Inuence of the normal force, abrasive slurry concentration andabrasive wear modes on the coefcient of friction in ball-crateringwear testsRonaldo Cmara Cozzaa,b,naCentro Universitrio da FEI Fundao Educacional InacianaPadre Sabia de Medeiros, Departamento de Engenharia Mecnica, Av. Humberto deAlencar Castelo Branco, 3972, So Bernardo do Campo, SP 09850-901, BrazilbCEETEPS Centro Estadual de Educao TecnolgicaPaula Souza FATEC-Mau, Av. Antnia Rosa Fioravante, 804, Mau, SP 09390-120, Brazila rti cle in foArticle history:Received 24 September 2012Received in revised form16 January 2013Accepted 13 September 2013Available online 8 October 2013Keywords:Ball-crateringAbrasive wearCoefcient of frictionabstractThe purpose of this work is to study the inuence of the normal force (N), abrasive slurry concentration(C) and abrasive wear modes on the coefcient of friction in ball-cratering wear tests. Experiments wereconducted with balls of AISI 52100 steel, an AISI H10 tool-steel specimen and abrasive slurries preparedwithblacksiliconcarbide(SiC) particlesdistilledwater. Thetangential (T) andnormal loadsweremonitored throughout the tests and the results have shown that: (i) the coefcient of friction behaviorwas independent of thenormal forceand(ii) boththeconcentrationsof abrasiveslurries andthesubsequent action of the abrasive wear modes, generally, did not affect the behavior or magnitude of thecoefcient of friction.& 2013 Elsevier Ltd. All rights reserved.1. IntroductionRecently, themicro-scaleabrasiveweartesthasgainedlargeacceptance at universities and research centers and is widely usedinstudiesfocusingontheabrasivewearofmaterials. Fig. 1a[1]presentsaschematicdiagramof theoperatingprinciplesof theabrasivewear test, wherearotatingball is forcedagainst thespecimen being tested in the presence of an abrasive slurry. Thereare two main equipment congurations that can be used toconductthistypeoftest:thefree-ballandxed-ballcong-urations. Fig. 1b[2]and1c[3,4]showexamplesoftheseequip-ment congurations.Theaimofthemicro-abrasiveweartestistogeneratewearcraters on the specimen being tested. Fig. 2 presents representa-tiveimagesof suchcraters, together withanindicationof thecraterdiameter(b)[5](Fig. 2a), thecraterdepth(h)(schematicillustration) (Fig. 2b) and the wear volume (V) [2] (Fig. 2c).Thediameter of thewear crater is commonlymeasuredbyoptical microscopy, but other methods are available. For example,computer aided design (CAD) software [3] has been used for thispurpose. The crater depth and the wear volume may bedetermined as a function of b, using Eqs. (1) and (2) [6],respectively, where R is the radius of the ball.hb28Rfor b{R 1V b464Rfor b{R 2Two abrasive wear modes are usually observed on the surfaceof the worn crater:grooving abrasion results when the abrasiveparticles slide onthe specimen(Fig. 3a [7]), while rollingabrasion is observedwhenthe abrasive particles roll onthesurface of the specimen (Fig. 3b [7]). Depending on testconditions, rollingabrasionandgroovingabrasioncanoccursimultaneously in a given crater [8]. Fig. 3c [1], 3d [8] and 3e [3]presents images of groovingabrasion, rollingabrasionandthesimultaneous action of rolling and grooving abrasion, respectively.In a previous work [3], At was dened as the total projected area ofthe crater and Ag as the projected area with grooving abrasion. Theprojected area with rolling abrasion (Ar) may be dened asArAtAg [9].The micro-abrasive wear test has been applied toward studyingtheabrasivewear of metallic [2,3,8] andnon-metallic [3,4,10]materials where, depending on the equipment conguration, it ispossibletoapplynormalloads(N)from0.01 N[11]to10 N[12]and ball rotational speeds (n) up to 80 rpm [13].The wear behavior of different materials can be analyzed basedonthedimensionsofthe wear crater and/oronthe wearmode.Contents lists available at ScienceDirectjournal homepage: www.elsevier.com/locate/tribointTribology International0301-679X/$ - see front matter& 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.triboint.2013.09.010nCorrespondenceaddress:CentroUniversitriodaFEI FundaoEducacionalInacianaPadre Sabia de Medeiros, Departamento de Engenharia Mecnica,Av.Humberto de Alencar Castelo Branco, 3972, So Bernardo do Campo, SP 09850-901,Brazil. Tel.: 55 11 4353 2900; fax: 55 11 4109 5994.E-mail addresses: rcamara@fei.edu.br, ronaldo.cozza@fatec.sp.gov.brTribology International 70 (2014) 5262NomenclatureAgprojected area with grooving abrasion [mm2]Arprojected area with rolling abrasion [mm2]Attotal projected area of the wear crater [mm2]b diameter of the wear crater [mm]C concentration of the abrasive slurryCAD Computer Aided DesignD diameter of the ball [mm]DSRW Dry Sand Rubber WheelDW distilled waterh depth of the wear crater [mm]k wear coefcient [mm3/(N m)]n ball rotational speed [rpm]N normal force [N]R radius of the ball [mm]S sliding distance [m]t test time [s]T tangential force (friction force) [N]TAStangential force (friction force) between abrasive par-ticles and specimen [N]TBAtangential force (friction force) between ball andabrasive particles [N]v tangential sliding velocity [m/s]V wear volume (volume of the wear crater) [mm3]Greek letters coefcient of frictionAScoefcientoffrictionbetweenabrasiveparticlesandspecimenBAcoefcient of friction between ball and abrasiveparticlesShaftSpecimenBall ShaftSpecimenBall Tangential forceTNormal forceNSpecimen Ball Abrasive slurry n Fig. 1. Micro-abrasive wear testing using the rotating ball method: (a) schematic diagram of the operating principle [1], (b) free-ball conguration [2] and (c) xed-ballconguration [3,4].V h Specimen Abrasive particles (in green) Ball Fig. 2. Representative images of wear craters: (a) diameter b [5], (b) crater depth h (schematic illustration) and (c) wear volume V [2].R.C. Cozza / Tribology International 70 (2014) 5262 53Since the early works of Hutchings [6,8], several other importantcontributions have been documented in terms of this type of test,suchasthewearmodetransition[9,14], the wear coefcient(k)[13,15], micro-abrasive wear of coatedsystems [16,17], micro-contact modeling of abrasive wear [18,19], ridge formation [20,21]and angularity of abrasive particles [22,23].Manyworksonfrictioncoefcient () duringabrasivewearand other types of tests are available in the literature [2431], butonlyafewwerededicatedtothefrictioncoefcient inmicro-abrasive wear tests with a rotating ball [3234].The purpose of this work is to study the inuence of the normalforce,abrasive slurry concentration andabrasive wearmodesonthe friction coefcient with the intent of contributing toward theunderstandingof thefrictioncoefcientbehaviorinmicro-scaleabrasion wear tests by the rotative ball method.2. Experimental details2.1. Micro-abrasive wear test equipmentThe rotative ball method withxed-ball conguration (Fig. 4)wasusedin themicro-scaleabrasive weartests. Thisequipmentwas designed and assembled with some mechanical and electricaldifferences from thexed-ball equipment conguration found inthe literature [8,1012,35].Differingfromthecommerciallyavailablesystemswheretheball isxedbytwoshafts[11](Fig. 5a), theball isstuckontheshaft withadhesivematerial inthetest apparatususedinthiswork (Fig. 5b). The surface of the tip is concave, with a tip radiusequal tothatofthesphere(R12.7 mm). Thissurfaceiscoatedwith adhesive CIANO CM1, and the sphere is pressed against thisadhesive, xing the sphere in position.Tocancel theeccentricity(misalignment) of thesphere, theshaft was designed to allow its displacement in a single plane intwo perpendicular directions, as schematized in Fig. 5c.Therotationofthesphereiscontrolledbyacoupledservo-motor/servo-controller, bought fromtheRexrothBoschGroup;this systemallows theselectionof ball rotational speeds from105rpmupto9103rpm, inboththeclockwiseandcounter-clockwise directions. A load cell controls the normal force (Fig. 4b),whichis appliedonthe specimenwiththe helpof a secondcoupled servo-motor/servo-controller (also provided by theRexrothBoschGroup)thatrotatesascrewfortheapplicationofthe normal force. A second load cell located below the specimen(Fig. 4b) measures the tangential force (T) generated duringthe tests.2.2. MaterialsTheexperimentswereconductedwithoneball madeof AISI52100 steel (nominal chemical composition of 1.04 wt% C, 0.35 wt%Abrasive particles 100m AgArAtFig. 3. Principle of (a) grooving abrasion [7] and (b) rolling abrasion [7]. Abrasive wear modes: (c) grooving abrasion [1], (d) rolling abrasion [8] and (e) the simultaneousaction of rolling abrasion and grooving abrasion [3].R.C. Cozza / Tribology International 70 (2014) 5262 54Mn, 0.25 wt%Si, 1.45 wt%Cr, bal. Fe), witha diameter (D) of25.4 mm(1). Thetestedspecimenwas composedof quenchedand tempered AISI H10 tool steel (nominal chemical compositionof0.42 wt% C, 0.81 wt% Si, 0.42 wt% Zr, 0.26 wt% S, 0.19 wt% V, 0.65 wt% Cr, 1.68 wt% Mn, 0.46 wt% Ni, bal. Fe), with dimensions of 5 mm(thickness) 20 mm (width) 50 mm (length).Fig.6a [3] and 6b presentthe microstructures ofthe ball andspecimen, respectively. The ball exhibited carbides homoge-neouslydistributedinthematrix, whilethetest specimendis-played a martensitic microstructure with retained austenite.The abrasive used was black silicon carbide (SiC) with an averageparticle size of 3 m. Fig. 7 [1] presents a micrograph of the abrasiveparticles (Fig. 7a) and the particle size distribution (Fig. 7b).Table 1 presents the hardness of the materials used in this work(specimen, ball [3,4] and abrasive particles [3,4]).2.3. Micro-abrasive wear testsTable 2 shows the test conditions selected for the experimentsconducted in this work.Two values of normal force were dened for the wear experi-ments: N10.5 NandN21.25 N; threevalues of theabrasiveslurry concentrationwere dened: C15%SiC 95%distilledwater, C225%SiC 75%distilledwater andC337.5%SiC 62.5%distilledwater(volumetricvalues). AftertheN1, N2, C1, C2and C3, values were established, six different combinations of thenormal force and the abrasive slurry concentration werecompiledintosixdifferent test conditions: N1C1, N1C2,N1C3, N2C1, N2C2 andN2C3.Theball rotational speedwas n37.6 rpm, whichwas pre-viouslyselectedbyTrezonaetal. [8]andAdachi andHutchings[11,14]. Forn37.6 rpmandD25.4 mm (R12.7 mm), thetan-gential sliding velocity at the external diameter of the ball is equalto v0.05 m/s, which potentially reduces or eliminates the occur-rence of hydrodynamic effects during the tests [11].Thetests wererunfor threedifferent slidingdistances (S),S110 m, S232 mandS3100 m. Thesevalueswerebasedonthe Renard's series R20/4 [36] and the correspondent test timesare t1200 s (3 min 20 s), t2640 s (10 min 40 s) and t32000 s(33 min 20 s), as presented in Table 2.Three repetitions were conducted for each S value, and the testsequence was thesameforthesixdifferentcombinationsofthenormal force (N) and the abrasive slurry concentration (C) (NC).This sequence was randomly dened as follows: 32, 10, 32, 10, 10,Vertical displacement of the ball. Horizontal displacement of the ball. Fig. 5. (a) Ballxing in the Plint TE 66 test equipment courtesy photo: Prof. Amilcar Lopes Ramalho University of Coimbra POR, (b) ballxing used in this work and(c) displacement in the two perpendicular directions of the sphere.Sphere Load cell for normal force controll. Load cell for tangential force measurement. Specimen Displacement of the specimen. Fig. 4. (a) General viewof themicro-abrasivewear test equipment withthe xed-ball congurationusedintheexperiments for thisworkand(b) loadcells formeasurements of the normal and tangential forces.R.C. Cozza / Tribology International 70 (2014) 5262 5532, 100, 100and100 m, whichprovides fty-four experiments(nine tests for each combination ofNC).All tests were conducted without interruption, and the abrasiveslurrywascontinuouslyagitatedandfedbetweentheball andspecimen with the help of a peristaltic pump.At the end, after all tests, the abrasive slurries were analyzed bylaser interferometry to verify the possible occurrence of abrasiveparticles fragmentation.2.4. Data acquisition and result analysisBoththenormal force(N) andthetangential force(T) weremonitored and registered constantly with a data acquisitionsystem, under a frequency of 100 Hz.Then, the friction coefcient was determined using Eq. (3): TN33. Results and discussion3.1. Nulling the eccentricity of the ballDuring the design and construction of the micro-abrasive weartest mentioned in Section 2Experimental details, there was agreat degree of difculty involved in removing the eccentricity ofthe ball. Initially, the purpose was to cancel the eccentricity of theshaft but was unable to do so because there were some misalign-ment issues with the sphere positioning.Intotal, fourdesignsandvedifferent machiningoperationtests were conducted. It was eventually understood that theproblem was not in the machine design or the machining opera-tions, but rather on the sphere and not the shaft.Finally, theeccentricityof thesphereinthemicro-abrasivewear test equipment by the xed rotating ball method wasremovedwithanunconventional design, andtheexperimentscould be performed accurately and reliably.Instudies of micro-abrasive wear test by the rotating ballmethod, the annulment of sphere misalignment is critical since aneccentricity between 20 and 24 m, as reported by Gee and Wicks[37], caninuencetheresults, particularlythemeasurementofthecoefcient of friction. Thejusticationfor this assertionissupportedbytheillustrationshowninFig. 8. Inmanyresearcheffortsrelatedtotestingof micro-abrasivewearbytherotatingball method, theaveragesizeof theabrasiveparticlestakenbythe respective scientists is approximately 4 m[24,8,11,14],amisalignmentbetween20and24 mis56timeslargerthanthe abrasive particles themselves. Consequently, the ball will1005001010.10.04 0 80 100 60 40 20 Cumulative values [%] Histogram [.10] Fig. 7. SiC abrasive [1]: (a) scanning electron micrograph and (b) particle size distribution.Table 1Hardness of the materials.Material Hardness GPa (HV)Specimen H10 tool steel 7.1 (720)Ball AISI 52100 steel 8.4 (856) [3,4]Abrasive particles SiC 18.519 (18861937) [3,4]10m Fig. 6. Microstructure of:(a)AISI52100steel ball[3]and(b)AISI H10toolsteelspecimen. Both ball and specimen were chemically etched with 3% Nital.R.C. Cozza / Tribology International 70 (2014) 5262 56Table 2Test conditions selected for the wear experiments.Test condition ) 1 2 3Normal force N1 [N] 0.5 0.5 0.5Normal force N2 [N] 1.25 1.25 1.25Abrasive slurry concentration C1 (in volume) 5% SiC 5% SiC 5% SiC95% distilled water 95% distilled water 95% distilled waterAbrasive slurry concentration C2 (in volume) 25% SiC 25% SiC 25% SiC75% distilled water 75% distilled water 75% distilled waterAbrasive slurry concentration C3 (in volume) 37.5% SiC 37.5% SiC 37.5% SiC62.5% distilled water 62.5% distilled water 62.5% distilled waterSliding distance S [m] 10 32 100Ball rotational speed n [rpm] 37.6 37.6 37.6Tangential sliding velocityv [m/s] 0.05 0.05 0.05Test time t 200 s 640 s 2000 s(3 min 20 s) (10 min 40 s) (33 min 20 s)Number of repetitions 3 3 3 The sphere is always in contact with the specimen. SpecimenSphere Sphere trajectory In most of the time, the sphere will be out of touch with the specimen. Abrasive particles Fig. 8. (a) Sphere without eccentricity and (b) sphere with eccentricity.Test time t [min] Coefficient of friction 5 10 15 20 2530 001.0 1.5 2.0 2.5 3.0 0.5 1 2 3 4 5 6 0 7 8 9 1001.0 1.5 2.0 2.5 3.0 0.5 0.250.50.751 1.5 2 0 2.25 2.5 2.753 1.25 1.75 3.2501.0 1.5 2.0 2.5 3.0 0.5 Fig. 9. Horizontal axis: test time; vertical axis: coefcient of friction. Test conditions: N10.5 N and C15% SiC 95% DW. Sliding distances: (a) S110 m, (b) S232 m and(c) S3100 m.R.C. Cozza / Tribology International 70 (2014) 5262 57not alwaysbeincontact withthetribological system(abrasiveparticles and specimen) with each rotation of the ball. Theproblem worsens if the abrasive particles have an average size of12 m, as seen in literature [2].3.2. General analysis of the behavior of the coefcient of frictionInitially, bylaser interferometryanalysis, it wasrelatedthatthere was not the fragmentation of the abrasive particles.Figs. 914present thebehavior ofthe coefcient of friction asafunction of the test time (or, sliding distance S), for the different testconditions established. In thesegures, DW signies distilled water.InFig. 9, withtest conditions N10.5 NC15%SiC95%DW, itisnotedthatbetween2and3 min(Fig. 9aand9c)andfrom8 min (Fig. 9b), there was a turbulence in the contactspecimenabrasiveparticles ball. InFig. 9c, whichshowsthe graph of f(t) for t333 min 20 s (S3100 m), alsoexhibits periodic peaks of that are repeated approximately every2 min 30 s.Fig. 10 displays the results obtained withN21.25 NC25%SiC95%DWinwhichtherearealsoperiodicpeaksof, butwith slightly smaller intervals of approximately 2 min (Fig. 10c).Following this decline, increasing the concentration of the abrasiveslurry to 25% SiC, for both the normal forces of approximately 0.5 Nand1.25 N, thefrequencyof suchturbulenceinthecoefcient offriction decreases to intervals of less than 1 min (Figs. 11 and 12). Therewere practically no peaks in for the condition N10.5 NC225%SiC 75% DW S3100 m (Fig. 11c).Moreover, contrary to this downward trend, the occurrence ofperiodic peaks of began to occur approximately every 2 min 30 swhen the concentration of the abrasive slurry was altered to 37.5%SiC, as seen in Figs. 13 and 14.Test time t [min] Coefficient of friction 51015 20 25300 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 1234 5 6 07 89100 0.2 0.4 0.6 0.8 1.0 0.250.50.7511.5 2 02.25 2.5 2.7531.25 1.75 3.25Fig. 10. Horizontal axis: test time; vertical axis: coefcient of friction. Test conditions: N21.25 N and C15% SiC 95% DW. Sliding distances: (a) S110 m, (b) S232 m and(c) S3100 m.Test time t [min] Coefficient of friction 5 10 15 20 25 30 000.2 0.4 0.6 0.8 1.0 00.2 0.4 0.6 0.8 1.0 1 2 3 4 5 6 0 7 8 9 1000.2 0.4 0.6 0.8 1.0 0.250.50.751 1.5 2 0 2.25 2.5 2.753 1.25 1.75 3.25Fig. 11. Horizontal axis: test time; vertical axis: coefcient of friction. Test conditions: N10.5 N and C225% SiC 75% DW. Sliding distances: (a) S110 m, (b) S232 m and(c) S3100 m.R.C. Cozza / Tribology International 70 (2014) 5262 58Even with thesepeaks,theaverage valueofthecoefcient offriction was between 0.18 and 0.20.The periodic rises in the coefcient of friction can be explained byanalyzing the types of movement that the abrasive particles acquireduring the wear process, which will be discussed in the next section.3.3. Inuence of the abrasive particle motion on the coefcient offriction3.3.1. Contactball abrasive particles andabrasive particlesspecimenManyfactorscontributetowardthecomplexityinanalyzingthecoefcientoffrictionbymicro-abrasiveweartestsusingtherotatingball method. Thesevariables, besidesthemovementofthe abrasive particles, must be considered assuming that there isno direct contact between the ball and the specimen [33] due tothe presence of the three main elements involved in the tribolo-gicalsystem:(i)ball, (ii)abrasiveparticlesand(iii)specimen, asschematized in Fig. 15.Closeanalysisofthissystem duringthetests actstoseparatethe coefcients of friction1: (i) the coefcient of friction betweenthe ball and theabrasive particles (BA) and (ii) the coefcientof frictionbetweentheabrasiveparticlesandthespecimenTest time t [min] Coefficient of friction 0 0.2 0.4 0.6 0.8 1.0 2468 10 12 014 16180 0.2 0.4 0.6 0.8 1.0 1234 5 6 07 89100 0.2 0.4 0.6 0.8 1.0 0.250.50.7511.5 2 02.25 2.52.7531.25 1.75 3.25Fig. 12. Horizontal axis: test time; vertical axis: coefcient of friction. Test conditions: N21.25 N and C225% SiC 75% DW. Sliding distances: (a) S110 m, (b) S232 mand (c) S3100 m.Test time t [min] Coefficient of friction 51015 20 25300 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 12 34 5 6 07 89100 0.2 0.4 0.6 0.8 1.0 0.250.50.7511.5 2 02.25 2.52.7531.25 1.75 3.25Fig. 13. Horizontal axis: test time; vertical axis: coefcient of friction. Test conditions: N10.5 N and C337.5% SiC 62.5% DW. Sliding distances: (a) S110 m, (b) S232 mand (c) S3100 m.1In this analysis, theuid of the abrasive slurry is not being considered.R.C. Cozza / Tribology International 70 (2014) 5262 59(AS), whiledisregardingtheuidof theabrasiveslurry. Conse-quently, there are two tangential forces, (i) TBA and (ii) TAS (Fig. 15),that have impact on BA and AS, respectively.Followingthis procedural analysis, thetangential forces (T)measuredandthecoefcients of friction() calculatedinthiswork are the sums (the result) of TBA and TAS (Eq. (4)) and BAandAS (Eq. (5)), respectively.T TBATAS4 BAAS53.3.2. Dispersion of the coefcient of friction depending on the typeof motion of the abrasive particlesDepending on the type of movement that the abrasive particlesacquire during wear, the quantities TBA, TAS, BAandASareconsiderably inuenced. In abrasive wear tests with a singleabrasiveparticle, Fangetal. [38]reportedthatthedispersionofthe coefcient of friction as a function of test time was relativelygreater for the rolling whencomparedtothe movement ofslide of the abrasive particle. In fact, Fang et al. [38] haveobserved periodic peaks of .Forasingleabrasiveparticle, itispossibletoassumethatthequotient T/Nina specic contact (local charge) is signicantlydependent onthe type of movement of the abrasive particle,rolling or sliding [18], as concluded by Fang et al. [38]. However,due toa greater discrepancyinthecoefcient of frictionas afunction of test time for the rolling abrasion condition, Fang et al.[38] considered the possibility of accepting the hypothesis that, dueto this behavior, thehigherfriction coefcientisrecordedfortherollingabrasivewearcondition. Thisisnot theconsensusdis-cussed within this work; the movement pattern is not exactly thesameforall theabrasiveparticlesparticipatingintheprocessofwear, and uctuations intheoverall valueof thecoefcient offriction can be expected depending on the number of particles thatroll and slide. It is therefore reasonable to assume that the localvalueof T/N equals toglobal valueof T/N incircumstanceswherethemovement patternisthesamefor thevast majorityof theabrasiveparticlesthataresupportingtheloadatagivenmoment.3.4. Inuence of the normal force on the coefcient of friction3.4.1. Literature reviewIn 1999, Dube and Hutchings [39] showed that the coefcientof friction is proportional to normal force. As test specimens, theauthors used carbon-steel AISI 1020 with silica abrasive particlesof sizes ranging between 125 and 150 m.Priorto Nahvi et al. [40],Ramos[41] studiedtheinuenceofthenormalforceonthebehaviorofthecoefcientoffrictionindrysandrubberwheel weartests; inhisM.Sc. Dissertation, heusedspecimensofcarbon-steel AISI1004andAISID2 tool steel,sandANB50andANB100(averageparticlesizeof150 mand300 m, respectively, 99.80%SiO2) andnormal forces of 15 N,30 N, 50 Nand100 N. Ramos [41] notedthat, for all materialcombinationsof specimenandabrasive materials,thecoefcientof friction increased with increase in the normal force.Ramos[41]attributesthisbehaviortoahighdegreeofstrain-hardening of the testedmaterial withthe maximizationof thenormal force, an adverse outcome to the one published by StevensonFig. 14. Horizontal axis: test time; vertical axis: coefcient of friction. Test conditions: N21.25 N and C337.5% SiC 62.5% DW. Sliding distances: (a) S110 m, (b) S232 mand (c) S3100 m.T TASTBANNormal Force Fig. 15. Schematic depictingT, TBA andTAS.R.C. Cozza / Tribology International 70 (2014) 5262 60and Hutchings [42] where the coefcient of friction remainedindependent of the normal force (application range: 24.5122.6 N).3.4.2. Results of this workDubeandHutchings [39], Nahvi et al. [40] andRamos [41]show that the coefcient of friction is proportional to the normalforce, asdescribedintheliteraturereviewabove. However, theresults of this study showed that the average coefcient of frictionremained between 0.18 and 0.20, independent of the normal force,as published by Stevenson and Hutchings [42].The explanation for this inalterability will be developed basedon different magnitudes of the normal forces applied to the weartests, which were discussed above, and the values adopted in thisresearch.In the literature review, which focused mainly on testingconguration of dry sand rubber wheel (DSRW), the normal forcesvalues ranged between 15 N and 100 N, considerably higher thanthe 0.5 N and 1.25 N values established for this study.For abrasive wear, higher normal forces result in deep groovescausedbythe abrasive particles whichhinder the removal ofmaterialandthereforeincreasesthetangentialforce. Takingtheexample from the research conducted by Dube and Hutchings [39]andRamos[41], itispossibletonotethatvariationsinnormalforce between 15 N and 100 N caused relatively minor increase inthecoefcient of friction, nearlya multiple of two, whilethenormal force increased more than six times its original value.Furthermore, another important aspect worth noting is the strain-hardening of the specimen, emphasized by Ramos [41]. Applications ofnormal forces in the range of 15100 Ncause greater degrees of strain-hardening than normal forces below 0.5 N and 1.25 N, as used in thiswork. Therefore, inmicro-abrasivewear testsbytherotatingballmethod, itispossibleandconsistenttoconcludethattheeffectofspecimen strain-hardening on the coefcient of friction is negligible.In particular, in studies of micro-abrasive wear tests by rotatingball where both the values of normal forces and abrasive particlessizes are relatively much smaller, there are lower penetrations andtherefore constant (or approximate) values of coefcient offriction.3.5. Inuence of the abrasive slurry concentration and abrasive wearmodes on the coefcient of frictionThe predominance of grooving abrasion was observed for bothnormal forceswiththeabrasiveslurriesconcentrationof5%SiC(consideredrelatively low) and 25%SiC(consideredrelativelyhigh). It isimportant toemphasizethat highconcentrationsofabrasive slurries favor the action of rolling abrasion, but this is notadenitive rule;in reality, theabrasion modedependson otherfactors, as discussed in detail in the works of Adachi and Hutchings[11,14]. Finally, rollingabrasionwas observedwhentheslurryconcentration was 37.5% SiC.Micro-rolling abrasion, whichis the occurrence of abrasivewear byrolling onthesurfaceor betweenthegrooves, wasobserved. Fig. 16 shows a wear crater generated during theexperiments and an image which relates the occurrence ofmicro-rolling abrasion.The literature did not report on any work that focused on theinuence of the abrasive slurry concentration on the coefcient offriction. In the beginning of this research, it was hypothesized thatincreasingtheabrasiveslurryconcentrationwoulddecreasethecoefcient of frictionbecausethequantityof abrasiveparticleswouldincreaseand, consequently, thewearmodewouldtransi-tion from grooving to rolling, as demonstrated in the literatureby the classic work of Trezona et al. [8]. However, fromtheobtainedresults, it is concludedthat theconcentrationof theabrasive slurry has no inuence on the coefcient of friction.Theconservationofthefrictioncoefcientdespiteincreasingthe abrasive slurry concentration can be attributed to the presenceof distilled water between abrasive SiC particles. Lower concentra-tions of abrasive slurry favor the action of abrasive wear bygrooving, butthereisagreatervolumeofliquid(inthiscase,distilled water) that tends to lubricate, and not wear, the tribolo-gical system. On the other hand, higher concentrations of abrasiveslurry favor the occurrence of abrasive wear by rolling as there islessdistilledwater that canact aslubricant inthetribologicalsystem, leadingto agreatervolume ofabrasiveparticlescominginto contact with the specimen.4. ConclusionsThe results obtained in this work have indicated the following:(1) It is necessary to take caution in comparing the behavior of thecoefcient of friction fromdifferent congurations of testequipment forabrasivewear. Dependingontheequipmentconguration and test conditions, the coefcient of friction canpresent a constant or variable behavior.(2) Thecancellationof theeccentricityof thesphereprovidedreliableandreproducibleresultsbecauseanyinuenceofitsmisalignment on the coefcient of friction was nullied.(3) In this research, the coefcient of friction was independent ofthenormalforceforthematerialsand test conditionsestab-lished. These results are related to the actual magnitude of thenormalforces andaverage size oftheabrasive particles, thatcause relatively small penetrations anda lower degree ofstrain-hardening on the specimen.(4) Boththeconcentrationsof abrasiveslurriesandthesubse-quentactionsof theabrasivewearmodesdidgenerallynotaffect the behavior or magnitude of the coefcient of friction.Fig. 16. (a) Crater obtained in this work and (b) occurrence of micro-rolling abrasion. N21.25 N, S232 m and C225% SiC 75% distilled water.R.C. Cozza / Tribology International 70 (2014) 5262 61However, itis important to emphasizethat the conditionsofthe normal force andconcentrationof the abrasive slurrycausedperiodic peaks onthecoefcient of friction, whichwere related to the types of movements of the abrasiveparticles.AcknowledgmentsThe author gratefully acknowledges Prof. Ana Helena deAlmeida Bressiani, fromthe Nuclear andEnergetic ResearchesInstitute, for thehelpintheabrasiveparticlesizedistributionanalysis; as well as Paulo Zanini, Rafael Rozolen and Vitor BenkardLira, from Rexroth Bosch Group for helping with the start-up of theservo-motors and servo-controllers.Appendix A. Supplementary materialSupplementary data associated with this article can be found intheonlineversionat http://dx.doi.org/10.1016/j.triboint.2013.09.010.References[1] Cozza RC, Tanaka DK, Souza RM. Friction coefcient and abrasive wear modesin ball-cratering tests conducted at constant normal force and constantpressure Preliminary results. Wear 2009;267:6170.[2] da Silva WM, Binder R, de Mello JDB. Abrasive wear of steam-treated sinterediron. Wear 2005;258:16677.[3] CozzaRC, deMelloJDB, TanakaDK, SouzaRM. Relationshipbetweentestseverity and wear mode transition in micro-abrasive wear tests. Wear2007;263:1116.[4] CozzaRC, TanakaDK, SouzaRM. Micro-abrasivewearofDCandpulsedDCtitaniumnitride thin lms withdifferent levels of lmresidual stresses.Surface and Coatings Technology 2006;201:42426.[5] Kelly DA, Hutchings IM. A new method for measurement of particle abrasivity.Wear 2001;250:7680.[6] Rutherford KL, Hutchings IM. Theory and application of a micro-scale abrasivewear test. Journal of Testing and Evaluation JTEVA 1997;25(2):25060.[7] Hutchings IM. Mechanisms of wear in powder technology: a review. PowderTechnology 1993;76:313.[8] TrezonaRI, AllsoppDN, HutchingsIM. Transitionsbetweentwo-bodyandthree-body abrasive wear: inuence of test conditions inthe microscaleabrasive wear test. Wear 1999;225229:20514.[9] CozzaRC. Effect of pressureonabrasivewear modetransitionsinmicro-abrasive wear tests of WC-Co P20. Tribology International 2013;57:26671.[10] BoseK, WoodRJK. Optimuntestsconditions for attaininguniformrollingabrasion in ball cratering tests on hard coatings. Wear 2005;258:32232.[11] Adachi K, Hutchings IM. Sensitivity of wear rates in the micro-scale abrasiontest to test conditions and material hardness. Wear 2005;258:31821.[12] Gee MG, Gant A, Hutchings IM, Bethke R, Schiffman K, Van Acker K, Poulat S,GachonY, vonStebut J. Progresstowardsstandardisationof ball cratering.Wear 2003;255:113.[13] Batista JCA, Godoy C, Matthews A. Micro-scale abrasive wear testing of duplexand non-duplex (single-layered) PVD (Ti,Al)N, TiN and Cr-N coatings. Tribol-ogy International 2002;35:36372.[14] Adachi K, HutchingsIM. Wear-modemappingforthemicro-scaleabrasiontest. Wear 2003;255:239.[15] Bello JO, Wood RJK. Micro-abrasion of lled and unlled polyamide 11coatings. Wear 2005;258:294302.[16] Kusano Y, Van Acker K, Hutchings IM. Methods of data analysis for the micro-scaleabrasiontest oncoatedsubstrates. SurfaceandCoatings Technology2004;183:31227.[17] Allsopp DN, Hutchings IM. Micro-scale abrasion and scratch response of PVDcoatings at elevated temperatures. Wear 2001;251:130814.[18] Shipway PH. Amechanical model for particle motioninthe micro-scaleabrasion wear test. Wear 2004;257:98491.[19] Sinnett-Jones PE, Wharton JA, Wood RJK. Micro-abrasion-corrosion of a CoCrMoalloy in simulated articial hip joint environments. Wear 2005;259:898904.[20] Trezona RI, Hutchings IM. Three-body abrasive wear testing of soft materials.Wear 1999;233235:20921.[21] Allsopp DN, Trezona RI, Hutchings IM. The effects of ball surface condition inthe micro-scale abrasive wear test. Tribology Letters 1998;5:25964.[22] Stachowiak GW. Particle angularity and its relationship to abrasive and erosivewear. Wear 2000;241:2149.[23] Stachowiak GB, Stachowiak GW. Wear mechanisms in ball-cratering tests withlarge abrasive particles. Wear 2004;256:6007.[24] Neville A, Kollia-Rafailidi V. A comparison of boundary wearlm formation onsteel and a thermal sprayed Co/Cr/Mo coating under sliding conditions. Wear2002;252:22739.[25] Wang L, Wood RJK, Harvey TJ, Morris S, Powrie HEG, Care I. Wear performanceof oil lubricatedsiliconnitrideslidingagainstvariousbearingsteels. Wear2003;255:65768.[26] Dhanasekaran S, Gnanamoorthy R. Abrasive wear behavior of sintered steelsprepared with MoS2 addition. Wear 2007;262:61723.[27] Liu CS, Zheng ZY, Wu DW, Ye MS, Gao P, Peng YG, et al. Sliding friction and wearproperties of CNX/TiN compositelms. Tribology International 2004;37:7215.[28] Iliu I. Wear and micropitting of steel ball sliding against TiN coated steel platein dry and lubricated conditions. Tribology International 2006;39:60715.[29] Ceschini L, Palombarini G, Sambogna G, FirraoD, ScavinoG, Ubertalli G.Friction andwear behaviour of sintered steels submitted to sliding andabrasion tests. Tribology International 2006;39:74855.[30] Yan Y, Neville A, Dowson D. Tribo-corrosion properties of cobalt-basedmedical implant alloys in simulated biological environments. Wear2007;263:110511.[31] Jianxin D, Jianhua L, Jinlong Z, Wenlong S, Ming N. Friction and wear behaviorsof the PVDZrNcoated carbide in sliding wear tests and in machiningprocesses. Wear 2008;264:298307.[32] RutherfordKL, Hutchings IM. Amicro-abrasivewear test, withparticularapplication to coated systems. Surface and Coatings Technology1996;79:2319.[33] KusanoY, HutchingsIM. Sourcesof variabilityinthefree-ball micro-scaleabrasion test. Wear 2005;258:3137.[34] Cozza RC. Astudy onfrictioncoefcient andwear coefcient of coatedsystems submitted to micro-scale abrasion tests, Surface and CoatingsTechnology. http://dx.doi.org/10.1016/j.surfcoat.2012.06.88.[35] Gee MG, Gant AJ, Hutchings IM, Kusano Y, Schiffman K, Van Acker K, Poulat S,GachonY, vonStebut J, HattoP, Plint G. Results fromaninterlaboratoryexercise to validate the micro-scale abrasion test. Wear 2005;259:2735.[36] Standard DIN 804 Edition: March/1977.[37] Gee MG, Wicks MJ. Ball crater testing for the measurement of the unlubricatedsliding wear of wear-resistant coatings. Surface andCoatings Technology2000;133134:37682.[38] Fang L, Kong XL, Su JY, Zhou QD. Movement patterns of abrasive particles inthree-body abrasion. Wear 1993;162164:7829.[39] DubeNB,HutchingsIM.Inuence of particlefracture in thehigh-stress andlow-stress abrasive wear of steel. Wear 1999;233235:24656.[40] Nahvi SM, Shipway PH, McCartney DG. Particle motion and modes of wear inthe dry sand-rubber wheel abrasion test. Wear 2009;267:208391.[41] Ramos LV. Construction and instrumentation of a dry sand rubber wheel weartest equipmentotostudythebehaviour of steels[M.Sc. dissertation]. SoPaulo SP, Brazil: Polytechnic School of the University of So Paulo; 2005: 82p. (in Portuguese).[42] StevensonANJ, HutchingsIM. Development of thedrysand/rubber wheelabrasion test. Wear 1996;195:23240.R.C. Cozza / Tribology International 70 (2014) 5262 62