Enhanced wear resistance of high-density polyethylene compositesreinforced by organosilane-graphitic nanoplatelets

Tian Liu, Bin Li, Brooks Lively, Allen Eyler, Wei-Hong Zhong n

School of Mechanical and Materials Engineering, Washington State University, Pullman, WA 99164, USA

a r t i c l e i n f o

Article history:Received 7 May 2013Received in revised form21 October 2013Accepted 26 October 2013Available online 13 November 2013

Keywords:Sliding wearPolymer–matrix compositeGraphite nanoplateletSolid lubricants

a b s t r a c t

It is known that graphite has excellent lubricant properties due to the 2D graphene layers bonded via van derWaals forces. Thus, graphite nanoplatelets (GNPs) should have high lubricating efficiency during contactfrictional movement of sliding parts. However, GNPs have rarely been used to improve the tribologicalproperties of polymeric materials. In this study, we aimed to improve wear resistance of GNP–reinforced high-density polyethylene (HDPE) composites via the synthesis of organosilane-modified GNPs. Wear resistanceof the HDPE/GNP composites was examined on a pin-on-disc wear testing apparatus under various slidingvelocities. The results revealed that compared to the composites without filler surface modification, significantenhancements in wear resistance under different sliding velocities were realized in silanized-GNP–reinforcedHDPE composites. In particular, 97% wear resistance improvement under 1.3 m s-1 sliding velocity wasobtained. Furthermore, the organosilane-modified GNPs minimized the influence of sliding velocity on wearresistance of the composite and thereby maintained excellent wear resistance in a broad range of slidingvelocities, and even at very high velocity (2.0 m s�1). Both superior wear resistance and good stability of wearperformance at various sliding velocities suggest silanized-GNP/HDPE composites are promising materials withgreat potential for a wide range of tribological applications.

& 2013 Elsevier B.V. All rights reserved.

1. Introduction

Tribological properties of polymeric materials are significant tothe performance and service life in many applications, includingload-bearing total joint replacements, airplane interiors, machin-ery parts of automobiles, medical devices, and other grindingsystems. However, many polymers do not possess satisfactorytribological properties [1,2]. In order to improve tribologicalperformance of polymers including the increase in wear resistanceand the reduction of the friction coefficient, numerous attemptshave been undertaken, typically involving the addition of differentadditives into the base polymers [3,4]. Although the reportedresults showed that better tribological performances could beachieved for composite materials under general wear conditions,the increased elastic modulus and the reduced crack resistance ledto failure of the composites under severe wear [4]. In addition, thethermal effect caused by wear process is another potential threatto the performance of polymeric materials. The increasing tem-perature during wear, especially at high wear sliding velocity, canaffect the structures and properties of polymer matrices, owing tothe nature of their viscoelasticity. The generated heat will speedup the damage of the material. Therefore, to meet the demand forgood wear performance stability and maintain life expectancy of

friction elements, development of reinforced polymeric compo-sites by adding appropriate reinforcements is necessary.

Graphite nanoplatelets (GNPs) are an ideal reinforcement for poly-meric materials because of their superior physical and mechanicalproperties, multi-functionalities, and the abundance of the necessaryraw materials in nature. The resulting GNP–reinforced polymernanocomposites benefit the creation of next-generation materials forbroad applications, such as automotive, aeronautical and biomedicalmanufacturing [5,6]. In recent years, studies of polymeric nanocom-posites with GNPs have primarily focused on the development ofelectrically and thermally conducting composites, gas storage, sensors,etc., showing extraordinary accomplishments in enhancing electrical,mechanical and thermal properties [7–9]. To date, very little work hasbeen demonstrated on the tribological benefits of GNP–polymernanocomposites.

With unique structure and exceptional properties, GNPs havethe potential to improve wear resistance of polymer compositeswhile minimizing the thermal effect generated by wear. However,very few studies have related to the resultant tribological perfor-mance of GNP–reinforced polymer composites thus far. GNPsconsist of stacked 2D graphene layers bonded via van der Waalsforces. These multilayered 2D structures allow strong interlayersliding motion with GNPs under external force, which providesGNP with an outstanding lubricant property, resulting in lesserwear [10]. Moreover, on each graphene basal plane, there existsplenty of sp2 hybridized carbons which are conducive to thestorage and dissipation of frictional heat. The high thermal

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/wear

Wear

0043-1648/$ - see front matter & 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.wear.2013.10.013

n Corresponding author. Tel.: þ1 509 335 7658; fax: þ1 509 335 4662.E-mail address: katie_zhong@wsu.edu (W.-H. Zhong).

Wear 309 (2014) 43–51

stability and conductivity of GNP–reinforced polymer compositescan be exploited to diminish the negative effect on polymer matrixresulting from heat generated during wear [11]. Furthermore,functional groups on commercial GNP surfaces, which are pro-duced as a result of oxidation processes during manufacturing, canserve as active sites for surface modification to strengthen inter-facial bonding and reinforce the composites. Hence, GNP is anideal nano-reinforcement to improve tribological properties ofpolymers.

To make high performance polymer composites, reinforce-ments that can be homogenously dispersed and strongly bondedto the polymer matrix are the essential prerequisites. Research ontribological properties of polymer composites has demonstratedthat effective filler treatment can contribute to the improved wearresistance of the composites through uniform filler dispersion andenhanced interfacial interaction [12]. For GNP–reinforced polymercomposites, the strengthened interactions between GNP and polarpolymer matrices can be easily achieved through the reactive sitesalong the edges of commercial GNPs, such as hydroxyl, carbonyland epoxide groups [13]. Hence, the fabrication of the compositeswith uniform GNP dispersion and strong GNP–matrix interactionis more facile. For non-polar polymers, the functional groups oncommercial GNP surfaces are incompatible with hydrocarbonpolymer matrices, such as polyethylene (PE), which have highwear resistance. In this case, the surface modification for GNPshas been commonly applied to improve the interfacial bonding inpolymer composites [11,13]. In general, various chemicals can beselected, such as diaminoalkanes, poly(sodium 4-styrene sulfo-nate), aryl diazonium salts, etc. [14–16] However, these agents donot significantly contribute to a homogeneous GNP dispersion. Therestacking of GNP sheets can compromise their effectiveness [8].

In the current work, we studied various GNPs to reinforce high-density polyethylene (HDPE) for tribological applications. As-received,thermally purified and surface-modified GNPs were selected. Theorganosilane coupling agent was used to modify GNPs. As a popularsurface modification approach, a variety of silane coupling agentshave been applied to various fillers [17,18]. To date, the studies ontribological performance of polymer composites with silane-modifiedcarbon-based nanofillers have been mainly focused on carbon nano-fibers (CNF) and carbon nanotubes (CNT) [18,19]. The relevantinvestigations of GNPs have been rarely reported to authors' knowl-edge. In this study, octadecyltrimethoxysilane (ODMS) was selected asa modifier to treat GNPs to achieve the desired uniform fillerdispersion and strong interaction with the HDPE matrix. Qualitativemorphological study and quantitative filler dispersion analysis showedthat the HDPE/GNP composite with homogenerously dispersed nano-fillers and enhanced interfacial interaction was realized by silanemodification of GNPs. The wear testing results revealed that the useof silanized GNP in composites was capable of improving wearresistance dramatically and maintaining the outstanding wear resis-tance at various sliding velocities. A significant enhancement on wearresistance was found in the silanized GNP composites, which was upto 97% at 1.3 m s�1 wear sliding velocity. This work suggests that thelubrication effect of GNPs can be successfully transferred to thenanocomposites after silane surface modification.

2. Experimental procedures

2.1. Materials

High-density polyethylene (HDPE) (HP54-60 Flake) was sup-plied by Bamberger Polymers Inc. with a density of 0.954 g cm�3.The as-received graphitic nanoplatelet (as-GNP) used in this studywas purchased from XG Science Inc., which was pre-exfoliatedwith dimensions approximately 5 μm in diameter and 5 nm

thickness. Octadecyltrimethoxysilane (ODMS) (90% technicalgrade) was used as the organosilane coupling agent and waspurchased from Sigma-Aldrich. Sodium hydroxide (NaOH) (5% w/vaq. Soln.) was supplied by Alfa Aesar Johnson Matthey.

2.2. Method of synthesis for silanized GNPs

2.2.1. High-temperature purification of as-received GNPsThe as-received GNPs (as-GNPs) were pre-exfoliated via an acid

intercalation approach by the supplier. This is a highly oxidativeprocess and can create an abundant amount of chemically activesites on the particle surface, typically hydroxyl and carbonylgroups [13]. After exfoliation, small amounts of acid residuals stillexist in GNP particles. To remove the acid impurities for furthersurface modification and composite fabrication, the as-GNPs werethermally purified in a muffle furnace at 1000 1C for 5 min toobtain high-temperature purified GNPs (ht-GNPs).

2.2.2. Surface modification of thermally purified GNPsOctadecyltrimethoxysilane (ODMS) was used as the coupling

agent to modify the ht-GNPs further, as its end groups with longhydrocarbon chains could lead to good interaction with HDPEmatrix. The ht-GNPs were first bath sonified in 95% ethanol for 1 h.Next, 5�10�5 mol/L NaOH to be used as a catalyst was then addedand mixed with the ht-GNPs/ethanol solution. ODMS was diluted in100% ethanol and added to the ht-GNP suspension slowly in 0.5 mlincrements after the suspension started to reflux. The whole reactionwas under reflux for 5 h. Finally, the silane-modified GNPs werevacuum dried, washed with 100% ethanol to remove the unreactedsilane coupling agent, and then the silanized-GNPs (s-GNPs) weredried in oven at 70 1C for 24 h.

2.3. Manufacturing of HDPE/GNP composites

The 3 wt% as-GNPs, ht-GNP and s-GNPs were melt-blendedwith HDPE by a Haake Torque Rheometer. Mixing was set at 170 1Cwith a rotator speed of 7.3 rad s�1 for 15 min. The resultingnanocomposites are named as Comp-asGNP, Comp-htGNP andComp-sGNP, respectively. A reference sample of neat HDPE wasalso prepared under identical conditions. All samples were subse-quently hot-pressed into 2.5 mm thick panels at 180 1C for 15 minon a hydraulic compression molding machine. The panels wereallowed to cool down to room temperature naturally after turningoff the heat. All manufacturing parameters for graphitic nanofillerreinforced HDPE composites were selected based on our previousstudies [19]. The specimens for pin-on-disk wear testing were cutfrom the compression molded panels. The detailed descriptions ofeach prepared sample are shown in Table 1.

2.4. Characterization

2.4.1. TGA (Thermogravimetric analysis)A Thermogravimetric Analysizer (TA Instruments, SDT Q600,

USA) was used to verify the removal of acid residuals in the as-GNPs after thermal purification, as well as the wt% of organosilanecoating on the surface of s-GNPs. The temperature range for TGA

Table 1Detailed description for each sample.

Sample Reinforcement Filler content

Pure HDPE – –

Comp-asGNP As-received GNPs 3 wt%Comp-htGNP High-temperature purified GNPsComp-sGNP Silanized GNPs

T. Liu et al. / Wear 309 (2014) 43–5144

tests were from ambient temperature to 800 1C, with a heatingrate of 10 1C/min in an N2 environment and a flow rate of around100 mL/min. Approximately 10 mg nanoparticles were used ineach test.

2.4.2. FTIR (Fourier transform infrared spectroscopy)FTIR (Thermo Fisher Scientific, Nicolet Nexus 670 E.S.P., USA)

was applied to both ht-GNPs and s-GNPs for the confirmation ofsuccessful silanization and the characterization of functionalgroups from ODMS modification of the nanofillers. 0.02 g of GNPswere mixed with 1 g of potassium bromide (KBr) by mechanicalgrinding. Then, the mixtures were compressed to disks with athickness of ca. 0.5 mm.

2.4.3. TEM (transmission electron microscope)To determine the existence of organosilane coating on s-GNP

surfaces, TEM (FEI Tecnai, G2 T20, USA) was used to observe themorphologies of ht-GNPs and s-GNPs. For TEM sample prepara-tion, GNP particles were first suspended in ethanol followed by3 h bath sonication. Subsequently, the middle layer of the GNPsuspension was collected and dispersed onto a 300 mesh copperTEM grid for overnight vacuum drying.

2.4.4. FESEM (field emission scanning electron microscope)FESEM (FEI Tecnai, Quanta 200F, USA) was used to investigate

the fillers dispersions, the interaction between filler and polymermatrix, as well as the worn surface of the nanocomposites. Thefractured surfaces of the composites were obtained by fracturingthe samples after immersion in liquid nitrogen for 10 min beforefracture. The surfaces of all samples were sputter coated with goldto improve electrical conductivity.

2.4.5. Light microscopeA light microscope (Olympus, BX51TRF, USA) equipped with a

camera (Olympus, U-CMAD 3, USA) was also applied to analyze thedispersion of different GNP nanoparticles in the polymer matrix.The GNP aggregate size distribution was quantitatively carried outthrough the radius measurement for each GNP aggregate. Anin-house prepared MATLAB program was used to binarize thelight microscope images and measure the visible agglomerates.Agglomerate radii were recorded as half of the largest point-to-point distance on the agglomerate. In order to assure the accuracy

of quantitative studies, fifty light microscope images were ran-domly taken for each type of the composites. Subsequently, the2-D GNP-aggregation size distribution of the composites wasplotted. The dispersion analysis accomplished by both FESEMand light microscopy could provide comprehensive dispersioninformation for GNPs. In particular, the effects of surface modifica-tion of GNPs on dispersion will be understood.

2.4.6. Wear testingThe wear tests were performed in a custom-built rig with a

vertical 1020 carbon steel disk (Hitachi, L200, USA). The effectiveradius of the disk (the radius of wear track) was 70 mm. A normalforce of 36 N was applied to the specimen. The schematic diagramof the components part of the pin-on-disk wear testing apparatuswas shown in Scheme 1. In this study, the wear testing wasconducted under three different sliding velocities: 0.3 m s�1,1.3 m s�1 and 2.0 m s�1. In order to keep the same slidingdistance, the corresponding wear times were 27 h, 6 h and 4 h,respectively. All the tests were performed at room temperature.And five repeated tests were conducted for each type of sample.The weight loss was recorded and the wear coefficient wassubsequently calculated after each testing period. Specific volu-metric wear rate was calculated by following equation:

w¼ ΔmρFd

ð1Þ

where w is specific volumetric wear rate, Δm refers to weight lossduring each wear period, ρ is the density of the sample, F is normalforce and d is linear sliding distance. In order to avoid theinfluence of previous test, the wear disk was cleaned with 100%ethanol and dried. After that, it was polished by fine sand paper ofthe type silicon carbide 220b (220 grit) and then conduct 4 h wearfor pure HDPE. Finally, the disk was cleaned thoroughly with 100%ethanol again and dried prior to further use.

3. Results and discussion

3.1. FTIR

As a surface treatment approach to enhancing filler dispersionand load transfer in polymer composites, silanization has beenused to modify oxidized carbon nanofibers (CNF) and oxidized

Scheme 1. (a) Schematic diagram of pin-on-disk wear testing apparatus; (b) Detailed description of wear disk, sample holder, and sample during wear process.

T. Liu et al. / Wear 309 (2014) 43–51 45

carbon nanotubes (CNT) [18,19]. In the current study, the chemicalreaction between ODMS and ht-GNPs during the silane treatmentcan be described by Scheme 2. Composed of sp2 hybridizedcarbons, ht-GNPs with functional groups can react with hydro-lyzed silane molecules to form strong covalent bonds. First,numerous reactive silanol groups were produced by the hydrolysisof ODMS in a boiling ethanol-water solution. Subsequently, thecondensation reaction of hydrolyzed ODMS with ht-GNPs occurred.The organosilane was finally coated onto GNP surfaces through theformation of strong chemical bonds.

In Fig. 1, the FTIR spectrum shows the constituent comparisonbetween ht-GNP and s-GNP. First, the absorption peaks at around1620 cm�1 and 3410 cm�1 are associated with carbonyl (CQO)and hydroxyl (O–H) bonds, respectively. These groups on the GNPsurface were abundantly produced during the process of pre-exfoliation, and they served as chemically active sites in thesilanization reaction. Second, the distinct difference between ht-GNP and s-GNP is the absorbance in the range of 800 cm�1 and1100 cm�1. Additional peaks in this range appeared in s-GNP.A strong peak at 925 cm�1 is ascribed to the creation of Si–O–GNP,confirming the chemically connected ODMS-GNP. Additionally,there exists an inconspicuous peak at around 830 cm�1, indicatingsome unreacted Si–O–CH3 groups remaining on the GNP surface.The new absorption peak formed around 1010 cm�1 can beassigned to Si–O–Si bonding, representing the establishment ofan ODMS self-crosslinking network structure [19]. Thus, the FTIRresults exhibit compelling evidence for the presence of self-crosslinked silane coating chemically bonded with GNP, suggest-ing a successful silane treatment.

3.2. Thermogravitimetric analysis of GNPs

GNPs with and without purification and surface modificationcould show different thermal stabilities. Thus, the thermal decom-position behaviors of GNPs could be an effective measurement ofsurface modification. To assess the removal of impurities (acidresiduals) in as-GNPs and to quantify the amount of organosilanecoating on s-GNPs, the thermal decomposition behaviors for as-GNP,

ht-GNP, and s-GNP are presented in Fig. 2. For as-GNPs, a 9 wt%weight loss occurred from 200 1C to 600 1C, caused by the removal ofacid residuals on as-GNPs. In comparison, this 9 wt% loss disappearedin ht-GNPs, suggesting the successful removal of impurities via thethermal purification process. The further organosilane modificationon ht-GNPs made a substantial weight loss degraded at 500 1C,which represents the degradation of coated organosilane on the GNPsurface. In addition, the s-GNPs yielded a high weight percentageof silane coating (48 wt%), accordingly, the average silane coatingthickness on s-GNP was calculated by following equation:

ab¼ VGNPaðbþ2τÞ ð2Þ

where a and b are diameter and thickness of GNP, respectively, VGNPis the volume fraction of GNPs, and τ is the average thickness of thesilane coating. Assuming that the densities of GNPs and ODMS are2 g/cm3 and 0.88 g/cm3, respectively, the estimated average silanecoating thickness is shown in the inset table of Fig. 2.

Scheme 2. Chemical reaction mechanism for organosilane coupling agent coated onto thermally purified GNP (ht-GNP). (a) ht-GNP; (b) hydrolyzed ODMS; and (c) silanizedGNP (s-GNP).

1000 1500 2000 2500 3000 3500

Abs

orba

nce

(A.U

.)

Wavenumber(cm-1)

ht - GNPs - GNP

Si-O-GNP

C=O

O-HSi-O-CH3

Si-O-Si

Fig. 1. FTIR spectra for thermally purified GNP (ht-GNP) and silanized GNP (s-GNP).

T. Liu et al. / Wear 309 (2014) 43–5146

3.3. Morphological study

FESEM was applied to qualitatively observe the morphologiesof various GNP composites (Fig. 3). The images with low magni-fication were shown to represent the dispersion and distributionstatus of different GNPs in HDPE (Fig. 3A–C). For Comp-asGNP,there were observable filler aggregations in the polymer matrix(Fig. 3A). For Comp-htGNP, although the GNP dispersion anddistribution were improved, a few small filler aggregations stillexisted (Fig. 3B). Additionally, most of the GNPs in Comp-asGNPand Comp-htGNP were exposed on the HDPE surface. In compar-ison, Comp-sGNP showed uniformly dispersed and well distribu-ted s-GNPs in HDPE matrix (Fig. 3C). In addition, almost all GNPswere able to remain embedded in the polymer matrix. Theinteractions between filler and polymer matrix can be clearly seen

in Fig. 3(a–c). For both Comp-asGNP and Comp-htGNP, visible gapsexist between the two phases, reflecting the unsatisfactory inter-facial interaction. In contrast, no gaps were created at filler-matrixinterface in Comp-sGNP. The morphologies of single ht-GNP(Fig. 3b-inset) and s-GNP (Fig. 3c-inset) were also observed byTEM. The ht-GNP showed a flaky morphology with noticeablecleavage features characteristic of the exfoliated structure. Thes-GNP displayed a smooth plate surface attributed to the coveredsilane coating. Also, most of the portions of s-GNP plate weresurrounded by a visually darker substance, indicating the locationof coated organosilane.

3.4. Dispersion analysis

In order to gain comprehensive understanding of the disper-sion of various GNPs, the quantitative nanofiller dispersion withinthe polymer matrix was characterized through light microscopeimage analysis. Fig. 4 displays representative light microscopeimages of three HDPE/GNP composites. Noticeably, Comp-asGNPand Comp-htGNP had large visible GNP aggregates, especially forthe as-received GNP composites. The quantitative nanofiller dis-persion was described by counting aggregated GNPs in eachradius range.

The statistical analysis on nanofiller dispersion within the GNPcomposites was performed via an in-house prepared MATLABscript through binarization of the acquired light microscopeimages [20]. Fig. 5(a–c) provides the distributions of GNP aggre-gate size in a broad range from several nanometers to 200 μm. Ineach radius range, the proportion of filler aggregation in Comp-asGNP was the highest compared to the other two composites.In the size range below 40 μm, a lower number of agglomerateswas present in the two treated-GNP–reinforced HDPE composites,even for the tiny agglomerations with the radius below 10 μm(Fig. 5a and inset table). This phenomenon was more obviousin the size range of 40–200 μm (Fig. 5b and c). Compared to

0 200 400 600 800

50

60

70

80

90

100

Wei

ght (

%)

Temperature(°C)

as - GNPht - GNPs - GNP

48 wt%

9 wt%

Estimated average coating

thickness of s-GNP5.8 nm

Fig. 2. TGA curves comparing the as-received GNP (as-GNP), thermally purifiedGNP (ht-GNP), and silanized GNP (s-GNP) (inset table: estimated average silane-coating thickness of s-GNP).

Cleavages

Fig. 3. FESEM micrographs (scale bars: 20 μm for 1st row and 1 μm for 2nd row) of the fracture surfaces of HDPE nanocomposites reinforced by as-received GNP (as-GNPs)(A) and (a), thermally purified GNP (ht-GNP) (B) and (b), silanized GNP (s-GNP) (C) and (c). The filer loading of all composites was 3 wt%. TEM images (scale bar: 1 μm) ofht-GNP (b-inset) and s-GNP (c-inset).

T. Liu et al. / Wear 309 (2014) 43–51 47

Comp-asGNP, the filler dispersion of Comp-htGNP was obviouslyimproved. Even so, the large aggregated GNPs existed sparsely,and distributed in the radius range of 120–200 μm (Fig. 5c). Incontrast, Comp-sGNP had the least filler aggregates, and thenumber of aggregates decreased dramatically as the aggregateradius increased. In particular, the amount of large aggregates withsizes above 130 μm was zero. This suggested that the satisfactorilydispersed GNP composite system with reduced number and sizeof aggregations can be realized by the organosilane modification ofnanofillers, which was in agreement with the morphologicalanalysis results in Fig. 3.

3.5. Wear resistance

Wear tests at three sliding velocities were conducted for pureHDPE and its nanocomposites. Fig. 6 shows the comparison oftheir specific volumetric wear rates, which were calculated byEq. (1). For each sliding velocity, Comp-asGNP and Comp-htGNP

200 μm 200 μm 200 μm

Fig. 4. Typical examples of light microscope images for (a) Comp-asGNP; (b) Comp-htGNP; and (c) Comp-sGNP.

40 50 60 70 80 90 100 110 1200

30

60

90

120 Comp- as GNP Comp- ht GNP Comp- s GNP

120 130 140 150 160 170 180 190 2000

3

6

9

12

15 Comp- as GNP Comp- ht GNP Comp- s GNP

0 4 8 12 16 20 24 28 32 36 400.0

2.0x103

4.0x103

6.0x103

8.0x103

1.0x104

1.2x104

No.

of N

anof

iler A

ggre

gatio

ns /m

m2

Agglomerate Radius (μm)

Comp- as GNP Comp- ht GNP Comp- s GNP

No.

of N

anof

iler A

ggre

gatio

ns /m

m2

No.

of N

anof

iler A

ggre

gatio

ns /m

m2

Agglomerate Radius (μm) Agglomerate Radius (μm)

Fig. 5. Aggregation radius distributions of nanofillers for all composites. The size range of radius is (a) below 40 μm, (b) 40–120 μm, and (c) above 120 μm.

0

20

40

60

120

130

140

Spec

ific

Volu

met

ric W

ear R

ate

(10-6

mm

3 /Nm

)

Pure HDPE Comp-as GNP Comp-ht GNP Comp-s GNP

0.3 1.3 2.0

Sliding Velocity (meters per second)

Fig. 6. Specific volumetric wear rates of pure HDPE, Comp-asGNP, Comp-htGNP,and Comp-sGNP under Pin-on-Disk wear tests with three sliding velocities.

T. Liu et al. / Wear 309 (2014) 43–5148

displayed very high wear rates, even higher than pure HDPE. Thedeterioration of wear resistance in nanocomposites can be attrib-uted to two primary factors: the existence of GNP aggregates, andthe residual acid and exposed reactive groups on GNP which resultin poor interfacial interaction with the matrix, leading to loadtransfer failure from polymer matrix to nanofillers. This phenom-enon was more obvious for Comp-asGNP. In our previous studies,the addition of as-received CNF with graphitic structures in HDPEimproved the wear resistance of the composites compared to pureHDPE [21]. Although both CNF and GNP are composed of graphenelayers, their effects on wear resistance of their HDPE compositesare opposite. We may conclude that the structure of nano-reinforcement can make a great impact on the wear resistance ofcomposites. Unlike the HDPE/CNF composite system, as-GNP or ht-GNP with their 2-D plate structures have a smaller opportunity toentangle with the PE chains. Thus, the un-modified GNP cannotcontribute to reinforcing the whole system for the creation of highwear resistance composites. In contrast, Comp-sGNP maintainedthe lowest specific volumetric wear rate for each sliding velocity.This reflects that s-GNP can stabilize the wear resistance ofthe composite, even at very high velocity (2.0 m s�1). In addition,the reactive groups on the GNP surface can be covered bythe organosilane coating. The long silane hydrocarbon tails onthe coating may also provide sufficient entanglement with the PEchains, and then induce a significantly enhanced filler-matrixinterfacial region in the composites [17].

The effect of nanofiller treatments on the wear resistance ofthe resulting GNP composites under various wear sliding velocitiescan be viewed clearly from Fig. 7. With the increase of slidingvelocity, the specific volumetric wear rate of the three compositesincreased. The increased interface temperature between sampleand rotating wear disc caused by higher sliding velocity was themajor influencing factor on the increase in wear rate [22]. Due tothe viscoelastic nature of polymer matrices, the internal structureand overall performance of a polymer is highly susceptible totemperature. The rise in temperature during a high speed wearprocess can lead to breaking and reforming of the secondarybonds in the polymer matrix and decrease wear property of thecomposites [22].

Additionally, the increased wear rates with respect to slidingvelocity for the three nanocomposites were different (Fig. 7). Theslope values of these curves quantitatively show the differentinfluence of sliding velocity on the wear rates of differentcomposites. Among the three nanocomposites, Comp-asGNP pre-sented the highest slope, followed by Comp-htGNP and the lowestvalue was found in Comp-sGNP. For the HDPE/GNP composite

system in this study, the different effects of sliding velocity onwear rate were representative of both nanofiller dispersion andfiller-matrix interfacial interaction in the different compositesamples. Comp-sGNP yielded the lowest slope, reflecting thestability of wear resistance of the composite at various slidingvelocities, due to homogeneous GNP dispersion and improvedGNP-HDPE interfaces. As prerequisites, uniform dispersion andstrong interfacial bonding with the polymer allowed s-GNPs to actas effective heat storage and emission channel which can extractand release friction heat. This can protect the polymer matrix fromoxidation and degradation caused by heat generated during wear[11]. In particular, the capability to maintain a low wear rate athigh velocity demonstrates evidence of the contributions of silanemodification on GNPs for enhanced wear resistance of thecomposites.

To intuitively explore the effect of GNP treatments on wearresistance improvement of the composites, the variation of wearresistance improvement of both Comp-htGNP and Comp-sGNPcompared to Comp-asGNP were calculated by the followingequation:

IWR ¼Was�W

Was� 100% ð3Þ

In Eq. (3), IWR is the improvement in wear resistance, Was refers tothe specific volumetric wear rate of Comp-asGNP, and W repre-sents the specific volumetric wear rate of either Comp-htGNP orComp-sGNP. Table 2 displays the calculation results of IWR for thethermally purified and further organosilane-modified GNP com-posites under three sliding velocities. For both treated GNPcomposites, the highest wear resistance improvements werefound at 1.3 m s�1. However, at the other two sliding velocities(0.3 and 2.0 m s�1), their wear resistance improvements wererelatively low. The reasons for this phenomenon can be under-stood by the following two aspects: (1) at low speed (0.3 m s�1),although less heat is produced during the wear process, posingless of a threat to the thermal stability of the polymer matrix,ample time is allowed for viscoelastic response, making wear-generated stress the major culprit in the velocity range withlimited improvement in wear resistance; (2) at high speed(2.0 m s�1), wear-generated heat was a dominant factor in therestriction of the increase in IWR. Consequently, compared toComp-asGNP, the wear resistances of the two treated GNP compositesat 0.3 m s�1 and 2.0 m s�1 sliding speeds were not improved sig-nificantly. When comparing two treated GNP composites, the wearresistance improvements of Comp-sGNP were much more remarkablethan that of Comp-htGNP, especially at 1.3 m s�1 which were up to97%. This indicates that filler dispersion and filler-matrix interface playcrucial roles in maintaining wear resistance of the composites undervarious sliding velocities.

The worn surface of pure HDPE and its nanocomposites canprovide morphological evidence for the wear resistance of unmo-dified and silane-modified GNP–reinforced HDPE nanocomposites,

1800

20

40

60

80

100

120

140

160 Comp-as GNP Comp-ht GNP Comp-s GNP

Sliding Velocity (meters per second)2.01.30.3

Slope: 0.55 (R=0.90)

Slope: 0.23 (R=0.91)

Slope: 0.07 (R=0.82)

Spec

ific

Volu

met

ric W

ear R

ate

(10-6

mm

3 /Nm

)

Fig. 7. Influence of sliding velocity on specific volumetric wear rate for three GNPcomposites.

Table 2Influence of treatments on wear resistance improvement of GNP composites.

Composite sample Treatmentmethod

SlidingVelocityvelocity (m s�1)

Improvement inwear resistance(compared toComp-asGNP) (%)

Comp-htGNP High temperature 0.3 39.91.3 78.42.0 53.0

Comp-sGNP Silane coupling agent 0.3 73.41.3 96.62.0 85.1

T. Liu et al. / Wear 309 (2014) 43–51 49

as presented in Fig. 8. For pure HDPE, Comp-asGNP, and Comp-htGNP, obvious defects were observed on their worn surfaces,including PE melting and cracks resulting from the frictional heatgenerated on the sample surface. Nevertheless, a relatively smoothsurface with few asperities for Comp-sGNP demonstrated thecontribution of organosilane coating to GNPs. These SEM imagesdisplay the evidence of reduced damage in Comp-sGNP, resulting

from uniform GNP dispersion and strong GNP-HDPE interfacialinteraction.

Combing the results obtained across this study, the wearmechanism of HDPE/GNP composites can be described by Fig. 9.For Comp-asGNP and Comp-htGNP, the addition of GNPs drama-tically decreased wear resistance of the composites. Due to poorfiller dispersion and inferior interfacial interaction between phases

Fig. 8. FESEM micrographs (scale bar: 300 μm; inset: 20 μm) for the typical worn surface of all samples after wear testing at 2.0 m s�1 sliding velocity.

Fig. 9. Schematic illustration of sliding wear mechanism for (a) un-silanized GNP composites and (b) silanized GNP composite.

T. Liu et al. / Wear 309 (2014) 43–5150

in the composites, the creation of cracks and crack propagationcaused by wear was the primary influencing factor. This damageresulted from ineffective load transfer from filler to matrix. Incontrast, an effective GNP network was established in Comp-sGNP,which was attributed to uniform dispersion and good distributionof s-GNPs in HDPE. During the continuous wear process, thegenerated frictional heat can be stored and dissipated throughthe effective s-GNP network. In addition, the strengthened filler-matrix interface can restrain the formation of cracks betweenphases caused by wear produced stress. In this case, GNP inter-layer sliding can occur; accordingly, GNP acting as a solid lubricanthas the ability to significantly improve wear resistance of thecomposite. Therefore, the analysis results strongly suggest that theorganosilane modification on GNPs have a great impact on reinforcingthe whole composite system and thereby can successfully maintainsuperior wear resistance of the composite at various sliding velocitiesranging from low to high.

4. Conclusions

Graphite nanoplatelets (GNPs) with 2D multilayered structureswere silanized and then used to fabricate high-density polyethy-lene (HDPE) nanocomposites for tribological applications. Thelubrication effects of GNPs were successfully achieved after ther-mal purification and organosilane surface modification of GNPs.It was found that both ht-GNPs and as-GNPs led to decreased wearresistance of the nanocomposites due to the poor dispersion andinterfaces in these two nanocomposites. Uniform GNP dispersionand strong physical interaction in filler-matrix interface wereascertained from thermal purification and organosilane surfacemodification for GNPs, correspondingly, enhanced wear resistancewas achieved. In particular, a considerable enhancement up to97% was found in Comp-sGNP under 1.3 m s�1 wear sliding velocity.Furthermore, superior wear resistance of the composite at varioussliding velocities was maintained after silanization of GNPs. Therefore,the silane treatment of GNP particles is favorable route for reinforcingthe HDPE nanocomposites with prominent and stable wear resistancein a broad range of wear sliding velocities.

Acknowledgments

The authors gratefully acknowledge the support from theNational Science Foundation (Civil, Mechanical and ManufacturingInnovation 0856510).

References

[1] M. Goldman, L. Pruitt, Comparison of the effects of gamma radiation and lowtemperature hydrogen peroxide gas plasma sterilization on the molecular

structure, fatigue resistance, and wear behavior of UHMWPE, J. Biomed. Mater.Res. 40 (1998) 378–384.

[2] W.X. Chen, F. Li, G. Han, L.Y. Wang, J.P. Tu, Z.D. Xu, Tribological behavior ofcarbon-nanotube-filled PTFE composites, Tribol. Lett. 15 (2003) 275–278.

[3] L. Fang, Y. Leng, P. Gao, Processing of hydroxyapatite reinforced ultrahighmolecular weight polyethylene for biomedical applications, Biomaterials 26(2005) 3471–3478.

[4] T.M. Wright, T. Fukubayashi, A.H. Burstein, The effect of carbon fiber reinforce-ment on contact area, contact pressure, and time-dependent deformation inpolyethylene tibial components, J. Biomed. Mater. Res. 15 (1981) 719–730.

[5] H. Hu, G. Chen, Electrochemically modified graphite nanosheets and theirnanocomposites films with poly(vinyl alcohol), Polym. Compos. 31 (2010)1770–1775.

[6] L. Wang, J. Hong, G. Chen, Comparison study of graphite nanosheets andcarbon black as fillers for high density polyethylene, Polym. Eng. Sci. 50 (2010)2176–2181.

[7] Y.C. Li, S.C. Tjong, R.K.Y. Li, Electrical conductivity and dielectric response ofpoly(vinylidene fluoride)–graphite nanoplatelet composites, Synth. Met. 160(2010) 1912–1919.

[8] S. Stankovich, D.A. Dikin, G.H.B. Dommett, K.M. Kohlhaas, E.J. Zimney,E.A. Stach, Graphene-based composite materials, Nature 442 (2006) 282–286.

[9] S. Ganguli, A.K. Roy, D.P. Anderson, Improved thermal conductivity forchemically functionalized exfoliated graphite/epoxy composites, Carbon 46(2008) 806–817.

[10] B. Li, W.H. Zhong, Review on polymer/graphite nanoplatelet nanocomposites,J. Mater. Sci. 46 (2011) 5595–5614.

[11] E. Planes, J. Duchet, A. Maazouz, J.F. Gerard, Characterization of new formula-tions for the rotational molding based on ethylene–propylene copolymer/graphite nanocomposites, Polym. Eng. Sci. 48 (2008) 723–731.

[12] Y.J. Shi, X. Feng, H.Y. Wang, X.H. Lu, The effect of surface modification on thefriction and wear behavior of carbon nanofiber-filled PTFE composites, Wear264 (2008) 934–939.

[13] B. Li, E. Olson, A. Perugini, W.H. Zhong, Simultaneous enhancements indamping and static dissipation capability of polyetherimide composites withorganosilane surface modified graphene nanoplatelets, Polymer 52 (2011)5606–5614.

[14] H.A. Margarita, A.A. Ahmed, J.M. Michael, A. Ilhan, K.P. Robert, Intercalationand stitching of graphite oxide with diaminoalkanes, Langmuir 23 (2007)10644–10649.

[15] Y.C. Si, E.T. Samulski, Synthesis of water soluble graphene, Nano Lett. 8 (2008)1679–1682.

[16] J.R. Lomeda, C.D. Doyle, D.V. Kosynkin, W.F. Hwang, J.M. Tour, Diazoniumfunctionalization of surfactant-wrapped chemically converted graphenesheets, J. Am. Chem. Soc. 130 (2008) 16201–16206.

[17] M. Abdelmouleh, S. Boufi, M.N. Belgacem, A. Dufresne, Short natural-fibrereinforced polyethylene and natural rubber composites: effect of silanecoupling agents and fibres loading, Compos. Sci. Technol. 67 (2007)1627–1639.

[18] P.C. Ma, H.K. Kim, B.Z. Tang, Functionalization of carbon nanotubes using asilane coupling agent, Carbon 44 (2006) 3232–3238.

[19] W. Wood, S. Kumar, W.H. Zhong, Synthesis of organosilane-modified carbonnanofibers and influence of silane coating thickness on the performance ofpolyethylene nanocomposites, Macromol. Mater. Eng. 295 (2010) 1125–1135.

[20] B. Lively, W.H. Zhong, An efficient quantified stereological macrodispersionanalysis approach for determining microscale influences on nanocompositematerial properties, Macromol. Mater. Eng. 298 (2013) 221–234.

[21] T. Liu, W. Wood, B. Li, B. Lively, W.H. Zhong, Effect of reinforcement on weardebris of carbon nanofiber/high density polyethylene composites: morpholo-gical study and quantitative analysis, Wear 294–295 (2012) 326–335.

[22] Y.Q. Wang, J. Li, Sliding wear behavior and mechanism of ultra-high molecularweight polyethylene, Mater. Sci. Eng. A 266 (1999) 155–160.

T. Liu et al. / Wear 309 (2014) 43–51 51