1-s2.0-S0261306913005724-main

Materials and Design 53 (2014) 129–136

Contents lists available at SciVerse ScienceDirect

Materials and Design

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

Effect of surface roughness on tribological properties of TiB2/Alcomposites

0261-3069/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.matdes.2013.06.038

⇑ Corresponding author at: School of Materials Science and Engineering, HarbinInstitute of Technology, Harbin 150001, China. Tel./fax: +86 451 86412164.

E-mail address: [email protected] (G.H. Wu).

S.F. Tian a, L.T. Jiang a,b, Q. Guo a, G.H. Wu a,b,⇑a School of Materials Science and Engineering, Harbin Institute of Technology, Harbin 150001, Chinab State Key Laboratory of Advanced Welding Production Technology, Harbin Institute of Technology, Harbin 150001, China

a r t i c l e i n f o a b s t r a c t

Article history:Received 25 November 2012Accepted 25 February 2013Available online 2 July 2013

Keywords:Aluminium matrix compositesTribological propertiesSurface roughness

55 vol.% TiB2/2024Al composites were fabricated by pressure infiltration method. The effect of surfaceroughness of GCr15 steel disc (Ra 0.606, 0.372, 0.023, 0.005 lm) on the tribological properties of compos-ites was investigated. Results showed that with the change of surface roughness, there is an optimal value(Ra 0.023 lm) under which the friction coefficient and wear rate is the lowest. The optimal surface rough-ness is in the same order of mixture of TiO2 and B2O3, observed on the surface of TiB2 particles after pre-heating process. During sliding, the filling of this oxidation layer into the asperity gap of GCr15 andgreatly reduces adhesion between aluminium and GCr15, furthermore, decreases the friction coefficientand wear rate.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction Sirico investigated the influence of surface roughness on wear.

TiB2 particles reinforced aluminium matrix (TiB2/Al) compositeshave been widely investigated due to their excellent properties. Inthe past, effects of load [1,2], temperature [3] and reinforcementvolume fraction [1,2,4] on wear properties of TiB2/Al compositeshave been reported. Kumar et al. [1] investigated Al–7Si/10TiB2

composites and concluded that adhesion and ploughing are pre-dominant at lower loads and delamination is more predominantat higher loads. Natarajan et al. [3] found that the wear mechanismof Al 6063/TiB2 composites at temperature less than 100 �C is abra-sive wear, above 200 �C is oxidative wear. Mandal et al. [2] inves-tigated effect of TiB2 particles on sliding wear behaviour of Al–4Cualloy, and found that wear resistance increases with increase in theamount of TiB2. However, the surface roughness of frictional pair(usually steel), which also plays an important role on tribologicalproperties, has not been well investigated yet. For convenience,the ‘‘surface roughness’’ in the following paragraphs represents‘‘surface roughness of frictional pair’’.

About the effect of surface roughness on tribological properties,Amontons (1699) and Bowden proposed the most simple frictionmodel: the friction coefficient f = tanu, where u denoted the anglebetween the asperity and the surface, from the formula it can beconcluded that the smoother the surface is, the lower the frictioncoefficient is. However, these theories are too simple to explainthe complex tribological phenomena, and the do not fully consistwith the situation encountered in practice [5]. In 1975, Bayer and

They found that wear was more sensitive to surface roughnessvariations for the finer surface (<V16) and wear was more sensitiveto orientation of the surface roughness for the coarser surface(>V16) [6]. Wang and Rack [7] investigated the friction and wearbehaviour of 2124 Al–SiCw/17-4 PH stainless steel systems underdry sliding wear using a pin-on-disc apparatus, and they concludedthat reducing the surface roughness of the sliding surfaces signifi-cantly reduced initial wear rates in both surfaces, but had no effecton the steady-state wear rate of either the pin or the disc. Ho-Chiehet al. [8] investigated the effect of surface roughness on friction ofceramics (Si3N4, SiC and Al2O3) sliding in water. They founded thatfor Si3N4 and SiC, smaller surface roughness provided smaller fric-tion coefficient in run-in process. However, for Al2O3, specimen of0.320 lm Rrms shows larger friction coefficient than other Al2O3

specimens of the same surface roughness. Sahin et al. [9] studiedeffect of surface roughness on friction coefficients during upsettingprocesses for different materials (steel, commercially pure alumin-ium and annealed CuZn40Pb2 brass). Results showed that the fric-tion coefficient depended on surface roughness where the roughersurfaces gave lower friction coefficients for all materials.

From the reports above, it can be founded that the effect of sur-face roughness on friction and wear is very complex, and it is notsure that reducing surface roughness is better or not for the specialfrictional system. At present, to find optimal surface roughness, alarge number of tests must be carried out [7–10]. However, fewpeople considered the effect of microstructure on the optimal sur-face roughness, and the relationship between the two in trilobogy.

In the present paper, 55 vol.% TiB2/2024Al composites were fab-ricated by pressure infiltration method and the effect of surface

130 S.F. Tian et al. / Materials and Design 53 (2014) 129–136

roughness of GCr15 steel disc on tribological properties of compos-ites was investigated to find the optimal surface roughness.

2. Experimental details

55 vol.% TiB2/2024Al composites were prepared by pressureinfiltration method. The average size of TiB2 is 1.6 lm. The pre-heating temperature of perform was between 600 and 650 �C, cast-ing temperature was between 750 and 770 �C and casting pressurewas between 70 and 80 MPa. Phase analysis was characterizedusing D/max-c b type X-ray diffraction instrument. FEI SIRION200scanning electron microscope (SEM) and JEOL200CX transmissionelectron microscope (TEM) were used to characterize themicrostructures.

The dry sliding wear tests were conducted on a pin-on-discwear testing machine (Model: CJS111A, Harbin, China) accordingto the ASTM: G99-05 (2010). The composite was first processedinto the dimensions of U3 � 13 mm of the rod. The end of samplewas then machined into hemispherical shape with the radius of1.5 mm. Finally, samples were soluted at 495 �C for 1 h, quenchedto room temperature in water, and then aged at 160 �C for 10 h.The counterface material used was GCr15 steel (China brand, cor-responding to AISI52100 bearing steel), hardened to HRC60.

In order to investigate the effect of surface roughness of GCr15steel disc on the tribological properties of composite, surfaceroughness with values of Ra 0.005, 0.023, 0.372 and 0.606 lm wereselected. Specially, Ra 0.005 lm was obtained by manual lappingand polishing, the other three were obtained by grinding machineswith different grinding wheel. Ra values were tested by atomicforce microscope (AFM) (NanoscopeIIIa, DI), according to the ISO:4288 [11], 3D profiles were shown in Fig. 1.

Friction and wear experiments, in which the load was fixed to50 g and the sliding speed, were from 1.0 to 2.0 m/s. The slidingtime for all the tests was 30 min. Before the test, the compositewas polished, and pin and disc were washed by acetone to removedirt. The worn surface of pin and disc was observed by scanning

Fig. 1. 3D profiles of disc for d

electron microscope. The worn surfaces of pin were also analyzedusing PHI5700 ESCA system X-ray photoelectron spectroscopy(XPS). Since one end of pin is hemisphere, after sliding, the shapeof wear volume is spherical segment. By SEM observation of wornsurface and corresponding mathematic equations, it is easily to cal-culate wear volume [12]. And the wear rate can be calculated as:

Wear rate ¼Wear volume=ðLoad� Distance TravelledÞ ð1Þ

3. Results and discussion

3.1. Microstructure of TiB2 after heat preservation

Fig. 2 gives TEM and HREM microstructure of the surface of aTiB2 particle, which was taken from the preform pre-heated at640 �C for 2 h. It can be seen that there was a layer with the thick-ness of 20–30 nm on the surface of TiB2 particle. Local amplifica-tion of the layer is showed in Fig. 2b. Some nanoparticles wereformed in this layer, which resulted from the heat preservationof perform. Fig. 2c is the HREM image of Fig. 2b. After calibration,it was found that the layer on the surface was made up of TiO2 andB2O3 nanoparticles with grain size of about 5 nm. The appearanceof TiO2 and B2O3 resulted from the reaction of TiB2 with oxygenduring the heat preservation at 640 �C of perform. This reaction fol-lowed the following equation,

TiB2 þ 5=2O2 ! TiO2 þ B2O3 ð2Þ

The reaction Gibbs free energy change at 640 �C is �1552.14 KJ/mol, based on JANAF Thermochemical Tables [13], which indicatesthat this reaction can take place spontaneously.

Although no reaction would be occurred between pure TiB2

phase and Al [14], the reaction between the formed nano-oxidelayer and molten Al would modify the TiB2/Al interface, and affectfriction behaviour of TiB2/2024Al composites.

ifferent surface roughness.

Fig. 2. TEM and HREM images of the surface of a TiB2 particle, which was taken from the preform pre-heated at 640 �C for 2 h. (a) TEM images of surface; (b) magnification of(a) and (c) HREM image.

Fig. 4. X-ray diffraction spectrum of 55 vol.% TiB2/2024Al composite.

S.F. Tian et al. / Materials and Design 53 (2014) 129–136 131

3.2. Microstructure of TiB2/2024Al composite

Representative morphology of TiB2/2024Al composite wasshown in Fig. 3, TiB2 particles were distributed uniformly withoutany particle clustering and no apparent porosity or significant cast-ing defects were observed in the composites.

It can be seen from Fig. 4 that there were only two phasesincluding TiB2 and a-Al in the composite, and almost no otherphase could be observed, especially Ti–Al compounds. This is quitedifferent from the composites that are fabricated by in situ reaction[4,15], because molten aluminium is directly pressed into the pre-pared TiB2 preform by the method of pressure infiltration methodunder external forces instead of synthesizing TiB2 by chemicalreaction, which is inevitably accompanied with by-product. Byobserving the high-resolution electron microscope image (Fig. 2),it could be found that a layer consisting of TiO2 and B2O3 couldbe observed on the surface of TiB2 particles after pre-heating, butthis could not be seen in XRD pattern. This was because the thick-ness (20–30 nm) of this surface layer was too small compared withthe size (1.6 lm) of TiB2 particle. As a result, the content of TiO2

Fig. 3. SEM microstructure of 55 vol.% TiB2/2024Al composite.

and B2O3 was rather low compared with the whole material, andlower than the XRD detection limit (5%). Consequently, TiO2 andB2O3 phases couldn’t be observed in XRD diffraction pattern.

Representative microstructure of TiB2/Al interface was shownin Fig. 5. Some interface of TiB2 particles and matrix were straightand clean. However, there were also some interface products, asindicated by the white arrow in Fig. 5a, different from Mitra’s re-port [14]. The reactant product in Fig. 5b is identified as TiO2,and the reactant product in Fig. 5c and d is identified as B2O3.The formation of TiO2 and B2O3 is from the oxidation of TiB2 (Eq.(2)), but we also find Al2O3 (Fig. 5c) and AlB2 (Fig. 5d), and bothof two are near B2O3, therefore we can deduce the formation ofAl2O3 and AlB2 are related to B2O3.

At the casting temperature (700–750 �C), the Gibbs free energychange of B2O3 reacts with Al (Eq. (3)) is �486.96 KJ/mol, based onJANAF Thermochemical Tables [13], which indicates that this reac-tion can take place, as shown in Eq. (3). Hence, the generation ofAl2O3 and AlB2 is explained. Moreover, the interface with partialreactant has higher bonding strength than TiB2–Al interface with-out any reactant, and maybe better wear resistance.

B2O3 þ 3Al ¼ Al2O3 þ AlB2 ð3Þ

Fig. 5. TEM and HREM images of 55 vol.% TiB2/2024Al composite (a) TEM image of composite and (b–d) HREM of image of interface reactant.

132 S.F. Tian et al. / Materials and Design 53 (2014) 129–136

3.3. Frictional behaviours of TiB2/2024Al composite

Figs. 6 and 7 show the curves of friction coefficients and wearrate of TiB2/Al composite against GCr15 bearing steel as a functionof sliding speed and surface roughness under dry sliding condition(P = 50 g). It can be seen that friction coefficient and wear rateshowed great difference with different surface roughness condi-tions. Friction coefficient changed in the range of 0.35–0.5 and0.25–0.55 for Ra 0.602 lm and Ra 0.372 lm, respectively. Itreached the lowest value of 0.08–0.15 if Ra declined to 0.023 lm.However it increased to 0.4–0.6 with the surface roughness furtherdecreased to 0.005 lm. Wear rate showed the same trend with the

Fig. 6. Curves of friction coefficients of TiB2/Al composite against GCr15 bearingsteel as a function of sliding speed and surface roughness under dry slidingcondition.

Fig. 7. Curves of wear rate of TiB2/Al composite against GCr15 bearing steel as afunction of sliding speed and surface roughness under dry sliding condition.

friction coefficient, possessing the lowest value with the Ra0.0023 lm surface roughness.

Under the same speed, with the decrease of Ra, the wear ratefirst declined (Ra 0.606 lm ? Ra 0.023 lm) and then ascended(Ra 0.023 lm ? Ra 0.005 lm), shown in Fig. 7. For friction coeffi-cient, it was a little complex, since variation tendency at 1.5 and2.0 m/s were not evident, but the overall trend was still firstlydown and then up, shown in Fig. 6. Friction coefficient and wearrate for Ra 0.023 lm was the lowest. Tribological properties withthe variation of Ra are different from other people’s works [6–9,16]. For example, Sahin et al. [9] suggested the rougher surfacesgave lower friction coefficients. Menezes et al. [16] indicated thecoefficient of friction does not vary much with surface roughness.

Fig. 8. worn surface of composite under the conditions of different surface roughness. (a) Ra 0.606 lm; (b) Ra 0.372 lm; (c) Ra 0.023 lm and (d) Ra 0.005 lm.

S.F. Tian et al. / Materials and Design 53 (2014) 129–136 133

The different variation of tribological properties with surfaceroughness was mainly caused by the complexity of tribology.

Under the same surface roughness, sliding speed had little effecton friction coefficient and wear rate. There was an optimal value ofsurface roughness, Ra 0.023 lm, under which friction coefficientand wear rates reached the lowest value (Figs. 6 and 7). This is inaccordance with the massive experience of practice, which is for a gi-ven tribological system, there is an optimal surface roughness to ob-tain the lowest friction coefficient and wear rate [17].

3.4. Friction mechanism

3.4.1. Analysis of TiB2/Al compositeIn order to investigate the wear mechanism, worn surfaces of

composite materials were observed, as shown in Fig. 8. For differ-ent surface roughness the worn surfaces were quite different.When Ra was 0.606 lm (Fig. 8a), there were obvious traces ofploughing, indicating a typical two-body abrasive wear, and the

Fig. 9. XPS spectrum of worn surface of pin sample under dry sliding co

reason was that GCr15 steel disc was rough and harder thanTiB2/Al composites, when GCr15 surface asperities sliding on thesurface of composite materials, lots of ploughing appeared, There-fore, the coefficient of friction and wear rate were high. When Rawas 0.372 lm (Fig. 8b), worn surface was smooth, but part of itwas covered with the plastic flow of composites, showing theadhesive wear (showed by arrow). This was the reason why thefriction coefficient was not high, but the wear rate was high. WhenRa was 0.023 lm (Fig. 8c), the worn surface is smooth and no obvi-ous adhesion, ploughing, showing the lowest friction coefficientand wear rate. The detail analysis of wear mechanism was in thenext paragraph. Further reduced surface roughness to Ra0.005 lm (Fig. 8d), there were lots of smearing and delaminationcaused by adhesion friction coefficient and wear rate was veryhigh. Shortly, for Ra 0.606, 0.372 and 0.005 lm, the wear mecha-nisms were adhesive and abrasive, similar to most reports [1–4,18,19], whereas for Ra 0.023 lm, only similar to Zhao’s work[12]. The variation of fabrication method, volume fraction and testconditions may give the corresponding interpretation.

ndition (Ra = 0.023 lm and V = 1.0 m/s): (a) boron and (b) titanium.

Fig. 10. EDS analysis of worn surface of GCr15 steel disc under conditions of different surface roughness (V = 1.0 m/s) (a) Ra 0.606 lm; (b) Ra 0.372 lm; (c) Ra 0.023 lm and(d) Ra 0.005 lm.

134 S.F. Tian et al. / Materials and Design 53 (2014) 129–136

S.F. Tian et al. / Materials and Design 53 (2014) 129–136 135

As seen from Fig. 8c, for Ra 0.023 lm, the wear mechanism wasnot adhesive and abrasive wear. In order to determine the wearmechanism, the worn surfaces were characterized by means ofXPS analysis; the results were shown in Fig. 9.

B2O3 and TiO2 were observed on worn surface, both of whichwere oxidational products of TiB2. H3BO3 was also observed, andself-lubricating, with layered triclinic crystal structure like graph-ite, MoS2, can reduce friction greatly [20–22]. The formation ofH3BO3 is from B2O3 reaction with water vapour in the air (B2O3 + -H2O = H3BO3), and the reaction is spontaneous [23]. Consideringthe smooth worn surface and existence of B2O3 and TiO2, for Ra0.023 lm, the main mechanism was oxidational wear.

There are two resources about boric acid. Firstly, generated inthe frictional process. During sliding, the temperature betweenthe contact surfaces was high, which can accelerate the tribo-chemistry reaction of TiB2, generating B2O3 and TiO2, furthermoreboric acid formed [12]; secondly, introduced by pre-heating. Afterpre-heating, there was one 20 nm oxidation layer on the surface ofTiB2 in composites, constituted of TiO2 and B2O3, furthermoreintroduced into the composite, as shown in Figs. 2c and 5. Andthe thickness of oxidation layer is coincided with the optimum sur-face roughness (Ra 0.023 lm). Therefore, at the initial frictionalprocess, the oxidation layer filled with the gap between the asper-ities of GCr15, weakening the adhesion of aluminium alloy andGCr15; meanwhile once B2O3 was exposed in the air, H3BO3 wasformed, resulting in low friction coefficient. Oxide introduced bypre-heating played a part in the running-in stage, and the oxida-tion of TiB2 during sliding can keep the low friction at the steadystage.

When surface roughness was high (Ra 0.606 lm and Ra0.372 lm), the size of oxidation products generated by pre-heatingwere too small to fill the gap between the asperities of GCr15 sur-face, therefore could not reduce friction. The wear mechanism wasadhesion and abrasion, as verified by Fig. 8a and b. Furthermore,once the adhesion and abrasion initiated, they would be sustain-able. Therefore the thin layer (20–30 nm) containing B2O3 cannotreduce friction, and friction coefficients and wear rates were high.When the surface roughness reduced to Ra 0.005 lm, the mole-cule-attraction played a major role during sliding, which enhancedadhesion, and friction coefficient is high, Fig. 8d verified this well.Ho-Chieh et a. [8] founded that for Si3N4 and SiC, smaller surfaceroughness provided smaller friction coefficient in run-in process,different from the present investigation. It was to be noted that,the lowest surface roughness they investigated was only0.092 lm, if they decreased the surface roughness to more low va-lue, the results may be different.

3.4.2. Analysis of GCr15The present work is focused on the surface roughness of GCr15

on tribological properties of TiB2/Al–GCr15 frictional pair. And tri-bological properties are systematic; therefore the investigation isvery important.

There were two aspects concerning the influence of GCr15.Firstly, the hardness of GCr15 was between matrix alloy andTiB2, the asperity on the surface of GCr15 would plough on thecomposites during sliding, as shown in Fig. 8a. Meanwhile, TiB2

would plough GCr15 surface as well. Secondly, the main compo-nent Fe element in GCr15 is 100% mutual soluble with aluminiumin liquid state [24], Al is adhesive with Fe greatly, which would af-fect the frictional and wear behaviour.

Element analysis of GCr15 worn surface on the condition ofsliding speed of 1.0 m/s and load of 50 g by EDS was shown inFig 10. Fig. 10a presents the worn surface of GCr15 steel disc whenthe wear roughness was Ra 0.606 lm. There was adhesive sub-stance on the worn surface. Surface energy spectrum analysisshowed that aluminium element content was high at the region

of adhesive substance and transferring of elements occurred, indi-cating the adhesive wear. Fig. 10b gives the worn surface of GCr15steel disc when the wear roughness was Ra 0.372 lm, EDS wassimilar to that of Ra 0.606 lm, showing adhesive wear. Fig. 10cgives the worn surface of GCr15 steel disc when the wear rough-ness was Ra 0.023 lm. The worn surface was smooth and adhesivewear didn’t occur. EDS analysis of the surface of GCr15 steel discshowed that there were mainly Fe, C, and Cr elements (all werecomposition of GCr15), no aluminium element. Al and Fe were easyto adhere [24], but adhesion was not observed on the surface ofGCr15 steel disc, therefore we can conclude that the compositeand steel disc were separated from each other by some substance,combing Fig. 9, the substance was identified as TiO2, B2O3 andH3BO3. They prohibited the occurrence of adhesive wear. Fig. 10dpresents the worn surface of GCr15 steel disc when the wearroughness was Ra 0.005 lm. Serious smear could be seen on wornsurface. Aluminium element content is high at this region andtransferring of elements occurred, indicating the severe adhesivewear occurred. Because the surface was too smooth, the two inter-molecular interactions between the surfaces played a major role,resulting in higher friction coefficient and wear rate [17].

4. Conclusions

55 vol.% TiB2/2024Al composites were fabricated by pressureinfiltration method. Effect of surface roughness was investigated.Results showed there was an optimal surface roughness (Ra0.023 lm), the size of which was in the same order of thicknessof oxide layer formed by pre-heating of TiB2. During sliding, theoxide was filled into the asperity gap of GCr15, greatly reducedadhesion between Al and GCr15, meanwhile H3BO3 was formed,therefore the friction coefficient and wear rate were low.

Acknowledgements

The microstructure analysis assistance of L.T. Jiang is greatlyappreciated. This work was supported by Program for New CenturyExcellent Talents in University (No. NCET-07-0234).

References

[1] Kumar S, Chakraborty M, Subramanya Sarma V, Murty BS. Tensile and wearbehaviour of in situ Al–7Si/TiB2 particulate composites. Wear2008;265:134–42.

[2] Mandal A, Chakraborty M, Murty BS. Effect of TiB2 particles on sliding wearbehaviour of Al–4Cu alloy. Wear 2007;262:160–6.

[3] Natarajan S, Narayanasamy R, Babu SPK, Dinesh G, Kumar BA, Sivaprasad K.Sliding wear behaviour of Al6063/TiB2 in situ composites at elevatedtemperatures. Mater Des 2009;30:2521–31.

[4] Yanfeng H, Xiangfa L, Xiufang B. In situ TiB2 particulate reinforced near eutecticAl–Si alloy composites. Compos Part A-Appl Sci 2002;33:439–44.

[5] Shizhu W, Ping H. Principles of tribology. 2nd ed. Beijing: Tsinghua UniversityPress; 2008.

[6] Bayer R, Sirico J. The influence of surface roughness on wear. Wear1975;35:251–60.

[7] Wang A, Rack H. Dry sliding wear in 2124 Al–SiCw/17-4 PH stainless steelsystems. Wear 1991;147:355–74.

[8] Ho-Chieh W, Noritsugu U, Koji K. The effect of surface roughness on friction ofceramics sliding in water. Wear 1998;218:237–43.

[9] Sahin M, Cetinarslan CS, Akata HE. Effect of surface roughness on frictioncoefficients during upsetting processes for different materials. Mater Des2007;28:633–40.

[10] Menezes Pradeep L, Kishore, Kailas Satish V. Influence of surface texture androughness parameters on friction and transfer layer formation during slidingof aluminium pin on steel plate. Wear 2009;267:1534–49.

[11] ISO 4288. Geometrical product specifications (GPS) surface texture: profilemethod on the assessment of surface texture. Int Stan Org; 1996.

[12] Zhao M, Wu GH, Jiang LT, Dou ZY. Friction and wear properties of TiB2P/Alcomposite. Compos Part A-Appl Sci 2006;37:1916–21.

[13] Chase Jr Malcolm W. JANAF thermochemical tables. 4th ed. Maryland: AIP;1998.

[14] Mitra R, Chiou WA, Fine ME, Weertman JR. Interfaces in as-extruded XD Al/TiCand Al/TiB2 metal matrix composites. J Mater Res 1993;8:2380–92.

136 S.F. Tian et al. / Materials and Design 53 (2014) 129–136

[15] Lü L, Lai MO, Su Y, Teo HL, Feng CF. In situ TiB2 reinforced Al alloy composites.Scripta Mater 2001;45:1017–23.

[16] Menezes Pradeep L, Kishore, Kailas Satish V. Influence of surface texture androughness parameters on friction and transfer layer formation during slidingof aluminium pin on steel plate. Wear 2009;267:1534–49.

[17] Jiashu S. Wear of metals. Beijing: Metallurgical Industry Press of China;1992.

[18] Kumar S, Subramanya Sarma V, Murty BS. High temperature wear behaviour ofAl–4Cu–TiB2 in situ composites. Wear 2010;268:1266–74.

[19] Niranjan K, Lakshminarayanan PR. Dry sliding wear behaviour of in situ Al–TiB2 composites. Mater Des 2013;47:167–73.

[20] Erdemir A, Bindal C, Zuiker C, Savrun E. Tribology of naturally occurring boricacid films on boron carbide. Surf Coat Technol 1996;86–87:507–10.

[21] Erdemir A, Halter M, Fenske GR. Preparation of ultralow-friction surface filmson vanadium diboride. Wear 1997;205:236–9.

[22] Deshmukh Pushkarraj, Lovell Michael, Gregory Sawyer W, Mobley Anton. Onthe friction and wear performance of boric acid lubricant combinations inextended duration operations. Wear 2006;260:1295–304.

[23] Bharat Bhushan. Modern tribology handbook: principles of tribology [M]. CRCPress; 2001.

[24] ASM Handbook. Friction, lubrication, and wear technology. vol. 18, 10thed. ASM International; 1992.