TheinuenceofabrasivebodydimensionsonsingleasperitywearM.Woldmana,c,n,E.VanDerHeidea,b,T.Tingac,M.A.MasenaaUniversityofTwente,P.O.
Box217,7500AE,Enschede,TheNetherlandsbTNO,P.O. Box6235,
5600HE,Eindhoven,
TheNetherlandscNetherlandsDefenceAcademy,P.O.Box10000,1780CA,DenHelder,TheNetherlandsarticle
infoArticlehistory:Received14September2012Receivedinrevisedform4December2012Accepted10December2012Availableonline21December2012Keywords:AbrasivewearWearrateSingleasperityscratchtestsabstractThiswork
focuses on the relationbetween the
propertiesofabrasivebodiesandtheweartheycause.By performing single
asperity scratch tests to simulate abrasive wear, the wear process
takes place in acontrolledenvironment, allowing the properties of
the abrasive body to be isolatedandstudiedindependently.
Thesetupusedisapin-on-atmachine,
enablingscratchingalongastraightlineandunderactiveloadcontrol.
Theabrasivebodiesweregeneralisedbyusingsinglecrystal
SiO2tipstoscratch DIN St-52 steel samples. The tipradiuswas
variedtostudythe inuenceof the abrasive
bodysizeontheabrasivewearprocess. Thenormalloadwasvariedaswell
toobtainitsinuence. Usingconfocal microscopy, the scratches were
analysed to identify the abrasion mechanisms, such
asploughingandcuttingandtodeterminethevolumetricwear.As expected,
the results show an expected increase in the wear volume with
increasing load. Moreinterestingly,
thewearratevariessignicantlyasafunctionofthesizeoftheabrasive,
enablingthepredictionof thewearratebasedontheabrasivebodysize.
Itisobservedthat, contrarytoearlierreported size effects, smaller
tips cause more wear than larger tips do. Moreover, at the low
values forthe degree of penetration studied in this work, a
regimeof limited plastic deformation was
identied,basedontheobservationthat thedegreeof wear parameter
decreases
withincreasingdegreeofpenetration.&2012ElsevierB.V.Allrightsreserved.1.
IntroductionExcessive wear can occur in machines, because of the
abrasiveaction
offorinstancesandparticlesinthecontactbetweensteelmachinecomponents.
Toprevent suchequipment fromfailing,detailed knowledge on the
abrasive wear process is required. Thiswork focuses on the relation
between the geometrical propertiesofabrasivebodiesandthewearthey
cause.In two- and three-body abrasive wear processes the
propertiesand characteristics of the abrasive bodies are an
important aspect.These abrasive bodies can be either hard particles
such as sand
orweardebristhatispresentinthesystem,orhardprotuberancesononeofthe
surfaces,forinstance dueto embeddingintoa softsurface of before
mentioned particles or surface roughnessasperities.
Propertiessuchasthesizeandthegeometryoftheseabrasive bodies have an
inuence on the predominant wearmechanisms,
meaningthatthereisnotnecessarilyastraightfor-wardrelationbetweenthesizeandshapeof
theabrasivebodyand the resulting wear volume. Examples are
described by [15],who state that in two-body abrasion, abrasive
particles orprotuberanceswithasizebelowapproximately100
mmshowaproportionate relation between size and wear caused, whereas
forlargerparticlesthewearcausedisindependentofsize. Explana-tions
of this size effect range from the scale dependent bluntnessof
particles and the benets of work hardening to the wearresistanceof
thetopmost nano- or micro-metres of asurface.Even though the
explanation for the observed size effect inabrasionis lacking, it
canbe concluded that the geometricalaspectsof
theabrasiveparticlesplayanimportant role. Inthepresent workthebasic
mechanisms behindabrasivewear
arestudiedbyfocusingontheabrasivewearof DINSt-52samplescaused by
single protuberances made of quartz (SiO2) as afunction oftheirsize
andnormal load.Inliterature,
severalmodelshavebeendescribedtocalculateabrasivewear [611].
However, improvements are neededforabrasion to be quantied to a
reasonable extent, e.g. by
incorpor-atingtheshapeandmotionoftheabrasivebodyintothemodel.Besides
the abrasive body characteristics, external conditions likethe
normal load inuence the abrasive wear rate too. According
toArchards wear equation [12], the abrasive wear volume
increaseslinearly withincreasingnormal load.Theaimof
theworkpresentedinthispaperistostudytheinuenceof
theabrasivemediumsizeonthewear process,
inContentslistsavailableatSciVerseScienceDirectjournal homepage:
www.elsevier.com/locate/wearWear0043-1648/$ -
seefrontmatter&2012
ElsevierB.V.Allrightsreserved.http://dx.doi.org/10.1016/j.wear.2012.12.009nCorresponding
author at: University of Twente, Drienerlolaan 5, P.O. Box
217,7500AE Enschede,TheNetherlands.Tel.:
31534892463.E-mailaddress:[email protected](M.Woldman).Wear301
(2013)7681particular when the abrasive medium is sand. To clearly
separatethe size effect, this property needs to be
isolatedfromotherpropertieslikeshape, hardnessandnumberofparticles.
Experi-mentalmethodsthatareoftenappliedtostudytheparticlesizeeffect
are using a multiple-particle setup, e.g. a pin-on-disk
setupwithadiskofabrasivepaper.Thisenhancesacouplingbetweenparticlesizeandthenumberofparticlesonthecontactandtheapplied
normal load, making it difcult to capture the actual sizeeffect.
Further, because of the multiple pass situationinsuchtests, the
scratches will overlap, thereby inuencing the wear rate[9,13].
Hence, theoriginalcontributionoftheworkpresentedinthis paper is
that the particle size effect is studied by performingsinglepass,
singleasperityscratchtestsemployingapin-on-atsetup. Usingvarious
tipradii theinuenceof thesizeontheabrasive wear can be studied
directly. Moreover, rather thanquantifying
theabrasivewearresistanceofthecounter material,this paper focuses
ontheproperties of theabrasiveitself; byutilising quartz tips the
test conditions will resemble a
wearprocesswheresandparticlesaretheabrasivemedium.Using this method
for studying the particle size effect
isconsideredanessentialstartingpoint,
todescribethescratchingbehaviour of particles in a situationof
sliding contact. Later,multiple particle and multiple pass
situations will be included
toapproachtherealabrasivewearsituationasmuchaspossible.2. Single
asperity scratch testsOnecommonway ofstudyingabrasioninacontrolled
wayisby sliding a pin or prismatic specimen over a surface (e.g.
[6,14]).Sliding the pin over a softer surface at a constant normal
load andsliding speedenablesthestudy ofthewear
processbyexamina-tion of the wear scar. The setup used in this
research is of the pin-on-at type. A photograph of the setup is
shown in Fig. 1. The tipismountedinasampleholder,
suspendedinelastichingesandconnected to a z-stage to make sure the
pin moves in the vertical/z-directiononly. Thesampletobescratchedis
mountedonalinear xy-stage. Theloads anddisplacements
arecontrolledbypiezo-electric actuators, force transducers and a
PID(Propor-tionalIntegralDerivative)controlsystem.At the onset of a
scratch test, the tip is lowered slowly towardsthe surface until
contact is achieved. Then, the predened normalload is applied and
the x-stage is moved to provide the
scratchingmotion.ThePIDsystemcontrolstheappliednormalloadwithin70.05
N.They-stageisusedtoperformconsecutivescratches.Sandisatypical
abrasivemediumthat causesconsiderablewearissues in many practical
applications. Current researchintoabrasivewear, however,
mainlyfocusesoneither steel or dia-mondabrasivebodiesandhence,
theknowledgeaboutabrasionbysandislacking. Theuseof
sandtostudyabrasionexperimentallycauses some issues though.
Firstly, sand particles have randomshapes, while the shape of the
abrasive should be known andunambiguously described to investigate
the basic geometrical effectson wear effectively. Secondly, it is
difcult to properly mountindividual sandparticles. Hence, inthis
workscratches aremadeusing tips produced of mono-crystalline quartz
(SiO2). The quartz tipshave a cone angle of 601 and tip radii of
50, 100 and 200 mm. Fig. 2shows photographs of the tips with the
three sizes used; thephotographs are obtained with a stereo
microscope.As the counter material, standard DIN St-52 steel is
selected asit is commonly used in machine components. The steel
specimencontactsurfacesarepolishedtoaroughnessvalueofRa6
mm.ThemechanicalpropertiesofboththequartzandthesteelusedarereportedinTable1.Experimentsareperformedusingthethreedifferenttipsizeswithnormalloadsbetween0.1and11
Nandaconstantslidingspeedof0.5 mm/s. Byvisual
inspectioninbetweentestsitwasveriedthat thegeometryof thetipsdidnot
changeprogres-sively. Moreover, the tips were cleanedusing a ne
brushinbetweenteststoremovepossiblesteelweardebris.
Thecreatedscratchesareanalysedusingaconfocal microscope,
measuringthegroovetodeterminethemechanism(scratching,
ploughing)andtheamountof wear. Aschematiccross-sectionof
atypicalwear scar is shown in Fig. 3, with Ag the groove area, As
the area ofthe shoulders, w the groove-width and d the depth of
indentation.MultiplicationofAgandAswiththescratchlengthlgivesthegroove
volume Vg and shoulder volume Vs. Then, the total
amountNomenclatureAggroovearea[mm2]Asshoulderarea[mm2]Dpdegreeofpenetration
[dimensionless]Dp,kdegreeofpenetration,according
toKato[dimensionless]Dp,mdegreeofpenetration,measured[dimensionless]d
depthofindentation [mm]HvHardness,Vickers[N/mm2]k
specicwearrate[mm3/N m]l scratchlength[mm]N normalload[N]R
tipradius[mm]R2coefcientofdetermination[dimensionless]Rasurfaceroughness[mm]Rggrooveradius[mm]s
slidingdistance[m]Vggroovevolume[mm3]Vsshouldervolume[mm3]w
groove-width[mm]b degreeofwear[dimensionless]y angle describing the
geometry of the abrasive tip[dimensionless]Fig.1.
Photographoftheexperimental setup.M.Woldmanetal./Wear301(2013)7681
77ofremovedmaterialisequaltoVgVsandthedegreeofwearisdetermined
usingthefollowing equation[15]:b VgVsVg1with
b1meaningidealmaterialremoval(purecuttingandnoploughing)
andb0indicatingideal ploughingor nomaterialremoval. Using this
parameter, the specic wear ratek (mm3/N m)derived by Archard [12]
can be expressed ask bVgNs2with N the normal load (N)and s the
sliding distance (m). Finally,
therelativeindentationdepthcanbecharacterisedwiththedegreeofpenetrationDp(dimensionless),
whichisequal totheratiooftheindentation depth and the half-width of
the groove. When obtainedfrom measured data, the degree of
penetration (Dp,m) is equal toDp,mdw=2 2dw3Fromthisdenition,
HokkirigawaandKato[14]derivedatheore-tical equation for the degree
of penetration, based on the geometryof the scratching
tipDp,tRpHv2NrpHv2NR21r4whereR is the tip radius (mm) and Hv the
hardness of the wearingmaterial (N/mm2). Note that higher values of
Dp are linked to cuttingwear and lower values indicate ploughing
wear.3. Resultsand discussion3.1.
ParticlesizeeffectThecorrelationbetweentheexperimentallydeterminedandtheoretical
degree of penetration is veried in Fig. 4, plotting
thedegreeofpenetrationpredictedbyEq.
(4)againstthemeasureddegreeofpenetrationaccordingtoEq. (3). Fig.
4showsthatthemeasured values are lower than the predicted ones.
Although thedata points couldbe representedby a straight line, its
slopeclearly deviates from the value of 1 represented by the dotted
linein Fig. 4. The theoretical parameter represents the degree
ofpenetration when the indenter is in the material, while
theexperimentalparameterisbasedonthegrooveafterremovaloftheindenter.
Astheelasticdeformationofthecountermaterialdisappears upon load
removal, that part of the indentation depthcannot be retrieved
fromthe groove [16]. As a result, themeasureddegree of
penetrationis lower thanthe theoreticalvalue that includes the
elastic part. Since the elastic deformationFig. 2. Examples of
quartz tips used, (a) R50 mm, (b) R100 mm and (c) R200
mm.Table1Mechanicalpropertiesofthe materialstested.Material
Hardness(GPa)Youngsmodulus(GPa)Density(kg/mm3)Poissonsratio(dimensionless)Quartz(SiO2)
9.8 7274 2200 0.17Steel(DINSt-52) 0.2 210 7800 0.30Fig. 3.
Schematicillustrationofthe cross-sectionofawearscar.Fig.4.
MeasureddegreeofpenetrationagainsttheoreticaldegreeofpenetrationasdenedbyHokkirigawaandKato[14].M.Woldmanetal./Wear301(2013)7681
78has a linear relation with the applied load, the relative
differencebetweenthetwoparametersisconstant andeitherof
thetwoparameterscanbeused.Thescratchresultsinthispaperwillbecompared
to Dp,t rather than Dp,m. Note that to give an indicationof
theaccuracyof theexperiment, repeatedscratchtestswereperformed
under identical conditions (tip radius R100 mm).Moreover,
theaccuracyofthewearvolumemeasurementswiththeconfocal
microscopewasestablishedbyrepeatingconfocalmeasurements on 1 wear
track. This is why Fig. 4 shows multipleresults at a value of Dpof
around 0.14 (repeated
confocalmeasurements)and0.21(repeatedscratchtests),
correspondingwithnormalloadsof1and5 N.
Thevariationisexpectedtobecaused by a transitional region for the
active wear
regime(alternatingbetweenploughingandscratching)athigherloads.ReferringtoFig.
4, thiswouldexplaintherelativehighvariationatavalueof Dpof
around0.21, indicatingthatcertainlyinthehigher load regime location
for the confocal measurement
isimportant.Notethatinthefollowinggures, theaveragevaluesforthetwo
abovementionedconditionswillbeused.When the measured wear volume is
plotted against thecorrespondingnormal load, Fig. 5(a)isobtained.
Inconsecutiveexperiments, thenormal loadwasincreaseduntil
thetipfailedduetobrittlefracture.
Thisexplainswhythemaximumappliedloadsaresmallerforthesmallertipradii,
becausetheyfailatalower load.Assuming linear behaviour according to
Eq. (2) (i.e. taking thatk, b ands are constant), the data points
for the different tip
radiiarettedwithstraightlinesandtheslopesofthelinesthenareequaltothespecicwearratevalueskreportedinTable2.
Thedifferences in the wear rate values indicate a size effect; the
wearrate of the steel caused by tips with a small radius is about
twicethewearrateforthelargertips.Tostudythedevelopmentof
thedifferentwearmodesovertheloadregime, thetotal wear volumeis
plottedagainst thedegreeofpenetrationinFig. 5(b).
Alsoheredifferenttrendsareobservedfor the different tipradii. For
all tipradii the wearvolume increases with increasing degree of
penetration, followingparabolictrends.
Thewearvolumeincreaseswithincreasingtipsize, because for constant
Dpthe cross-sectional area of thegrooveislargerforlargertips.Fig.
5clearly illustrate the effects of size onthe wear
volume.Theslopesforthettedcurvesforthetipswithradius100and200 mmare
comparable, whereas the line for the tip withR50 mmisdifferent.
Fromliterature, indeedasizeeffect hasbeen described for abrasive
bodies with sizes up to around100 mm. Similarly, the wear rate is
expectedtoincrease withparticle size. According to Fig. 5(a)
however, the smaller tipresults inahigher wear ratethanlarger tips
do. This canbeexplained with a geometrical analysis. The
cross-sectional area AgofthegroovecanbeexpressedasAgR2g2ysin y
5with Rg the groove radius, which is approximately equal to the
tipradius and y the angle between the left- and right contact point
asshowninFig.3.Angle yisequaltoy 2cos1RdR 6withEq. (5), thewear
volumecanbecalculatedtheoretically,dependingonthetipradiusRandindentation
depthd.Inplasticallydeformingcontacts,
thecontactpressureequalsthehardnessofthedeformingmaterialandthusthesizeofthecontactareacanbeexpressedastheratiooftheappliednormalload
and the hardness. This means that the contact area isindependent of
thetipradius, andtheresultingwear volumecan be plotted against the
contact area for the different tip radii asinFig. 6.
Thisgureshowsthatforaconstantcontactarea, andthus a constant load,
the volume of the resulting groove increaseswith decreasing tip
radius. Therefore, the wear rate increases withdecreasingtipradius
ataconstant load.Note that this effect opposes the size effect
reported inliterature, where it is observed that the abrasive wear
rateincreaseswithincreasingsize.
Apossibleexplanationisthatinliterature, see e.g. [5], the particle
size effect is commonly
studiedusingagridofparticlesortips,whileinthispapertheeffectsofonlyonetiparestudied.Whenusingagridandaconstanttotalnormal
load, theloadpertipdecreaseswithdecreasingtipsizeFig. 5.
Representativewearvolumeversusthenormal load(a)
andthedegreeofpenetration(b).Table
2Specicwearratesforthedifferenttipradii.Tipradius(mm)
Wearrate(104mm3/N m) R250 4.3 0.87100 1.8 0.92200 2.2
0.94M.Woldmanetal./Wear301(2013)7681
79sincethenumberoftipsperunitareaincreasesforsmallertipswhen the
height distribution of the bearing area curve of the gridis kept
constant. This results ina decreasedamount of
wearcausedbyeachsingletip.Itisexpectedthattheresultsofthepresentwork,
combinedwithearlierwork[17], canbeusedthoroughlytoexaminethereal
three-bodywear process andpredict theassociatedwearrates.3.2.
WearmechanismsAccording to literature, plotting the degree of
penetration(Eq. (4))againstthedegreeofwear(Eq.
(1))shouldresultinans-shaped curve, as shown in Fig. 7(a) based on
the work ofHokkirigawaandLi [18]. Thecurveshowsadegreeof
wearof0atlowdegreesofpenetration,
andanincreasewithincreasingDpfollowingans-shapedcurve. Typically,
resultsat degreesofpenetration below0.1 are not reported, possibly
because
oflimitationsoftheaccuracyofthemeasurementequipment;thetypical
lateral dimensions of the wear scratch at such low valuesfor Dp are
in the order of microns and the depth is in the order ofnanometres.
Byemployingaconfocal microscopewithaheightresolution of less than 1
nm it is possible to measure the
resultingwearscratchesobtainedatlowdegreesofpenetration. Fig.
7(b)shows degrees of wear obtained for low values of Dp. At very
lowvalues of the degree of penetration, a value of almost 1
isobservedforthedegreeofwearandwithincreasingindentationdepths the
degree of wear drops towards 0. The gure is anexpansion of the
existing curves, showing that the degree of wearrst decreases with
increasing degree of penetration. This isbecause of plastic
deformationof the counter surface. At lowloads,
andthuslowdegreesofpenetration, thecountermaterialtends to
deformplastically, pushing the deformed
materialdownwardsratherthanintotheshoulders.
ThismeansthatthegroovevolumeVghasarelativelyhighvaluewhilenoshouldersareformedandthustheshouldervolumeVsispracticallyzero,leading
to a high value for the degree of wear parameter b. As thedegreeof
penetrationincreases, thematerial will nolonger bepushed down but
will go upwards and sideways,
causingshoulderstoformalongthescratchanddecreasingthedegreeofwearvalue.The
load range that could be used in the current set
ofexperimentsislimitedbecauseof brittlefractureof thequartztip
material. This means that the degrees of penetration at whichthe
wear has been studied are limited too. Note that the fractureof
abrasive sand particles at high indentation depths andsignicant
shearloadsalsooccur inpracticalapplications.4.
ConclusionsTheinuenceof abrasivebodysizeontheabrasivewearinsingle
asperity scratch tests has been studied. The work presentedin this
paper demonstrates the benets of performing single pass,single
asperity scratch tests in helping to understand the
particlesizeeffect.Contrarytothecommonlyappliedweartestsusingamulti-particle
setup (e.g. by running a tip against a grid ofparticles),
performingsingleasperityscratchtestswithdifferenttipradii
enablestoisolateandquantifyatipsizeeffect. It hasbeen found that
both the amount of wear and the wear
mechan-ismvarywithsizeatagivenload.Smallertipscausemorewearthanlargertipsdo,
becauseatconstantload, thegroovevolumeincreases
withdecreasingtipradius. This observedsize
effectopposesthesizeeffectsreportedbyothers.Atverylowvaluesforthedegreeof
penetrationithasbeenobserved that plastic deformation causes high
values of theFig. 6. Contactareaversusgroovearea,calculated.Fig. 7.
Degreeof wearversusdegreeof
penetrationadaptedfromHokkirigawaandLi[18](a)andaccordingtothemeasurements(b).M.Woldmanetal./Wear301(2013)7681
80degreeof wear parameter. At theselowloads
andindentationdepthsthesurfacedamagemechanismisnot wear, but
plasticdeformation,withoutshoulderformation.AcknowledgementsThe
authors wish to thank Walter Lette and Erik de Vries fromthe
supporting staff of the Laboratory for Surface Technology
andTribologyattheUniversityofTwentefortheirhelprunningtheexperiments.References[1]
A. Misra, I. Finnie, Onthesizeeffectinabrasiveanderosivewear,
Wear65(1980)359373.[2] I.M. Hutchings, Tribology: Friction and Wear
of Engineering Materials,EdwardArnold,London,1992.[3] D.V. de
Pellegrin, A.A. Torrance, E. Haran, Wear mechanisms and scale
effectsintwo bodyabrasion,Wear266(12)(2009)1320.[4]
H.Sin,N.Saka,N.P.Suh,Abrasivewearmechanismsandthegritsizeeffect,Wear55(1)(1979)163190.[5]
R. G ahlin, S. Jacobson, The particle size effect in abrasion
studied bycontrolledabrasivesurfaces,Wear258(14)(1999)118125.[6]
M.A. Masen, M.B. deRooij, D.J. Schipper,
Micro-contactbasedmodellingofabrasivewear,Wear258(14)(2005)339348.[7]
K.H. ZumGahr, Modellingoftwo-bodyabrasivewear,
Wear124(1)(1988)87103.[8]
A.A.Torrance,Modellingabrasivewear,Wear258(14)(2005)281293.[9]
Y.Xie,J.A.Williams, Thepredictionoffriction
andwearwhenasoftsurfaceslidesagainstaharderroughsurface,Wear196(12)(1996)2134.[10]
L. Fang, W. Liu, D. Du, X. Zhang, Q. Xue, Predicting three-body
abrasive wearusingMonteCarlomethods,Wear256(78)(2004)685694.[11] Y.
Xie, B. Bhushan, Effects of particle size, polishing pad and
contact pressureinfreeabrasivepolishing,Wear200(1996)281295.[12]
J.F.Archard, Contactandrubbingofatsurfaces,
JournalofAppliedPhysics24(8)(1953)981988.[13] A.A. Torrance, A new
approach to the mechanics of abrasion, Wear 67 (1981)233257.[14] K.
Hokkirigawa, K. Kato, Anexperimental andtheoretical
investigationofploughing, cuttingandwedgeformation,
TribologyInternational 21(1988)5157.[15] K. Kato, Wearmechanisms,
NewDir. Tribol., publishedat theFirst
WorldTribologyCongress,1997pp.3956.[16] A.A. Torrance, The
correlation of abrasive wear tests, Wear 63 (1980)359370.[17] M.
Woldman, E. VanDerHeide, D.J. Schipper, T. Tinga, M.A. Masen,
Inves-tigating the inuence of sand particle properties on abrasive
wear behaviour,Wear294-295(2012)419426.[18] K. Hokkirigawa, Z.Z.
Li,
Theeffectofhardnessonthetransitionofabrasivewearmechanismsofsteels,
in:ProceedingsoftheInternationalConferenceonWearofMaterials1987585593.M.Woldmanetal./Wear301(2013)7681
81
