1172

63 V Jamieson New Sci 15 Mar 2003 30

64 H Zengin W S Zhou J Y Jin W R Czerw D W Smith

L Echegoyen D L Carroll S H Foulger and J Ballato Adv

Mater 2002 14 1480ndash1483

65 J Deng X Ding W Zhang Y Peng J Wang X Long P Li and

A S C Chan Eur Polym J 2002 38 2497ndash2501

66 G B Blanchet C R Fincher and F Gao Appl Phys Lett 2003

82 1290ndash1292

67 S A Curran P M Ajayan W J Blau D L Carroll J N

Coleman A B Dalton A P Davey A Drury B McCarthy

S Maier and A A Strevens Adv Mater 1998 10 1091ndash1093

68 N S Sariciftci L Smilowitz A J Heeger and F Wudl Science

1992 258 1474ndash1476

69 N Camaioni L Garlaschelli A Geri M Maggini G Possamai

and G Ridolfi J Mater Chem 2002 12 2065ndash2070

70 D B Romero M Carrard W De Heer and L Zuppiroli

Adv Mater 1996 8 899ndash902

71 H Ago K Petritsch M S P Shaffer A H Windle and R H

Friend Adv Mater 1999 11 1281ndash1285

72 E Kymakis and G A J Amaratunga Appl Phys Lett 2002 80

112ndash114

73 Y-C Chen N R Raravikar L S Schadler P M Ajayan Y-P

Zhao T-M Lu G-C Wang and X-C Zhang Appl Phys Lett

2002 81 975ndash977

74 S A OrsquoFlaherty R Murphy S V Hold M Cadek J N

Coleman and W J Blau J Phys Chem B 2003 107 958ndash964

75 A Peigney C Laurent F Dobigeon and A Rousset J Mater

Res 1997 12 613ndash615

76 E Flahaut A Peigney C Laurent C Marliere F Chastel and

A Rousset Acta Mater 2000 48 3803ndash3812

77 A Peigney E Flahaut C Laurent F Chastel and A Rousset

Chem Phys Lett 2002 352 20ndash25

78 J W An and D S Lim J Ceram Process Res 2002 3 201ndash204

79 G-D Zhan J D Kuntz J L Wan and A K Mukherjee Nature

Mater 2003 2 38ndash42

80 G-D Zhan J D Kuntz J E Garay and A K Mukherjee Appl

Phys Lett 2003 83 1228ndash1230

81 T Seeger T Kohler T Frauenheim N Grobert M Ruhle

M Terrones and G Seifert Chem Commun 2002 34ndash35

82 T Seeger G De La Fuente W K Maser A M Benito M A

Callejas and M T Martınez Nanotechnology 2003 14 184ndash187

83 H Han S Vijayalakshmi A Lan Z Iqbal H Grebel E Lalanne

and A M Johnson Appl Phys Lett 2003 82 1458ndash1460

Harris Carbon nanotube composites

42 International Materials Reviews 2004 VOL 49 NO 1

84 P Vincent A Brioude C Journet S Rabaste S T Purcell J Le

Brusq and J C Plenet J Non-Cryst Sol 2002 311 130ndash137

85 J S Sakamoto and B Dunn J Electrochem Soc 2002 149 A26ndashA30

86 R Z Ma J Wu B Q Wei J Liang and D H Wu J Mater Sci

1998 33 5243ndash5246

87 R Andrews D Jacques A M Rao T Rantell F Derbyshire

Y Chen J Chen and R C Haddon Appl Phys Lett 1999 75

1329ndash1331

88 E T Thostenson W Z Li D Z Wang Z F Ren and T W

Chou J Appl Phys 2002 91 6034ndash6037

89 D D L Chung lsquoCarbon fiber compositesrsquo chap 7 1994 Boston

Butterworth-Heinemann

90 S Rawal JOM 2001 53 14ndash17

91 T Kuzumaki K Miyazawa H Ichinose and K Ito J Mater

Res 1998 13 2445ndash2449

92 T Kuzumaki O Ujiie H Ichinose and K Ito Adv Eng Mater

2000 2 416ndash418

93 W X Chen J P Tu L Y Wang H Y Gan Z D Xu and X B

Zhang Carbon 2003 41 215ndash222

94 K Liao and S Li Appl Phys Lett 2001 79 4225ndash4227

95 A Star J F Stoddart D Steuerman M Diehl A Boukai E W

Wong X Yang S W Chung H Choi and J R Heath

Angewandte Chemie-Int Ed 2001 40 1721ndash1725

96 A Star Y Liu K Grant L Ridvan J F Stoddart D W

Steuerman M R Diehl A Boukai and J R Heath

Macromolecules 2003 36 553ndash560

97 P J F Harris S C Tsang J B Claridge and M L H Green

J Chem Soc Faraday Trans 1994 90 2799ndash2802

98 A A Setlur S P Doherty J Y Dai and R P H Chang Appl

Phys Lett 2000 76 3008ndash3010

99 S P Doherty D B Buchholz B J Li and R P H Chang

J Mater Res 2003 18 941ndash949

100 D B Buchholz S P Doherty and R P H Chang Carbon 2003

41 1625ndash1634

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 43

Page 11: Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P. J. F. Harris * Carbon nanotubes are molecular-scale tubes of graphitic carbon with

metal composites may have potential in tribologicalapplications

DiscussionThe subject of carbon nanotube composites has grownrapidly over the past 5 years or so into a vibrant andexciting area of research It is important to recognisehowever that this is still a new subject and the workcarried out so far may have hardly scratched the surfaceof what is possible For example the great majority ofthe studies that have been published to date haveinvolved the use of polymer matrices there is great scopefor more work on incorporating nanotubes into cera-mics metals and other forms of carbon Also most ofthe work up to now has been aimed at exploiting themechanical properties of carbon nanotubes rather thantheir electronic or optical properties although interest inthose is growing

When considering the mechanical performance ofcarbon nanotube composites it is clear that these dependstrongly on whether arc grown tubes or catalyticallyproduced tubes have been used In considering thecharacteristics of the two types of tube it is instructive tocompare the work of Qian et al34 with that of Watts andHsu57 Both incorporated the nanotubes into polymersand then used TEM to characterise the composites andstudy their fracture behaviour Qian et al used cataly-tically produced tubes and these were found to bewell dispersed throughout the polymer matrix with noclustering together of tubes However during compositefracture the tubes were often observed to break sug-gesting that they have relatively poor strength Wattsand Hsu on the other hand employed much moreperfect nanotubes produced by arc evaporation andTEM studies revealed superior mechanical propertieswith no observed breaking However the nanotubestended to cluster together so that during crack growththe tubes could quite easily slide past one another Thuswhile it is clear that the best mechanical performancewill be achieved with arc grown nanotubes it is impor-tant that the tubes be well dispersed throughout thematrix in order fully to exploit their properties

Equally important is achieving good bonding betweennanotubes and matrix Many studies have shown that thisinterfacial bonding can be enhanced by functionalising thenanotubes before embedding them in the matrix Theeffect of this can be seen in Fig 10 from the work ofGojny et al59 where bonding between MWNTs and anepoxy matrix is so strong that the outer shells of the tuberemain embedded in the matrix following pull-out Apossible disadvantage of functionalisation is that it couldweaken the nanotubes by disrupting the perfect cylindricalstructure This might be expected to be a particularproblem with single-walled tubes However functionali-sation of the nanotubes may not be necessary to producebonding with a polymer matrix Liao and Li havemodelled the interfacial characteristics of a nanotubepolystyrene composite assuming no covalent bondingbetween the tubes and the matrix94 It was found thatelectrostatic and van der Waals interactions contributedsignificantly to the interfacial stress transfer A nanotubepull-out simulation suggested that the interfacial shearstress of the nanotubepolystyrene system is about160 MPa significantly higher than for most carbon fibrereinforced polymer composite systems It is also worth

mentioning recent work by Star and co-workers9596

in which it was demonstrated that poly(5-alkoxy-m-phenylenevinylene)-co-[(25-dioctyloxy-p-phenylene)vinylene] (PAmPV) derivatives could be induced towrap around SWNTs forming strong non-covalentbonds

The quality of the nanotubes and the nature of theinterfacial bonding also has a strong influence on theelectrical properties of nanotubepolymer compositesArc grown tubes can have resistivities lower than that ofsingle crystal graphite but as already discussed theirelectronic properties depend on their precise structureAt present we do not have a reliable method of separat-ing metallic and semiconducting tubes so it is notpossible to exploit their full potential Catalytically pro-duced tubes are likely to be poorer conductors thanmetallic arc-grown tubes owing to their relatively dis-ordered structure Nevertheless catalytically grown tubesare sufficiently conducting to be useful in some com-mercial applications such as electrically conductive paintsThe influence of interfacial bonding on electronic pro-perties was demonstrated in the work of Cochet et al44

These workers prepared a MWNTpolyaniline com-posite by in situ polymerisation and found it to haveexcellent electrical transport properties This was attrib-uted to strong interaction between tubes and polymer

As has been shown in this review the possible appli-cations of carbon nanotube composites range widelyfrom electrically conductive paints to optical devicesOnly a relatively small number of these new materialshave been used commercially however and there arestill many problems to overcome before the full potentialof nanotube containing composites can be realised Notleast of these problems is the high cost of good qualitycarbon nanotubes Currently the commercial price ofunpurified arc-grownMWNTs is approximately $10000gwhile purified MWNTs cost around $30000g Cataly-tically grown MWNTs are much cheaper but as empha-sised throughout this review have generally inferiorproperties As far as SWNTs are concerned unpurifiedsamples cost around $5000g while purified SWNTs are$50000g As a comparison the price of gold is cur-rently about $1200g The development of new techni-ques for the low cost synthesis of high quality nanotubesis clearly essential This might be achieved throughimproved catalytic processes and there is currently alarge amount of work being carried out towards thisend Alternatively it may be possible to develop newnon-catalytic techniques for the synthesis of fullerene-related nanotubes In this connection it is interesting tonote that nanotubes can be produced by the high tem-perature heat treatment of conventional carbons such asmicroporous non-graphitising carbons9798 or carbonblack99100 Further work in this area would be of value

Acknowledgement

I thank Milo Shaffer for helpful comments

References1 P J F Harris lsquoCarbon nanotubes and related structures ndash new

materials for the twenty-first centuryrsquo 1999 Cambridge

Cambridge University Press

2 S Iijima Nature 1991 354 56ndash58

3 W Kratschmer L D Lamb K Fostiropoulos and D R

Huffman Nature 1990 347 354ndash358

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 41

4 S Iijima and T Ichihashi Nature 1993 363 603ndash605

5 D S Bethune C H Kiang M S De Vries G Gorman R Savoy

J Vasquez and R Beyers Nature 1993 363 605ndash607

6 H Dai A G Rinzler P Nikolaev A Thess D T Colbert and

R E Smalley Chem Phys Lett 1996 260 471ndash475

7 M M J Treacy T W Ebbesen and J M Gibson Nature 1996

381 678ndash680

8 E W Wong P E Sheehan and C M Lieber Science 1997 277

1971ndash1975

9 M R Falvo G J Clary R M Taylor V Chi F P Brooks

S Washburn and R Superfine Nature 1997 389 582ndash584

10 K Anazawa K Shimotani C Manabe H Watanabe and M

Shimizu Appl Phys Lett 2002 81 739ndash741

11 P Kim L Shi A Majumdar and P L McEuen Phys Rev Lett

2001 87 paper 215502

12 E T Thostenson Z F Ren and T W Chou Compos Sci

Technol 2001 61 1899ndash1912

13 K T Lau and D Hui Composites Part B Engineering 2002 33

263ndash277

14 Y Huh J Y Lee J Cheon Y K Hong J Y Koo T J Lee and

C J Lee J Mater Chem 2003 13 2297ndash2300

15 M Endo T Koyama and Y Hishiyama Jpn J Appl Phys 1976

15 2073ndash2076

16 S C Lyu B C Liu T J Lee Z Y Liu C W Yang C Y Park

and C J Lee Chem Commun 2003 734ndash735

17 M-F Yu O Lourie M J Dyer K Moloni T F Kelly and R S

Ruoff Science 2000 287 637ndash640

18 P Poncharal Z L Wang D Ugarte and W A De Heer Science

1999 283 1513ndash1516

19 J-P Salvetat A J Kulik J-M Bonard G A D Briggs

T Stockli K Metenier S Bonnamy F Beguin N A Burnham

and L Forro Adv Mater 1999 11 161ndash165

20 M-F Yu B S Files S Arepalli and R S Ruoff Phys Rev Lett

2000 84 5552ndash5555

21 M Terrones Int Mater Rev to be published

22 J W Mintmire B I Dunlap and C T White Phys Rev Lett

1992 68 631ndash634

23 R Saito M Fujita G Dresselhaus and M S Dresselhaus Phys

Rev B 1992 46 1804ndash1811

24 N Hamada S Sawada and A Oshiyama Phys Rev Lett 1992

68 1579ndash1581

25 T W Ebbesen H J Lezec H Hiura J W Bennett H F Ghaemi

and T Thio Nature 1996 382 54ndash56

26 J E Riggs D B Walker D L Carroll and Y-P Sun J Phys

Chem B 2000 104 7071ndash7076

27 S C Tsang Y K Chen P J F Harris and M L H Green

Nature 1994 372 159ndash162

28 R M Lago S C Tsang K L Lu Y K Chen and M L H

Green Chem Commun 1995 1355ndash1356

29 Y-P Sun K F Fu Y Lin and W Huang Acc Chem Res 2002

35 1096ndash1104

30 S B Sinnott J Nanosci Nanotechnol 2002 2113ndash123

31 M S P Shaffer X Fan and A H Windle Carbon 1998 36

1603ndash1612

32 M S P Shaffer and A H Windle Adv Mater 1999 11 937ndash941

33 D E Hill Y Lin A M Rao L F Allard and Y-P Sun

Macromolecules 2002 35 9466ndash9471

34 D Qian E C Dickey R Andrews and T Rantell Appl Phys

Lett 2000 76 2868ndash2870

35 R Andrews D Jacques D Qian and T Rantell Acc Chem Res

2002 35 1008ndash1017

36 R Haggenmueller H H Gommans A G Rinzler J E Fischer

and K I Winey Chem Phys Lett 2000 330 219ndash225

37 P Potschke T D Fornes and D R Paul Polymer 2002 43 3247ndash

3255

38 M Sennett E Welsh J B Wright W Z Li J G Wen and Z F

Ren Appl Phys A 2003 76 111ndash113

39 P M Ajayan O Stephan C Colliex and D Trauth Science 1994

265 1212ndash1214

40 J Sandler M S P Shaffer T Prasse W Bauhofer K Schulte

and H Windle Polymer 1999 40 5967ndash5971

41 A Allaoui S Bai H M Cheng and J B Bai Compos Sci

Technol 2002 62 1993ndash1998

42 X Xu M M Thwe C Shearwood and K Liao Appl Phys Lett

2002 81 2833ndash2835

43 D Puglia L Valentini and J M Kenny J Appl Polymer Sci

2003 88 452ndash458

44 M Cochet W K Maser A M Benito M A Callejas M T

Martınez J-M Benoit J Schreiber and O Chauvet Chem

Commun 2001 1450ndash1451

45 M S P Shaffer and K Koziol Chem Commun 2002 2074ndash

2075

46 C Park Z Ounaies K A Watson R E Crooks J Smith S E

Lowther J W Connell E J Siochi J S Harrison and T L St

Clair Chem Phys Lett 2002 364 303ndash308

47 K Jurewicz S Delpeux V Bertagna F Beguin and E

Frackowiak Chem Phys Lett 2001 347 36ndash40

48 M Hughes G Z Chen M S P Shaffer D J Fray and A H

Windle Chem Mater 2002 14 1610ndash1613

49 A A Mamedov N A Kotov M Prato D M Guldi J P

Wicksted and A Hirsch Nature Mater 2002 1 190ndash194

50 J H Rouse and P T Lillehei Nano Lett 2003 3 59ndash62

51 K Lozano and E V Barrera J Appl Polym Sci 2001 79 125ndash

133

52 L Valentini J Biagiotti J M Kenny and S Santucci Compos

Sci Technol 2003 63 1149ndash1153

53 J Sandler P Werner M S P Shaffer V Demchuk V Altstadt

and A H Windle Composites Part A Appl Sci Manuf 2002 33

1033ndash1039

54 R Czerw Z X Guo P M Ajayan Y P Sun and D L Carroll

Nano Lett 2001 1 423ndash427

55 B P Grady F Pompeo R L Shambaugh and D E Resasco

J Phys Chem B 2002 106 5852ndash5858

56 M Cadek J N Coleman V Barron K Hedicke and W J Blau

Appl Phys Lett 2002 81 5123ndash5125

57 P C P Watts and W K Hsu Nanotechnology 2003 14 L7ndashL10

58 L S Schadler S C Giannaris and P M Ajayan Appl Phys Lett

1998 73 3842ndash3844

59 F H Gojny J Nastalczyk Z Roslaniec and K Schulte Chem

Phys Lett 2003 370 820ndash824

60 P M Ajayan L S Schadler C Giannaris and A Rubio Adv

Mater 2000 12 750ndash753

61 M J Biercuk M C Llaguno M Radosavljevic J K Hyun A T

Johnson and J E Fischer Appl Phys Lett 2002 80 2767ndash2769

62 K Lozano J Bonilla-Rios and E V Barrera J Appl Polym Sci

2001 80 1162ndash1172

63 V Jamieson New Sci 15 Mar 2003 30

64 H Zengin W S Zhou J Y Jin W R Czerw D W Smith

L Echegoyen D L Carroll S H Foulger and J Ballato Adv

Mater 2002 14 1480ndash1483

65 J Deng X Ding W Zhang Y Peng J Wang X Long P Li and

A S C Chan Eur Polym J 2002 38 2497ndash2501

66 G B Blanchet C R Fincher and F Gao Appl Phys Lett 2003

82 1290ndash1292

67 S A Curran P M Ajayan W J Blau D L Carroll J N

Coleman A B Dalton A P Davey A Drury B McCarthy

S Maier and A A Strevens Adv Mater 1998 10 1091ndash1093

68 N S Sariciftci L Smilowitz A J Heeger and F Wudl Science

1992 258 1474ndash1476

69 N Camaioni L Garlaschelli A Geri M Maggini G Possamai

and G Ridolfi J Mater Chem 2002 12 2065ndash2070

70 D B Romero M Carrard W De Heer and L Zuppiroli

Adv Mater 1996 8 899ndash902

71 H Ago K Petritsch M S P Shaffer A H Windle and R H

Friend Adv Mater 1999 11 1281ndash1285

72 E Kymakis and G A J Amaratunga Appl Phys Lett 2002 80

112ndash114

73 Y-C Chen N R Raravikar L S Schadler P M Ajayan Y-P

Zhao T-M Lu G-C Wang and X-C Zhang Appl Phys Lett

2002 81 975ndash977

74 S A OrsquoFlaherty R Murphy S V Hold M Cadek J N

Coleman and W J Blau J Phys Chem B 2003 107 958ndash964

75 A Peigney C Laurent F Dobigeon and A Rousset J Mater

Res 1997 12 613ndash615

76 E Flahaut A Peigney C Laurent C Marliere F Chastel and

A Rousset Acta Mater 2000 48 3803ndash3812

77 A Peigney E Flahaut C Laurent F Chastel and A Rousset

Chem Phys Lett 2002 352 20ndash25

78 J W An and D S Lim J Ceram Process Res 2002 3 201ndash204

79 G-D Zhan J D Kuntz J L Wan and A K Mukherjee Nature

Mater 2003 2 38ndash42

80 G-D Zhan J D Kuntz J E Garay and A K Mukherjee Appl

Phys Lett 2003 83 1228ndash1230

81 T Seeger T Kohler T Frauenheim N Grobert M Ruhle

M Terrones and G Seifert Chem Commun 2002 34ndash35

82 T Seeger G De La Fuente W K Maser A M Benito M A

Callejas and M T Martınez Nanotechnology 2003 14 184ndash187

83 H Han S Vijayalakshmi A Lan Z Iqbal H Grebel E Lalanne

and A M Johnson Appl Phys Lett 2003 82 1458ndash1460

Harris Carbon nanotube composites

42 International Materials Reviews 2004 VOL 49 NO 1

84 P Vincent A Brioude C Journet S Rabaste S T Purcell J Le

Brusq and J C Plenet J Non-Cryst Sol 2002 311 130ndash137

85 J S Sakamoto and B Dunn J Electrochem Soc 2002 149 A26ndashA30

86 R Z Ma J Wu B Q Wei J Liang and D H Wu J Mater Sci

1998 33 5243ndash5246

87 R Andrews D Jacques A M Rao T Rantell F Derbyshire

Y Chen J Chen and R C Haddon Appl Phys Lett 1999 75

1329ndash1331

88 E T Thostenson W Z Li D Z Wang Z F Ren and T W

Chou J Appl Phys 2002 91 6034ndash6037

89 D D L Chung lsquoCarbon fiber compositesrsquo chap 7 1994 Boston

Butterworth-Heinemann

90 S Rawal JOM 2001 53 14ndash17

91 T Kuzumaki K Miyazawa H Ichinose and K Ito J Mater

Res 1998 13 2445ndash2449

92 T Kuzumaki O Ujiie H Ichinose and K Ito Adv Eng Mater

2000 2 416ndash418

93 W X Chen J P Tu L Y Wang H Y Gan Z D Xu and X B

Zhang Carbon 2003 41 215ndash222

94 K Liao and S Li Appl Phys Lett 2001 79 4225ndash4227

95 A Star J F Stoddart D Steuerman M Diehl A Boukai E W

Wong X Yang S W Chung H Choi and J R Heath

Angewandte Chemie-Int Ed 2001 40 1721ndash1725

96 A Star Y Liu K Grant L Ridvan J F Stoddart D W

Steuerman M R Diehl A Boukai and J R Heath

Macromolecules 2003 36 553ndash560

97 P J F Harris S C Tsang J B Claridge and M L H Green

J Chem Soc Faraday Trans 1994 90 2799ndash2802

98 A A Setlur S P Doherty J Y Dai and R P H Chang Appl

Phys Lett 2000 76 3008ndash3010

99 S P Doherty D B Buchholz B J Li and R P H Chang

J Mater Res 2003 18 941ndash949

100 D B Buchholz S P Doherty and R P H Chang Carbon 2003

41 1625ndash1634

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 43

Page 12: Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P. J. F. Harris * Carbon nanotubes are molecular-scale tubes of graphitic carbon with

4 S Iijima and T Ichihashi Nature 1993 363 603ndash605

5 D S Bethune C H Kiang M S De Vries G Gorman R Savoy

J Vasquez and R Beyers Nature 1993 363 605ndash607

6 H Dai A G Rinzler P Nikolaev A Thess D T Colbert and

R E Smalley Chem Phys Lett 1996 260 471ndash475

7 M M J Treacy T W Ebbesen and J M Gibson Nature 1996

381 678ndash680

8 E W Wong P E Sheehan and C M Lieber Science 1997 277

1971ndash1975

9 M R Falvo G J Clary R M Taylor V Chi F P Brooks

S Washburn and R Superfine Nature 1997 389 582ndash584

10 K Anazawa K Shimotani C Manabe H Watanabe and M

Shimizu Appl Phys Lett 2002 81 739ndash741

11 P Kim L Shi A Majumdar and P L McEuen Phys Rev Lett

2001 87 paper 215502

12 E T Thostenson Z F Ren and T W Chou Compos Sci

Technol 2001 61 1899ndash1912

13 K T Lau and D Hui Composites Part B Engineering 2002 33

263ndash277

14 Y Huh J Y Lee J Cheon Y K Hong J Y Koo T J Lee and

C J Lee J Mater Chem 2003 13 2297ndash2300

15 M Endo T Koyama and Y Hishiyama Jpn J Appl Phys 1976

15 2073ndash2076

16 S C Lyu B C Liu T J Lee Z Y Liu C W Yang C Y Park

and C J Lee Chem Commun 2003 734ndash735

17 M-F Yu O Lourie M J Dyer K Moloni T F Kelly and R S

Ruoff Science 2000 287 637ndash640

18 P Poncharal Z L Wang D Ugarte and W A De Heer Science

1999 283 1513ndash1516

19 J-P Salvetat A J Kulik J-M Bonard G A D Briggs

T Stockli K Metenier S Bonnamy F Beguin N A Burnham

and L Forro Adv Mater 1999 11 161ndash165

20 M-F Yu B S Files S Arepalli and R S Ruoff Phys Rev Lett

2000 84 5552ndash5555

21 M Terrones Int Mater Rev to be published

22 J W Mintmire B I Dunlap and C T White Phys Rev Lett

1992 68 631ndash634

23 R Saito M Fujita G Dresselhaus and M S Dresselhaus Phys

Rev B 1992 46 1804ndash1811

24 N Hamada S Sawada and A Oshiyama Phys Rev Lett 1992

68 1579ndash1581

25 T W Ebbesen H J Lezec H Hiura J W Bennett H F Ghaemi

and T Thio Nature 1996 382 54ndash56

26 J E Riggs D B Walker D L Carroll and Y-P Sun J Phys

Chem B 2000 104 7071ndash7076

27 S C Tsang Y K Chen P J F Harris and M L H Green

Nature 1994 372 159ndash162

28 R M Lago S C Tsang K L Lu Y K Chen and M L H

Green Chem Commun 1995 1355ndash1356

29 Y-P Sun K F Fu Y Lin and W Huang Acc Chem Res 2002

35 1096ndash1104

30 S B Sinnott J Nanosci Nanotechnol 2002 2113ndash123

31 M S P Shaffer X Fan and A H Windle Carbon 1998 36

1603ndash1612

32 M S P Shaffer and A H Windle Adv Mater 1999 11 937ndash941

33 D E Hill Y Lin A M Rao L F Allard and Y-P Sun

Macromolecules 2002 35 9466ndash9471

34 D Qian E C Dickey R Andrews and T Rantell Appl Phys

Lett 2000 76 2868ndash2870

35 R Andrews D Jacques D Qian and T Rantell Acc Chem Res

2002 35 1008ndash1017

36 R Haggenmueller H H Gommans A G Rinzler J E Fischer

and K I Winey Chem Phys Lett 2000 330 219ndash225

37 P Potschke T D Fornes and D R Paul Polymer 2002 43 3247ndash

3255

38 M Sennett E Welsh J B Wright W Z Li J G Wen and Z F

Ren Appl Phys A 2003 76 111ndash113

39 P M Ajayan O Stephan C Colliex and D Trauth Science 1994

265 1212ndash1214

40 J Sandler M S P Shaffer T Prasse W Bauhofer K Schulte

and H Windle Polymer 1999 40 5967ndash5971

41 A Allaoui S Bai H M Cheng and J B Bai Compos Sci

Technol 2002 62 1993ndash1998

42 X Xu M M Thwe C Shearwood and K Liao Appl Phys Lett

2002 81 2833ndash2835

43 D Puglia L Valentini and J M Kenny J Appl Polymer Sci

2003 88 452ndash458

44 M Cochet W K Maser A M Benito M A Callejas M T

Martınez J-M Benoit J Schreiber and O Chauvet Chem

Commun 2001 1450ndash1451

45 M S P Shaffer and K Koziol Chem Commun 2002 2074ndash

2075

46 C Park Z Ounaies K A Watson R E Crooks J Smith S E

Lowther J W Connell E J Siochi J S Harrison and T L St

Clair Chem Phys Lett 2002 364 303ndash308

47 K Jurewicz S Delpeux V Bertagna F Beguin and E

Frackowiak Chem Phys Lett 2001 347 36ndash40

48 M Hughes G Z Chen M S P Shaffer D J Fray and A H

Windle Chem Mater 2002 14 1610ndash1613

49 A A Mamedov N A Kotov M Prato D M Guldi J P

Wicksted and A Hirsch Nature Mater 2002 1 190ndash194

50 J H Rouse and P T Lillehei Nano Lett 2003 3 59ndash62

51 K Lozano and E V Barrera J Appl Polym Sci 2001 79 125ndash

133

52 L Valentini J Biagiotti J M Kenny and S Santucci Compos

Sci Technol 2003 63 1149ndash1153

53 J Sandler P Werner M S P Shaffer V Demchuk V Altstadt

and A H Windle Composites Part A Appl Sci Manuf 2002 33

1033ndash1039

54 R Czerw Z X Guo P M Ajayan Y P Sun and D L Carroll

Nano Lett 2001 1 423ndash427

55 B P Grady F Pompeo R L Shambaugh and D E Resasco

J Phys Chem B 2002 106 5852ndash5858

56 M Cadek J N Coleman V Barron K Hedicke and W J Blau

Appl Phys Lett 2002 81 5123ndash5125

57 P C P Watts and W K Hsu Nanotechnology 2003 14 L7ndashL10

58 L S Schadler S C Giannaris and P M Ajayan Appl Phys Lett

1998 73 3842ndash3844

59 F H Gojny J Nastalczyk Z Roslaniec and K Schulte Chem

Phys Lett 2003 370 820ndash824

60 P M Ajayan L S Schadler C Giannaris and A Rubio Adv

Mater 2000 12 750ndash753

61 M J Biercuk M C Llaguno M Radosavljevic J K Hyun A T

Johnson and J E Fischer Appl Phys Lett 2002 80 2767ndash2769

62 K Lozano J Bonilla-Rios and E V Barrera J Appl Polym Sci

2001 80 1162ndash1172

63 V Jamieson New Sci 15 Mar 2003 30

64 H Zengin W S Zhou J Y Jin W R Czerw D W Smith

L Echegoyen D L Carroll S H Foulger and J Ballato Adv

Mater 2002 14 1480ndash1483

65 J Deng X Ding W Zhang Y Peng J Wang X Long P Li and

A S C Chan Eur Polym J 2002 38 2497ndash2501

66 G B Blanchet C R Fincher and F Gao Appl Phys Lett 2003

82 1290ndash1292

67 S A Curran P M Ajayan W J Blau D L Carroll J N

Coleman A B Dalton A P Davey A Drury B McCarthy

S Maier and A A Strevens Adv Mater 1998 10 1091ndash1093

68 N S Sariciftci L Smilowitz A J Heeger and F Wudl Science

1992 258 1474ndash1476

69 N Camaioni L Garlaschelli A Geri M Maggini G Possamai

and G Ridolfi J Mater Chem 2002 12 2065ndash2070

70 D B Romero M Carrard W De Heer and L Zuppiroli

Adv Mater 1996 8 899ndash902

71 H Ago K Petritsch M S P Shaffer A H Windle and R H

Friend Adv Mater 1999 11 1281ndash1285

72 E Kymakis and G A J Amaratunga Appl Phys Lett 2002 80

112ndash114

73 Y-C Chen N R Raravikar L S Schadler P M Ajayan Y-P

Zhao T-M Lu G-C Wang and X-C Zhang Appl Phys Lett

2002 81 975ndash977

74 S A OrsquoFlaherty R Murphy S V Hold M Cadek J N

Coleman and W J Blau J Phys Chem B 2003 107 958ndash964

75 A Peigney C Laurent F Dobigeon and A Rousset J Mater

Res 1997 12 613ndash615

76 E Flahaut A Peigney C Laurent C Marliere F Chastel and

A Rousset Acta Mater 2000 48 3803ndash3812

77 A Peigney E Flahaut C Laurent F Chastel and A Rousset

Chem Phys Lett 2002 352 20ndash25

78 J W An and D S Lim J Ceram Process Res 2002 3 201ndash204

79 G-D Zhan J D Kuntz J L Wan and A K Mukherjee Nature

Mater 2003 2 38ndash42

80 G-D Zhan J D Kuntz J E Garay and A K Mukherjee Appl

Phys Lett 2003 83 1228ndash1230

81 T Seeger T Kohler T Frauenheim N Grobert M Ruhle

M Terrones and G Seifert Chem Commun 2002 34ndash35

82 T Seeger G De La Fuente W K Maser A M Benito M A

Callejas and M T Martınez Nanotechnology 2003 14 184ndash187

83 H Han S Vijayalakshmi A Lan Z Iqbal H Grebel E Lalanne

and A M Johnson Appl Phys Lett 2003 82 1458ndash1460

Harris Carbon nanotube composites

42 International Materials Reviews 2004 VOL 49 NO 1

84 P Vincent A Brioude C Journet S Rabaste S T Purcell J Le

Brusq and J C Plenet J Non-Cryst Sol 2002 311 130ndash137

85 J S Sakamoto and B Dunn J Electrochem Soc 2002 149 A26ndashA30

86 R Z Ma J Wu B Q Wei J Liang and D H Wu J Mater Sci

1998 33 5243ndash5246

87 R Andrews D Jacques A M Rao T Rantell F Derbyshire

Y Chen J Chen and R C Haddon Appl Phys Lett 1999 75

1329ndash1331

88 E T Thostenson W Z Li D Z Wang Z F Ren and T W

Chou J Appl Phys 2002 91 6034ndash6037

89 D D L Chung lsquoCarbon fiber compositesrsquo chap 7 1994 Boston

Butterworth-Heinemann

90 S Rawal JOM 2001 53 14ndash17

91 T Kuzumaki K Miyazawa H Ichinose and K Ito J Mater

Res 1998 13 2445ndash2449

92 T Kuzumaki O Ujiie H Ichinose and K Ito Adv Eng Mater

2000 2 416ndash418

93 W X Chen J P Tu L Y Wang H Y Gan Z D Xu and X B

Zhang Carbon 2003 41 215ndash222

94 K Liao and S Li Appl Phys Lett 2001 79 4225ndash4227

95 A Star J F Stoddart D Steuerman M Diehl A Boukai E W

Wong X Yang S W Chung H Choi and J R Heath

Angewandte Chemie-Int Ed 2001 40 1721ndash1725

96 A Star Y Liu K Grant L Ridvan J F Stoddart D W

Steuerman M R Diehl A Boukai and J R Heath

Macromolecules 2003 36 553ndash560

97 P J F Harris S C Tsang J B Claridge and M L H Green

J Chem Soc Faraday Trans 1994 90 2799ndash2802

98 A A Setlur S P Doherty J Y Dai and R P H Chang Appl

Phys Lett 2000 76 3008ndash3010

99 S P Doherty D B Buchholz B J Li and R P H Chang

J Mater Res 2003 18 941ndash949

100 D B Buchholz S P Doherty and R P H Chang Carbon 2003

41 1625ndash1634

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 43

Page 13: Carbon nanotube composites - personal.rdg.ac.ukscsharip/IMR3.pdf · Carbon nanotube composites P. J. F. Harris * Carbon nanotubes are molecular-scale tubes of graphitic carbon with

84 P Vincent A Brioude C Journet S Rabaste S T Purcell J Le

Brusq and J C Plenet J Non-Cryst Sol 2002 311 130ndash137

85 J S Sakamoto and B Dunn J Electrochem Soc 2002 149 A26ndashA30

86 R Z Ma J Wu B Q Wei J Liang and D H Wu J Mater Sci

1998 33 5243ndash5246

87 R Andrews D Jacques A M Rao T Rantell F Derbyshire

Y Chen J Chen and R C Haddon Appl Phys Lett 1999 75

1329ndash1331

88 E T Thostenson W Z Li D Z Wang Z F Ren and T W

Chou J Appl Phys 2002 91 6034ndash6037

89 D D L Chung lsquoCarbon fiber compositesrsquo chap 7 1994 Boston

Butterworth-Heinemann

90 S Rawal JOM 2001 53 14ndash17

91 T Kuzumaki K Miyazawa H Ichinose and K Ito J Mater

Res 1998 13 2445ndash2449

92 T Kuzumaki O Ujiie H Ichinose and K Ito Adv Eng Mater

2000 2 416ndash418

93 W X Chen J P Tu L Y Wang H Y Gan Z D Xu and X B

Zhang Carbon 2003 41 215ndash222

94 K Liao and S Li Appl Phys Lett 2001 79 4225ndash4227

95 A Star J F Stoddart D Steuerman M Diehl A Boukai E W

Wong X Yang S W Chung H Choi and J R Heath

Angewandte Chemie-Int Ed 2001 40 1721ndash1725

96 A Star Y Liu K Grant L Ridvan J F Stoddart D W

Steuerman M R Diehl A Boukai and J R Heath

Macromolecules 2003 36 553ndash560

97 P J F Harris S C Tsang J B Claridge and M L H Green

J Chem Soc Faraday Trans 1994 90 2799ndash2802

98 A A Setlur S P Doherty J Y Dai and R P H Chang Appl

Phys Lett 2000 76 3008ndash3010

99 S P Doherty D B Buchholz B J Li and R P H Chang

J Mater Res 2003 18 941ndash949

100 D B Buchholz S P Doherty and R P H Chang Carbon 2003

41 1625ndash1634

Harris Carbon nanotube composites

International Materials Reviews 2004 VOL 49 NO 1 43