anti-friction, wear-proof and self … · anti-friction, wear-proof and self-lubrication...

75Anti-friction, wear-proof and self-lubrication application of carbon nanotubes

© 2014 Advanced Study Center Co. Ltd.

Rev. Adv. Mater. Sci. 36 (2014) 75-88

Corresponding author: C.W. Wu, e-mail: [email protected]

ANTI-FRICTION, WEAR-PROOF AND SELF-LUBRICATIONAPPLICATION OF CARBON NANOTUBES

W. Zhang, G.J. Ma and C.W. Wu

State Key Laboratory of Structural Analysis for Industrial Equipment,Faculty of Vehicle Engineering and Mechanics, Dalian University of Technology, Dalian 116024, P. R. China

Received: March 09, 2013

Abstract The unique graphite-like layered structure of carbon nanotubes (CNTs) together withtheir high strength, high stiffness, high thermal conductivity and low density make them ideal fora wealth of technological applications. This article outlines the recent advances on the utilizationof CNTs in tribology with the emphases on anti-friction, wear-proof and self-lubrication. To achievesuch functions, CNTs can be either used directly as additives in various liquid lubricant media orembedded as fillers in polymer, metal and ceramic matrices. The critical issues such asprocessing technique, CNT dispersion and interfacial bonding are summarized. The mechanismswith respect to the improvement in tribological performance are highlighted as well. Ourconcentration throughout is on the exploration of the links among composition, structure andtribological property from viewpoint of fundamental scientific research. It is likely that this willdeepen our understanding of the tribological role of CNTs and lead to the proliferation of theirengineering applications.

1. INTRODUCTION

Although it has to be admitted that, under someoccasions, frictional force needs to be increasedfor safety reason, for instance, in automobile brakes,friction clutches, and tires on icy roadways, thereduction of friction and wear, in many cases, isstill the primary objective. Nowadays, as a matterfact, friction has become one of the major reasonsfor failure of vital engineering components andsystems used in aerospace, military, and industrialapplications. It is estimated that the annual cost offriction and wear-related energy and material lossesis over $700 billion, i.e. 5% to 7% of the UnitedStates’ $14 tri]]ion gross nationa] product [1]. Takingthe automobile engine as an example, 5% of thetotal energy generated is lost to frictional resistance.Obviously, it is desperately urgent to deepen ourunderstanding on friction and wear theories at macro/micron levels and develop new materials andmanufacturing techniques to make durable and low

friction surface coatings and materials. This will,undoubtedly, have an immediate impact on betterenergy efficiency, sustainability, and increasedmechanical performance worldwide.

Soon after the discovery of carbon nanotubes(CNTs), on the other hand, it was recognized thatthese seamless tubular materials could be an idealcandidate for a wealth of tribological applications,largely owing to their high strength, high stiffness,high thermal conductivity, high chemical inertnessand unique sp2 bonded structure. For instance, theYoung’s modu]us of a sing]e wa]]ed carbon nanotube(SWCNT) was theoretically predicted to be up to 5TPa [2]. The bending strength of an isolated multi-walled carbon nanotube is as high as 14.2 GPa [3].The tensile strength of CNTs can be 100 fold strongerthan steel, whereas its density is of only one-sixthto one-seventh of steel. Indeed, the last two decadeshave evidenced the proliferation of studies on thetribological applications of CNTs, however, by no

76 W. Zhang, G.J. Ma and C.W. Wu

means, a complete diagram has been depicted andreviews on this topic are actually still scarce. In lightof this, we review the latest documents in the past5 years, namely, published in the year of 2008 andthereafter, on the anti-friction, wear-proof and self-lubrication application of CNTs. The motivation is todepict a picture on the state of the art of this fieldand highlight the immense possibilities of researchand development in this area. It is likely that thiswill contribute to our understandings on thetribological role of CNTs, and may lead to designand manufacture of materials with improvedtribological performance.

This article is structured as the following,illustrated in Fig. 1. First, this article reviews themodeling and simulation, and atomic forcemicroscopy (AFM) experimental work on thetribological properties of individual CNTs or CNTarrays. Following this, the application of CNTs aslubricant additives is discussed. The incorporationof CNTs into polymer matrix to form surface coatingand bulk composite is another concern, in whateverform the research is summarized according toprocessing techniques. Analogously, studies onCNT/metal (ceramic) composite are addressed.Based on the discussion present, finally, this articleoutlines the summary, scope and direction for futurework.

2. Tribological properties of individualCNTs/CNT arrays

The wear resistance of materials is usually improvedby the addition of filler, and CNTs are no exception.

Fig. 1. Structure of this article.

CNTs are composed of multiple cylindrical shellsmade, in concept, by rolling graphene sheets into aseamless cylinder. This easily makes one speculatewhether the lubricating function of CNTs is relatedwith the mutual dislodgement of their cylindricalgraphene layers with one another, like graphite muchor less. This speculation has motivated manyresearch topics. Neglecting the electrostaticColumbic interactions, molecular mechanicssimulation was conducted to examine the pull-outprocess of some outer walls against other innerwalls in MWCNTs and double walled carbonnanotubes (DWCNTs) [4]. The pull out force wasfound to be proportional to the diameter of the criticalwall, i.e. the immediate out wall at the slidingsurface, and independent of nanotube length andchirality. This was explained in the increase ofnumber of carbon atoms in circumferential directionwith increasing diameter; accordingly, stronger vander Waa]’s interactions needed to be overcome forthe possible pull-out. The presence of broken sp2-sp2 carbon bonds may also affect the frictionproperties of CNTs. Based on molecular dynamicssimulation, Li et al. [5] showed DWCNT with 16%inter wall sp3 bonding have friction stresses 3 timeslarger than SWCNT and perfect structured DWCNTin composites with an amorphous carbon matrix,which may be interpreted in the enhancement ofinterlock or radial stiffness of DWCNTs. As for frictioncoefficient, DWCNTs with sp3 bonding are 4 timeslarger than SWCNTs, 0.016 vs. 0.0040, stemmingfrom much rougher surface of the former.

Concurrently, experiments have been designedand carried out to investigate the mechanism and


factors that are associated with the lubricating ef-fect of CNTs, in particular, with the advances of AFMand transmission electron microscopy (TEM)techniques. Zhang et al. proposed, at nanoscale,the friction force between aligned MWCNTs and AFMprobe was governed by the topography-induced forcecaused by the collision of the tip with an asperity ofa positive slope, and adhesion force did not play adominant role [6]. The friction coefficient of CNT filmswith roughness of RMS 2.3 nm and 1.6 nm againstAFM tip were 0.04 and 0.03 respectively. Atmicrometer scale, however, the elastic or plasticdeformation force prevailed and the correspondingfriction coefficients were 0.2 and 0.38, respectively.Sliding friction and adhesion properties of alignedMWCNT array were also quantitatively evaluatedusing AFM by Lu et al.[7]. It was shown friction ismuch higher in dry nitrogen (humidity <5%) than inambient condition (humidity 63%), whereas nosignificant difference in the adhesive forces betweenMWCNT arrays and AFM probe was observed. Againthe lack of correlation between friction and adhesionproperties at nanoscale was observed. But, cautionneeds to be taken in this case. In ambient condition,an active water molecule could be dissociated intoH+ and OH-, and the H+ could be attracted to anactive site on CNT surface (e.g. a carbon danglingbond site) and passivate it. This explained why thefriction was lower. However, as the fraction of activesites that are not passivated is small, the adhesionforce measured by point-to-point force spectroscopymapping actually may not reflect this effect. Usinga TEM equipped with a nanomanipulation system,Suekane et al. successfully measured the staticfriction force between two overlapped CNTs [8]. Theauthors pointed out that the friction force isindependent of overlap length only when the CNTshave a high crystallinity and the friction force ismainly the van der Waals force. While for as grownCNTs usually containing amorphous carbon anddefects, the static friction force increases withincreased overlap length, as was understood to beassociated with surface roughness of CNTs.

2. CNTS AS LUBRICANT ADDITIVE

2.1. Oil based lubricant

The idea of nanoparticle-assisted lubrication hasbeen proposed as a promising concept. Actually,many lubrication oils already contain solid particles.They may be present by accident as in the case ofsoot or debris contaminants or are addeddeliberately. For the latter, metal dichalcogenidesand their derivatives, for example, MoS

2 and WS

2,

are investigated almost exceptionally. From an ap-plication point of view, sometimes, the use of solidlubricants is severely restricted by the operationconditions such as high temperature and highhumidity. The heat arising from the friction processmay activate the oxidation of lubricant, resulting inthe malfunction of lubricant. Zhang et al. [9]electrodeposited MoS

2 on vertically aligned CNTs

to make nanocomposite lubricants. Strikingly, afriction coefficient of as low as 0.04 can be achievedwhen this nanocomposite film was rubbed againstalumina ball at 300 °C on dry sliding. This isinterpreted in terms of the extremely high thermalconductivity of CNTs, which hampers the localaccumulation of friction heat by instantly dissipatingthe generated energy within the matrix and therebydelays the oxidation process. When thesecomposites are involved in commercial machine oil,the friction coefficient can be further reduced byaround 15%, demonstrating the positive tribologicalrole of CNT composites.

It is well accepted that the lubrication effect ofmetal dichalcogenides can be attributed to agradually exfoliation of the external sheets of theparticles during the friction process, which issubsequently transferred onto the reciprocatingsurfaces and forms a tribofilm with a nano-scaledthickness [10].This material transfer mechanismmeans these lubricants have a finite, usually short,lifetime. And also, for environmental protectionreason, it is necessary to reduce or eliminate thepresence of sulphur and phosphorus. As analternative, CNTs have been attempted as novellubricant additives. Owing to high surface energy,CNTs have a great propensity to agglomeratetogether and form bundles/rope. This leads to theirpoor solubility in liquid media. To remove this barrier,1-butyl-3-methylimidazolim hexafluorophosphate[11] and 2-(1-butylimidazolium-3-yl)ethylmethacrylate [12] have been employed tofunctionalize MWCNTs. When added to the baselubricant, both friction coefficient and wear volumeare reduced considerably, see Fig. 2 for an example.During friction, the functionalized MWCNTs fill in themicro-gap of the rubbing surfaces and deposit there,then a self-assembly lubricating thin film forms,providing protection for the specimen surface fromsevere wear. But the addition of excessive MWCNTsdetriments the lubricating effect in that the populationof un-decorated MWCNTs will increase with theincreasing MWCNT concentration, bring about theoccurrence of agglomeration of MWCNTs. Thus themicroscopic rolling effect would become moredifficult to be realized. As reported, the critical CNT


concentration ranged from 0.01 wt. % to 0.06wt. %.

Another concern regarding the use of CNTs assolid additives is associated with the presence ofresidual metal catalyst particles. Whether and howthese particles will affect the tribological behavior ofCNTs is still a question open for arguing. Pottuz etal. reported that the addition of 1 wt.% Ni/Y-SWCNTsto poly-alpha-ole-fin oil can reduce the frictioncoefficient from 0.27 to 0.08, namely, a 70%reduction [13]. In sharp contrast, pure SWCNTs doesnot contribute substantially to the reduction of frictioncoefficient and a value of 0.24 is obtained. X-rayphotoelectron spectroscopy (XPS) analyses of thewear track show the generation of nickel oxides,NiO, and Ni

2O

3, and no carbide signatures are

identifiable; Raman spectra analyses evidence thedisappearance of the characteristic radical breathingmode of SWCNTs. A combination of the above dataenables the authors to hypothesize that, duringfriction process, Ni particles initially encapsulatedinside SWCNTs are exposed to environment in thecontact area and become oxidized, while theSWCNTs are flattened and move like a tank belt.The overall results reveal that residual metal catalyst

particles may contribute a lot to the formation oftribofilm, however, the exact mechanism is still un-certain and more intensive work is required.

2.2. Water based lubricant

Water-based lubricants are gaining popularity innumerous areas such as hydraulic fluid, cutting fluid,metal-foaming operation, oil extraction industry,miniaturized motion devices (e.g. MEMS andNEMS) and biological environments [14]. Comparedwith conventional oil and grease lubricants, waterbased lubricants offer the merit of being eco-friendly.In addition, the high heat capacity of water makes itideal for absorption and transfer of heat. Toincorporate CNTs into aqueous media, CNTs needto be surface modified first, as it is well establishedthat CNTs have very poor solubility. Pei et al.covalently grafted polyacrylamide onto MWCNTs byredox polymerization of acrylamide with cericammonium nitrate being the initiator [15]. The graftedMWCNTs could be dispersed in water and form ahomogeneous and stable dispersion. Whenemployed as solid additives in 2 wt.%triethanolamine and 1 wt.% S-(carboxypropyl)-N-

Fig. 2. 3D images of the worn surfaces lubricated: (a) base lubricant (1-methyl-3-butylimidaaoliumhexafluorophosphate); (b) 0.01 wt.% pristine MWCNTs/lubricant mixtures; and (c) 0.01 wt.% MWCNTs-g-poly 2-(1-butylimidazolium-3-yl)ethyl methacrylate/lubricant mixtures at 200 N, reprinted with permissionfrom [12], (c) 2008 Wiley Periodicals, Inc..


diethyl dithiocarbamic acid, the friction coefficientdrops precipitously and reaches a minimum at aconcentration of 2 wt.%. The authors suggested thatthe MWCNTs served as spacers, preventing roughcontact between sliding counterparts. Whereas theincrease of friction coefficient over 2 wt.% CNTs wasrelated with the formation of CNT lumps in the frictioninterface, decreasing the lubrication efficacy. Thiscovalent functionalization process is effective inimproving the dispersibility of CNTs, but it iscomplicated and the sidewall/end of CNTs needs tobe partly broken for the implantation of functionalmoieties.

As an alternative, non-covalent functionalizationmethod has been adopted to improve thedispersibility of CNTs with the prominent advantageof simplicity. Kristiansen et al. [16] used humic acidto promote the dispersibility of CNTs in water andstudied the tribological behavior of the resultantmixture between two mica surfaces. Fringes of equalchromatic order images display the build-up ofMWCNT layers between mica surfaces even at highloads (pressure about 10 MPa); this accumulationreduces the adhesion between the surfaces andthereby diminishes wear. Interestingly, the frictioncoefficient with either SWCNT or MWCNT dispersionfalls in the range of 0.30-0.55 and is independent ofthe load and sliding velocity.

Irrespective of covalent or non-covalentapproaches, the ideal functionalization is one: (i)after functionalization, the agglomeration of CNTscan be alleviated efficiently; (ii) the modified CNTsshould possess intermediate substrate-surfaceinteractions. The interaction needs to be strongenough to hold the CNTs between the substrates,rather than being squeezed out, during sliding, butat the same time is not too strong that it will makeCNTs adhere to the substrate.

3. CNT/POLYMER COMPOSITE

3.1. Surface coating

Noting that CNTs decorated with functional groupsthat are structurally similar with the polymer matrixshould have good compatibility, many methods havebeen proposed to enhance the interactions betweenCNTs and polymer matrix. Men et al. [17] usedfurfuryl alcohol (FA) to functionalize MWCNTs andfabricated MWCNT-FA/poly(furfuryl alcohol) (PFA)composite coatings on steel by spraying. Thesubstitution of pristine-MWCNT with FA-MWCNTsdecreased the friction coefficient and prolonged wear

life pronouncedly. They ascribed this to the non-homogeneous dispersion of pristine MWCNTs in thePFA composites, giving rise to the production ofcracks between MWCNTs and PFA matrix. Similarly,Zhang et al. [18] grafted maleic anhydride ontoMWCNTs’ surface to enhance the compatibi]ity ofMWCNTs with poly(tetrafluoroethylene) and cotton.

Analogously, Sinha et al. [19] used air plasmato treat SWCNTs with an attempt to improve theirinteractions with ultra-high molecular weightpolyethylene (UHMWPE). SWCNTs/UHMWPE wasthen dip-coated on steel. They examined the effectof counter face materials, namely Si

3N

4, steel and

brass, on the tribological performance. For brassball, the composite coating displayed a lower valueof steady state friction coefficient, and theyattributed this to the more hydrophobic nature ofthe brass ball, which provided less adhesive forceinteractions between the materials. The low wearrate of composite coatings were correlated with thepresence of SWCNTs, which have strong bondingwith polymer matrix and facilitate the improvementin loading bearing capacity and shear strength.Moreover, the authors observed the addition ofSWCNTs could effectively hinder the increment ofthe surface roughness of the wear track caused bythe plastic deformation of polymer matrix, inparticular at high temperature (up to 120 °C), andthey attributed this to the enhanced mechanical andthermal properties intrigued by SWCNTs [20]. Menet al. [21] postulated that the incorporation ofsulfonic acid groups onto MWCNTs would help PFAto coalesce onto the MWCNTs through electrostaticinteractions between the hydroxyls of PFA and thesulfonic groups, and the interactions with PFA woulddepend on the amount of sulfonic groups onMWCNTs. Indeed, the poly(m-aminobenzenesulfonic acid) functionalized short MWCNTs, 0.5-2m in length, illustrated lower fiction coefficient and

longer wear life in comparison with the pristineMWCNTs and long MWCNTs (50 m). They believedthis is a result of the adequate dispersion ofMWCNTs and their strong interfacial bonding withPFA matrix.

Although CNT based surface coatings havedemonstrated the potential to improve the anti-friction and wear-proof properties of certainsubstrates, one fundamental issue of this techniqueis yet not fully addressed. That is how firmly thecoating can be attached to the substrate. In a sense,this will determine whether this technique can gainany engineering application.


3.2. Bulk composite

3.2.1. In situ polymerization

In situ polymerization is one of the most popularmethods to incorporate CNTs into polymer matrix.By this method, poly (methyl methacrylate)/ sty-rene/MWCNTs co-polymer composites [22], poly-styrene/acrylonitrile/MWCNTs composites [23], andpolyacrylonitrile/ methylmethacrylate/MWCNT co-polymer composites [24] etc. have beensuccessfully fabricated. What is common in thesestudies is there exists a critical CNT concentrationwith respect to the tribological performance ofcomposites. The friction coefficient and wear ratedecrease substantially when CNT concentration islower than the critical concentration. When the CNTconcentration exceeds this value, both frictioncoefficient and wear rate increase. In general, thiscritical concentration lies in 1 wt.% to 1.5 wt.% ofpolymer matrix. The incorporation of CNTscontributes to the restraint of the scuffing andadhesion of polymer matrix in sliding against thecounter face, leading to the improvement oftribological performance. Whereas the detrimentaleffects of excessive CNTs can be accounted for bythe agglomeration of CNTs.

Huang et al. [25] functionalized MWCNTs withmaleic anhydride by Friedel-Crafts acrylation, andthe resulting MWCNTs were incorporated into poly(methyl methacrylate) by in situ solutionpolymerization. What is different here is the weightloss of the obtained composite increased withincreasing MWCNT load up to 1 wt.%. A furtherincrease in CNT concentration to 5 wt.% only resultsin a slight decrease in weight loss. The authorsdeduced, at this concentration, the nanoscale ofMWCNTs could not sustain the microscale damagecaused by the grinding wheel and the compositescould be scratched easily by the grinding wheel.

3.2.2. Compression molding

Another popular method to impart CNTs into matrixmaterials is compression molding [26]. Forinstance, SWCNTs were functionalized with 1-octyl,3-methylimidazolium tetrafluoroborate and blendedwith polystyrene (PS), polymethylmethacrylate(PMMA), and polycarbonate (PC) with a content of1 wt.% by compression molding [27]. The wear rateof PS, PMMA and PC reduces by 74%, 63%, and14%, respectively. This was ascribed to the welldispersion of SWCNTs in polymer matrix, whichincreases their resistance to crack propagation and

to the surface modification of SWCNTs, which im-proves the lubricating ability of the additive.Critical CNT concentrationAs aforementioned, the tribological performance ofCNT filled composite is highly dependent on CNTconcentration. For CNT filled carbon fabriccomposites, the critical MWCNT concentration wasreported to be 6 wt.% [28]. In the case of MWCNT-bismaleimide composites, it is about 2.5 wt.% [29].

But there are some controversy reports as well.The wear rate of CNTs/high density polyethylenecomposites decreases monotonically withincreasing CNT concentration up to 5 wt.%. Unlikewear rate, however, there was not a clear, monotonictrend of decreasing friction coefficient with theincreased CNT concentration [30]. As the MWCNT-reinforced polyphenylene sulfide composites areconcerned [31], the change in friction coefficientagainst MWCNT concentration was minute. Thespecific wear rate decreased with the addition of0.2 vol.% MWCNTs, but the wear rate became largerwith further increase in MWCNT concentration.Based on these data, the author even commentedthat, unlike graphite and MoS

2, CNTs may lack the

lubricating capability. Another possible explanationraised by the authors is the by-products containedin pristine MWCNTs may offset the desirable effectof CNTs.

As can be seen above, the critical CNTconcentration may vary from system to system, andsometimes controversy results were presented. Asa matter of fact, the true magnitude can only befound by trial and error. In this aspect, thedevelopment of theoretical model that can be usedto predict this value accurately represents anotherchallenge.Effect of operative variables in slidingThe operative variables such as load, speed andlubricating medium in sliding test will impose impacton the tribological properties of materials. When theload was increased from 50 N to 250 N, the frictioncoefficient of MWCNT/carbon composites exhibiteda descending trend and the wear volume showedan ascending trend [32]. If the sliding speed wasincreased from 0.42 m/s to 0.84 m/s, the scale offriction coefficient changed from 0.16-0.42 to 0.13-0.32, demonstrating the friction-reduction trend,however, the wear volume showed a reverse trendand increased greatly. The effect of load and speedon the tribological properties of MWCNT-carbonfabric composite were also discussed by Zhang etal [27]. When load was increased from 200 N to500 N, both friction coefficient and wear rate tendedto decrease. When the sliding speed was increased


from 0.431 m/s to 0.862 m/s, the authors observeda higher friction coefficient and a lower wear rate.

It is really not too much meaningful to talk aboutthe effect of load and speed without considering thespecific case, as several competitive effects willoccur simultaneous during sliding. The eventualoutcome is the balanced result and will be governedby the prevailing one. With the increase of load orspeed, big particle-shaped or flaky debris in the wearsurface would be crushed or sheared into smallparticles or thin flakes. The newly formed debriswould produce a more integrated but thin film onthe worn surface, which decreases the degree oftwo-body abrasive wear. Meanwhile, the MWCNTsreleased from the wear surface can largely bear theapplied load and hinder direct contact betweensliding counterparts and serve as a lubricant-like‘micro-ro]]er’, ]owering friction coefficient. Another factalso should be mentioned; that is CNTs caneffectively transfer the friction heat due to theirexcellent thermal conductivity and thus reduce thepossibility of adhesive wear. Concurrently, someadverse influence on tribological properties mayinitiated by high load and speed. For instance, theintegrated lubricating film can be formed easily, butit is also readily brushed away due to the strongertangential force caused by high speed and load. Inaddition, an increase in load and speed may resultin a higher contact temperature and decrease theadhesion between CNTs and polymer matrix. As aconsequence, parts of CNTs may fall out from thematrix and become ruptured, resulting in an increaseof friction coefficient.

There are numerous sliding applications whereliquid media are deliberately introduced as coolantsor are present as a working fluid; these media mayalso affect the tribological properties of materials.In comparison with dry sliding, the introduction ofwater decreases friction coefficient of polyamide 6(PA 6)/MWCNT composite, but increases wear rate[33]. Water was not only a good lubricant for thecomposite to reduce the direct contact zone betweenthe sliding counterparts, but a cooling fluid todissipate the frictional heat. As such, the frictioncoefficient was lower in water than in dry sliding.On the other hand, water can distill into theamorphous region of the composite, resulting in areduction in hardness and strength, which made thePA6 fibrils to be detached easily, causing it lesswear resistant. Additionally, an inhibition of theformation of transfer layers of PA6 on the slidingcounterpart may lead to higher wear rate as well.Similar trend was observed for MWCNTs/polytetrafluoroethylene/ polyimide composite [34].

What is more interesting is alkali-lubrication wasobserved to give the lowest wear rate, a nearly 60% reduction with respect to that under dry sliding,even much better than oil lubrication, let alone water.The differences in viscosity may take the mainresponsibility for this phenomenon: the alkali solutionhas a much higher viscosity (0.00115 Pas) thanwater (0.00065 Pas); this indicates the cooling andboundary lubricating effect of alkali solution issuperior to water [27,34]. The substitution of oil withwater generally shows better friction and wearproperties. This could be ascribed to the decreaseof shear strength of the composite and theimmaturely developed transfer film in waterlubrication.

In some occasions, materials need to work onharsh conditions. For example, automotive brakefriction materials must withstand frequent friction athigh temperature. Hwang et al. [35] produced brakefriction materials using CNTs (1.7-8.5 wt.%), phenolicresin, aramid pulp, graphite and zicon. The frictionstability as referenced to the variation of frictioncoefficient with time at 300-°C was improved, whichwas attributed to the reinforcement of CNTs tophenolic resin, reducing the smearing anddelaminating of the friction film due to the weak resinat elevated temperature. Since the torque amplitudeproduced during brake application is directly relatedto the brake induced vibration, the amplitude of thefriction force oscillation generated during the stoptests was compared. The torque variation in the lastrevolution of the disk during stop could be reducedfrom 4.6 N m to 2.5 N m with the introduction ofCNTs; this is because the huge surface area of CNTsincreased the level of interfacial friction between theCNTs and polymer, improving the dissipation anddamping capacity of the composites on the judderpropensity.

4. CNT/METAL (CERAMIC)COMPOSITE

4.1. Surface coating

Dip coatingSol-gel silica reinforced with 0.1 wt.% MWCNTs wasfabricated by mechanical mixing (MM) and ultrasonicprobe mixing (UM), and then dip-coated onmagnesium alloy [36]. The friction coefficient of allthe coatings was in the range of 0.4-0.46 anddifferences among coatings were marginal.Compared with corresponding silica coatings, areduction of 59.4% and 14.5% in wear rate wereobtained for MM and UM coatings, respectively.Scanning electron microscopy (SEM) observation


of worn track showed that many MWCNTs wereconnecting two parts of the worn coating in which acrack formed, implying MWCNTs were locking thecrack propagation by bridging mechanism andalleviated subsequent spalling of the coating bymechanical fatigue processes. UM coating is inferiorto MM one, and this was ascribed to lower innerdensification of UM coating, presumably owing tothe presence of acoustic cavitation bubbles formedby the ultrasonic treatment that speeded uphydrolysis and condensation reactions during thesol formation.Spray coatingA multiple length scale wear study on plasma-sprayed Al

2O

3-CNT composite coating was

conducted by Balani et al. [37]. The introduction ofCNTs can enhance the macro-wear resistance by49 times and nano-wear resistance of up to 18times. This reduction in wear volume loss was

Fig. 3 Wear surface of Al2O

3-4 wt. % CNT coating

showing (a) bimodal wear debris and micro crackarrest; and (b) sworded CNT bridges to renderenhanced wear resistance, reprinted with permis-sion from [37], (c) 2008 Elsevier.

attributed to the interactions of CNTs with the abrad-ing surface: (i) CNTs act as glue to restrain the Al

2O

3

particles, densifying the coating; (ii) CNTs restrictthe crack propagation by bridging mechanism, indi-cating the role of enhance fracture toughness in re-straining the cracks and reducing wear loss; (iii)CNT swording is observed in coating (see Fig. 3),showing the sliding of CNT layers upon wear. Stain-less steel (SS) /MWCNT composite coating wasprepared on SS by thermal spraying. Friction coef-ficient of the composite coating was 3 times lowerthan SS coating and wear rate was reduced bynearly 2 times. The improved wear resistance ofcomposite coating was attributed to the increase ofcoating hardness; the composite coating is muchharder than the SS coating, i.e. 480 HV

0.3 vs. 303

HV 0.3

[38].In situ growthCobalt colloid nanoparticles were coated on SS, onwhich CNTs were grown by plasma enhancedchemical vapor deposition [39]. The resultant filmhas good tribological properties with frictioncoefficient of approximately 0.1-0.2. Raman spectraof wear track and sliding ball scar indicated the G-band of CNTs (1585 cm-1) shifted to a higherfrequency (1605 cm -1) and simultaneously adownshift of D-band from 1350 cm-1 to 1340-1355cm-1. This variation corresponds to the progressivereduction in size of ordered graphite layers in thewalls of CNTs, whist retaining the aromatic rings.These soft materials adhered to mating surfaces,diminishing direct contact between asperities andplough. By studying the tribological properties ofaligned CNTs/alumina composite coating onaluminum substrate, Xia et al. reported that thick-walled CNTs (12.3 nm) demonstrated lower frictioncoefficient than thin-walled ones (4.5 nm) on bothmacro- and nano-scale friction tests [40]. This wasbecause thin-walled CNTs are relatively liable tobuckle upon loading, increasing the possibility ofdirect contact between the sliding tip (ball) andmatrix. Direct growth of CNTs from Ni-based coatingson copper substrate was accomplished through thehydrothermal approach [41]. After annealing at 500°C, the hardness of the coating could reach 656HV

0.1, much higher than that of copper being 105

HV0.1

, and wear loss and friction coefficient couldbe reduced by up to 3.5 times and 3.3 times (1.18vs. 0.35), respectively. SEM observations revealedthe grown CNTs firmly bridges and wraps Ni-basednanoparticles. Consequently, a good load transferringability is achieved during wear test, reducing plasticdeformation and adhesive wear.


Fig. 4. (a) Pure Ni and (c) Ni-CNT composite blended images of Raman spectroscopy color map andoptical micrograph taken inside the wear tracks from friction tests in Fig. 4b. Also shown are corresponding(b) and (d) Raman spectra taken inside (red circles and spectra) and outside (blue circles and spectra) thewear track. The blue circle in (c) shows the location of an unworn CNT. A reference Raman spectrum of CNTpowder in (d) is shown for comparison, reprinted with permission from [42], (c) 2009 American Institute ofPhysics.

Laser -engineering net shape processingScarf et al. used the laser-engineering net shapeprocessing to prepare MWCNT-Ni composite coat-ings on SS [42]. Raman spectroscopy mapping wasperformed inside the wear tracks to elucidate themechanism responsible for the solid lubrication pro-cess. When sliding against Si

3N

4, NiO peak at 553

cm-1, but no carbon peaks, are present for pure Nicoating, see Fig. 4.

Conversely, carbon D (1347 cm-1) and G (1604cm-1) peaks are clearly distinguishable for MWCNT-Ni coating. Compared with G peak at 1582 cm-1 forthe unworn MWCNT-Ni coating, this G peak shifted22 cm-1 toward high wave number. Coupled with theincrease in ratio of intensities I

D/I

G and the full width

at half maximum, the authors inferred that the slidingprocess increases graphitization and the structurallymodified layer has a more disordered structure. Inthe case of sliding against SS, similarly, thepresence and upshift of carbon G peak were alsoevident. By contrast, Fe

3O

4 peaks at 538 and 676

cm-1 become dominant, which is likely due to the

adhesive wear of SS counterface. Based on thesedata, the authors concluded the wear induced, highdisordered and low interfacial shear strength graphiticfilm accounts for the diminution in friction coefficient.Electrochemical methodMany researchers have co-deposited CNTs withmetal particles to modify the tribological performanceof substrate. The concentration of CNTs in depositswill definitely have a role to play, but as this topichas been addressed previously, this section mainlyfocuses on the underlying mechanism.

Ni-P-CNT coating was fabricated by both electro-and electroless deposition. The incorporation of CNTscan efficiently reduce friction coefficient and wearloss [43,44]. Several rationales were put forward forthe observed improvement: (i) CNTs function asintense obstacles to the movement of dislocationsthrough the Ni-P matrix, hindering the occurrenceof plastic deformation; (ii) CNTs serve as separators,preventing the close contact between sliding pairs;(iii) during sliding, cylindrical CNTs are separatedfrom the coating and roll easily between the surfaces,


as partially supported by the observation of CNTs inwear track.

Another striking effect of the incorporation ofCNTs is on the grain size and hardness. Praveen etal. found the introduction of MWCNTs could reducethe grain size of Zn-Ni in electrodeposits from 29nm to 18 nm, as evidenced by SEM images, seeFig. 5 [45]. In a similar study, MWCNTs were co-electrodeposited in nano Ni matrix with a mean grainsize of 28 nm on copper substrate [46]. This fine-grained structure of the composite, together withthe dispersing hardening effect of MWCNTs, wasinvoked to account for the less wear loss. Morerecently, Ni-MWCNTs were co-deposited on Ti-6Al-4V substrate [47]. The incorporation of MWCNTscould increase the coating hardness by 98.5% withregard to pure Ni coating, which was also explainedby the drop in the Ni crystallite size incurred byMWCNTs. The introduction of CNTs can reduce grainsize in that CNTs can provide more nucleation sitesfor crystal growth [48].

Still another explanation for the improvedtribological properties is the rearrangement of CNTsduring sliding. Arai et al.[49] observed that the frictioncoefficient of the Ni-0.5 mass% CNTs compositefilm increased rapidly at the very beginning of sliding,and then decreased gradually and reached a steadyvalue of 0.13, much lower than that of pure Ni being0.33. They related this with the sliding-inducedrearrangement of MWCNTs. Some MWCNTs mayroot in the film with one end protruding from the filmvertically, leading to a high friction coefficient on thecommencement of sliding. Once the sliding ballscratches the film, plastic deformation of the filmwill occur gradually and some MWCNTs aregradually rearranged transversely, resulting in a lowfriction coefficient.

As discussed, the explanations are fairly diverseand may contain contradictions in some cases.Carpenter et al. [50] analyzed the wear track of the

Fig. 5. SEM images of MWCNTs-Zn-Ni coated sample (A) and Zn-Ni coated samples (B), reprinted withpermission from [45], (c) 2009 Elsevier.

Ni-MWCNTs coating with SEM coupled with energydispersive X-ray spectrometry (EDX) and found athick oxide transfer layer covered the majority ofthe wear scar with the metallic coating exposed onlyat a few localized regions and no MWCNTs werevisible on the surface of this oxide transfer layer. Incontrast, the surface of the wear scar in Ni coatingwas largely bare metal. Therefore, the authors arguedthat MWCNTs become submerged within the oxidetransfer layer and can not act as a spacer betweenthe wearing surfaces, and they further deduced theMWCNTs encourage the formation of this transferlayer by bridging the interface between the metaland the transfer layer and adhering loose materialsto the underlying matrix, thereby preventing thetransfer of worn materials from the scar. This loose,oxidized material builds up and forms a protectivetransfer layer, reducing further wear.

4.2. Bulk material

Spark plasma sintering, microwave sintering andhot pressed sintering have been attempted to makeCNT-metal composites with improved tribologicalbehavior with regard to pure metal. Again, the criticalCNT concentration phenomenon was well observed[51], and will not be paid particular attentionrepeatedly. But several additional points are worthmentioning here. To enhance the wettability and thebonding strength between CNTs and metal matrix,CNTs were proposed to be coated with metal (e.g.Cu) by either electroless deposition or hydrationprinciple before sintering [52-54]. Another issue isabout the formation of metal carbide during sintering.XPS analyses of the worn surface of CNT-Alcomposite revealed the emergence of aluminumcarbide (Al

4C

3); the formation of Al

4C

3 may cause a

decrease in hardness and an increase in wear loss[55]. Lastly, to alleviate the damage to CNTs bychemical modification or mechanical blending and


obstruct the pull-out of CNTs from the matrix in slid-ing, porous silica was used templates to hold CNTsby either absorbing CNTs on its surface or embed-ding CNTs in its pores [56].

The improvement in tribological performanceoffered by CNTs could be attributed to thestrengthening of mechanical properties and theformation of graphene layer at contact surfaces.Actually in many cases both mechanisms maywork, although it is not easy to ascertain which oneis the prevai]ing contribution. Evan’s formu]a(Formula 1) was suggested by Ahmad et al. tocharacterize wear volume (V) of Al

2O

3-MWCNTs

[57], which relates the wear volume with the hardness(H), Young’s modu]us E) and fracture toughness(K

IC).

IC

F EV a L

K H H

4 / 59 / 8

1/ 2 5 / 8, (1)

wherein F is the applied load, a is the constantindependent of materials type, and L is the slidingdistance. Feng et al. [58] used the ratio of E over H(E/H) as an indicator to evaluate wear rate for CNTreinforced Ti-Ni composite, the lower E/H value is,the better wear resistance is. Both these calculationsare in consistence with the experimental results.This is not surprising. A material with high hardnessis more difficult to deform under high loads, while]ow Young’s modu]us cou]d ]ower the contactpressure, leading to a delay in plastic deformation[59].

The formation of graphene layer is largely affectedby the sliding operatives. For CNT reinforcedhydroxyapatite [60], at macro-scale, wear resistanceincreased by 66%, whereas friction coefficientdecreased by 60%. In case of nano-scale wear, a45% increase in wear resistance was observed,however, friction coefficient increased by 14%. Thedifferent effect of CNT on friction coefficient wasassociated with the formation of graphene layer. Theeffective shear stress during macro-wear wascalculated to be 22 GPa, while it was only about300 MPa in micro-wear. It was estimated that theremoval of a single graphene layer from MWCNTsrequires a tensile force in the range of 11-63 GPa[61]. Lateral force during sliding causes tensilestress on the newly exposed surface after massremoval. As the force applied in nano-wear is muchsmaller than the minimum stress required forgraphene layer peeling from CNT (11 GPa), nolubrication from graphene layer peeling is available.On the contrary, the available lateral stress in macro-wear (21 GPa) is sufficient to remove graphene layer,

causing a decrease in friction coefficient. ForDWCNT-Cu composites, Guiderdoni et al. [62] usedthe maximum Hertzian contact pressures (Formu-lae 2-5) to examine if the formation of a graphitizedlubricating tribofilm in contact could take place duringsliding.

Max

FP

a2

3,

2 (2)

where a is the contact radius and F is the appliedload.

FRa

E

1/ 33

.2

(3)

The equivalent contact radius R* and the equivalentYoung modulus E* are defined as

RR * (ball- plane contact),

2 (4)

ball plane

ball planeE E E

2 2

*

1 11, (5)

R is the ball radius, is Poisson coefficient, E iselastic modulus.

The calculated maximum Hertzian contactpressures ranged from 495 MPa to 950 MPa,depending on the type of sliding ball and appliedload, which are sufficiently high to shear the wallsof individual DWCNTs, since it was documented that215 MPa is a pressure high enough to incur theshearing [63]. This generated tribofilm during slidingmay account for the 4 order of decrease in frictioncoefficient.

5. SUMMARY AND CHALLENGESFOR FUTURE WORK

As stated above, CNTs have demonstrated greatpotential in reducing friction and improving wearresistance. However, to unambigiously understandthe tribological role of CNTs is truly a challengingtask, since friction between contact surfaces is sucha complex issue, which is characterized by theinterplay of energy, stress and chemistry at manylength scales. So many factors will affect thetribological performance, for instance, thecomposition and property of sliding pairs (such assurface roughness, hardness, elastic modulus,fracture toughness etc.) and the sliding operativeparameters (such as sliding load, speed,temperature and lubrication state etc.). As such,


diverse mechanisms have been proposed based ondiscrete cases. (i) CNTs can bridge crack and lockthe propagation of cracks; (ii) CNTs can restrainmetal/ceramic particles and densify materials; (iii)the dislodgement of individual graphene layers ofCNTs can provide lubrication effect; (iv) CNTs canstrengthen the mechanical properties of matricesthat they are embedded in; (v) CNTs can staybetween mating surfaces and diminish the directcontact between asperities and plough; (vi) CNTscan effectively dissipate the friction heat and reducetemperature induced wear. It is really hard to ascribethe positive tribological role of CNTs to any soleclause, as for most cases, if not all, more than onemechanism may work simultaneously.

CNTs continue to hold promise for applicationsin tribology but with significant challenges for realsuccess. (i) To achieve the uniform properties ofcomposites, CNTs need to be distributedhomogenously in matrices and consequently propermanufacturing method is required. Althoughnumerous approaches have been proposed, it is stillnot quite clear to what extent the perfect structuresof CNT have been preserved after experiencing theprocess, quite often associated with harshconditions such as high temperature and highshearing stress. So far, only a few methods toquantify the level or quality of dispersion of CNTs inthe microstructure have been documented. Hence,there is also a need to develop a universalquantification scheme for the quality of CNTdistribution in order to compare various processesfor their ability to disperse the CNTs in themicrostructure. (ii) CNT composites exhibit goodtribological performance at optimum concentration,which is widely ascribed to the agglomeration ofCNTs at high concentration. This concentrationvalue varies from one system to another dependingon many factors such as the aspect ratio of CNTs,their dispersion and orientation state in matrices,their interactions with the matrix at the interfaces,and the phase separation effect of a third phase, ifapplied. At current situations, this is usually doneby trial and error. Novel model needs to be formulatedand more simulation work should be accompanied.(iii) A strong interfacial connection between thematrix and CNTs is always essential. As a weakinterface connection with matrix would end up withCNTs being dragged out of the worn surface duringwear, losing their lubrication features and becomingdebris. CNTs have been grafted with monomer orcoated with metal to enhance their compatibility withmatrices. The detailed mechanism is not yet as-certained and more techniques are welcome. (iv)

The toxicity of CNTs is still a topic under debate. Ifthe unique tribological potential of CNTs is to beexploited on industry scale, toxicological studiesmust be conducted in parallel, before eventuallyconverging to provide a clear framework acceptableto regulatory authorities and the public.

In brief, the extremely excellent mechanicalproperties of CNTs coupled with their graphite-likestructure have stimulated tremendous studies ontheir tribological role and this filed is rapidly evolving.The up-to-date status of work has witnessed thepositive effect of CNTs solidly, and it is conceivablethat this effect and its engineering application willbe strengthened even more with further reduction ofprices and availability of better quality of CNTs.

ACKNOWLEDGEMENT

The National Natural Science Foundation of China(51105051, 11272080) and the FundamentalResearch Funds for the Central Universities of China(DUT11RC(3)79, DUT12ZD101) were acknowledgedfor the financial support of this work.

REFERENCES

[1] C. McKnight, NanoMech wins highlycompetitive department of energy grant, http://www.businesswire.com/news/home/20121120005239/en/NanoMech-Wins-Highly-Competitive-Department-Energy-Grant.Accessed on Jan. 25, 2013.

[2] W.E. Wong, P.E. Sheehan and C.M. Lieber// Science 277 (1997) 1971.

[3] M.M.J. Treacy, T.W. Ebbesen and J.M. Gibson// Nature 381 (1996) 678.

[4] Y. Li, N. Hu, G. Yamamoto, Z. Wang,T. Hashida, H. Asanuma, C. Dong, T. Okabe,M. Arai and H. Fukunaga // Carbon 48 (2010)2943.

[5] L. Li, J.B. Niu, Z.H. Xia, Y.Q. Yang and J.Y.Liang // Scripta Mater. 65 (2011) 1014.

[6] W. Zhang, B. Xu, A. Tanaka and Y. Koga //Carbon 47 (2009) 926.

[7] H. Lu, J. Goldman, F. Ding, Y. Sun, M.X.Pulikkathara, V.N. Khabashesku, B.I.Yakobson and J. Lou // Carbon 46 (2008)1294.

[8] O. Suekane, A. Nagataki, H. Mori andY. Nakayama // Appl. Phys. Express 1 (2008)064001.

[9] X. Zhang, B. Luster, A. Church, C. Muratore,A.A. Voevodin, P. Kohli, S. Aouadi andS. Talapatra // Appl. Mater. Inter. 1 (2009) 735.


[10] R. Rosentsveig, A. Gorodnev, N. Feuerstein,H. Friedman, A. Zak, N. Fleischer,J. Tannous, F. Dassenoy and R.Tenne //Tribol. Lett. 36 (2009) 175.

[11] B. Wang, X. Wang, W. Lou and J. Hao //J. Phys. Chem. C. 114 (2010) 8749.

[12] X. Pei, Y. Xia, W. Liu, B. Yu and J. Hao //J. Appl. Polym. Sci. A 46 (2008) 7225.

[13] L. Joly-Pottuz, B. Vacher, T.L. Mogne, J.M.Martin, T. Mieno, C.N. He and N.O. Zhao //Tribol. Lett. 29 (2008) 213.

[14] B. Bhushan, Nanotribology &Nanomechanics- An Introduction (Springer,Heidelberg, 2008).

[15] X. Pei, L. Hu, W. Liu and J. Hao // Eur. Poly.J. 44 (2008) 2458.

[16] K. Kritiansen, H. Zeng, P. Wang and J.N.Israelachvili // Adv. Funct. Mater. 21 (2011)4555.

[17] X.H. Men, Z.Z. Zhang, H.J. Song, K. Wangand W. Jiang // Compos. Sci. Technol. 68(2008) 1042.

[18] H.J. Zhang, Z.Z. Zhang, F. Guo, K. Wang andW. Jiang // Mater. Chem. Phys. 116 (2009)183.

[19] M. Abdul Samad and S.K. Sinha // Wear 271(2011) 2759.

[20] M. Abdul Samad and S.K. Sinha // Wear 270(2001) 395.

[21] X.H. Men, Z.Z. Zhang, J. Yang, K. Wang andW. Jiang // J. Appl. Polym. Sci. 118 (2010)2881.

[22] B. Dong, C. Wang, B.L. He and H.L. Li //J. Appl. Polym. Sci. 108 (2008)1675.

[23] C. Wang, T. Xue, B. Dong, Z. Wang and H.L.Li // Wear 265 (2008) 1923.

[24] C. Wang, B. Dong, G.Y. Gao, M.W. Xu andH.L. Li // Mater. Sci. Eng. A 478 (2008) 314.

[25] Y.L. Huang, S.M. Yuen, C.C.M. Ma, C.Y.Chuang, K.C. Yu, C.C. Teng, H.W. Tien, Y.C.Chiu, S.Y. Wu, S.H. Liao and F.B. Weng //Compos. Sci. Technol. 69 (2009) 1991.

[26] B. Likozar and Z. Major // Appl. Surf. Sci.257 (2010) 565.

[27] F.J. Carrion, C. Espejo, J. Sanes and M.D.Bermudez // Compos. Sci. Technol. 70(2010) 2160.

[28] X. Zhang, X. Pei and Q. Wang // Polym. Adv.Technol. 22 (2011) 2157.

[29] L. Liu, Z.P. Fang, A.J. Gu and Z.H. Guo //Compos. Part B-Eng. 42 (2011) 2117.

[30] B.B. Johnson, M.H. Santare, J.E. Novotnyand S.G. Advani // Mech. Mater. 41 (2009)1108.

[31] M. Cho // Mater. T. 49 (2008) 2801.[32] X.H. Men, Z.Z. Zhang, J. Yang, X.T. Zhu,

K. Wang and W. Jiang // Tribol. T. 54 (2011)265.

[33] M. Meng, G.X. Sui, G.Y. Xie and R. Yang //Compos. Sci. Technol. 69 (2009) 606.

[34] Y.J. Shi, L.W. Mu, X. Feng and X.L. Lu //J. Appl. Polym. Sci. 121 (2011)1574.

[35] H.J. Hwang, S.L. Jung, K.H. Cho, Y.J. Kimand H. Jang // Wear 268 (2010) 519.

[36] A.J. Lopez, A. Urena and J. Rams // ThinSolid Films 519 (2011) 7904.

[37] K. Balani, S.P. Harimkar, A. Keshri, Y. Chen,N.B. Dahotre and A. Agarwal // Acta Mater.56 (2008) 5984.

[38] D. Kaewsai, A. Watcharapasorn, P. Singjai,S. Wirojanupatump, P. Niranatlumpong andS. Jiansirisomboon // Surf. Coat. Tech. 205(2010) 2104.

[39] M.D. Abad, J.C. Sachez-Lopez,A. Berenguer-Murcia, V.B. Golovko,M. Cantoro, A.E.H. Wheatley, A. Fernandz,B.F.G. Johnson and J. Robertson // Diam.Relat. Mater. 17 (2008) 1853.

[40] Z.H. Xia, J. Lou and W.A. Curtin // ScriptaMater. 58 (2008) 223.

[41] F. Liu and Y. Zhang // Diam. Relat. Mater. 25(2012) 144.

[42] T.W. Scharf, A. Neira, J.Y. Hwang, J. Tileyand R. Banerjee // J. Appl. Phys. 106 (2009)013508.

[43] Y. Suzuki, S. Arai and M. Endo //J. Electrochem. Soc. 157 (2010) D50.

[44] M. Alishahi, S.M. Monirvaghefi, A. Saatchiand S.M. Hosseini // Appl. Surf. Sci. 258(2012) 2439.

[45] B.M. Praveen and T.V. Venkatesha // J. AlloyCompd. 482 (2009) 53.

[46] P.Q. Dai, W.C. Xu and Q.Y. Huang // Mat.Sci. Eng. A-Struct. 483-484 (2008)172.

[47] C.K. Lee // Tribol. Int. 55 (2012) 7.[48] Y.L. Yang, Y.D. Wang, Y. Ren, C.S. He, J.N.

Deng, J. Nan, J.G. Chen and L. Zuo // Mater.Lett. 62 (2008) 47.

[49] S. Arai, A. Fujimori, M. Murai and M. Endo //Mater. Lett. 62 (2008)3545.

[50] C.R. Carpenter, P.H. Shipway and Y. Zhu //Wear 271 (2011) 2100.

[51] P.V. Trinh, T.B. Trung, N.B. Thang, B.H.Thang, T.X. Tinh, L.D. Quang, D.D. Phuongand P.N. Minh // Comp. Mater. Sci. 49 (2010)S239.

[52] C.B. Lin, Z. Chang, Y.H. Tung and Y. Ko //Wear 270 (2011) 382.


[53] K. Rajkumar and S. Aravindan // Wear 270(2011) 613.

[54] H.J. Choi, S.M. Lee and D.H. Bae // Wear270 (2010) 12.

[55] I. Kim, J. Lee, G. Lee, S. Baik, Y. Kim andY. Lee // Wear 267 (2009) 593.

[56] J. Umeda, K. Kondoh and H. Imai // Mat. Sci.Eng. A Struct. 504 (2009) 157.

[57] I. Ahmad, A. Kennedy and Y.Q. Zhu // Wear269 (2010)71.

[58] X. Feng, J.H. Sui, W. Cai and A.L. Liu //Scripta Mater. 64 (2011) 824.

[59] W.Y. Yan // Mater. Sci. Eng. A- Struct. 427(2006) 348.

[60] D. Lahiri, V. Singh, A.K. Keshri, S. Seal andA. Agarwal // Carbon 48 (2010) 3103.

[61] M.F. Yu, O. Lourie, M.J. Dyer, K. Moloni, T.F.Kelly and R.S. Ruoff // Science 287 (2000)637.

[62] C. Guiderdoni, C. Estournes, A. Peigney,A. Weibel, V. Turq and C. Laurent // Carbon49 (2011) 4535.

[63] E. Bichoutskaia, O.V. Ershova, Y.E. Lozovikand A.M. Popov // Tech. Phys. Lett. 35(2009) 666.

anti-friction, wear-proof and self … · anti-friction, wear-proof and self-lubrication...

Documents