Measurement of sound, vibration and friction between soft materials under light loads

(1)

Wear276–277 (2012) 61–69

ContentslistsavailableatSciVerseScienceDirect

Wear

j o ur n a l ho me p a g e :w w w . e l s e v i e r . c o m / l o c a t e / w e a r

Measurement

of

sound,

vibration

and

friction

between

soft

materials

under

light

loads

夽

Adnan

Akay

a,∗

_,

_Blair

_Echols

b

_,

_Junqi

_Ding

c

_,

_Anne

_Dussaud

c

_,

_Alex

_Lips

c

a_Bilkent_University,_Mechanical_Engineering_Department,₀₆₈₀₀_Bilkent,_Ankara,_Turkey b_Carnegie_Mellon_University,_Mechanical_Engineering_Department,_Pittsburgh,_PA_15213,_USA c_Unilever_Research_&_Development,₄₀_Merritt_Blvd._Trumbull,_CT_06611,_USA

a

r

t

i

c

l

e

i

n

f

o

Articlehistory: Received25May2011

Receivedinrevisedform5November2011 Accepted14December2011

Available online 22 December 2011 Keywords:

Tactile Softmaterials Frictionmeasurement

a

b

s

t

r

a

c

t

Tactileperceptionofmaterialsandsurfacetextureinvolvesfrictionunderlightnormalloadsandis fundamentaltofurtheradvancingareassuchastactilesensing,hapticsystemsusedinroboticgripping ofsensitiveobjects,andcharacterizationofproductsthatrangefromfabricstopersonalcareproducts, suchaslotions,onskin.Thispaperdescribesanewapparatustomeasurefrictionsimultaneouslywith dynamicquantitiessuchasaccelerations,forces,andsoundpressuresresultingfromlightslidingcontact overasoftmaterial,muchlikeaﬁngerlightlytouchingasoftmaterial.Thepaperalsointroducesanovel frictionandadhesionmeasurementmethodthatcanbeparticularlyusefulforsoftmaterialsandlight normalloads.

1. Introduction

Inrecentyears,hapticperception,oneofthefundamentalareas of cognitive engineering, has advanced withsignificant contri-butions from researchin biology, engineering, and psychology. Althoughlesswellunderstoodthanauditionandvision,thetouch, orsomaticsensation,continuestohaveanincreasingsignificance inengineeringandmedicalfieldsalike.Forexample,inthe devel-opmentofroboticgripsorvirtualrealityhardware,tactilefeedback isanessentialelement.Inmedicine,tactilesensitivityatfingertips isthoughttohavea roleindiagnosisofand monitoringduring rehabilitationfordiabetes.

Haptic perceptionofmaterialsand surfacesrelieslargely on touch,ortactilesensing,butalsobenefitsfromvisualandauditory cuesand,insomecases,fromolfactoryandevengustatorysenses. Investigationshaveshownthatvisualorauditorycuescanhavea significantroleindiscriminatingsurfacecharacteristics[viz.,1–4]. Studiesontheroleofmultisensoryintegrationofthesesignalson perceptionofsurfacefinishortexturesuggestthatvisionandtouch

夽 ThisresearchwascarriedoutatCarnegieMellonUniversity.

∗ Correspondingauthorat:BilkentUniversity,MechanicalEngineering Depart-ment,06800Bilkent,Ankara,Turkey.Tel.:+903122903389;

fax:+903122664126.

E-mailaddresses:akay@bilkent.edu.tr,akay@cmu.edu(A.Akay).

actasindependentsources[4]althoughinsomecasesvisioncan improveperceptionbasedonsomaticsensation[5].

Multiplesensoryinputsareutilizednotonlyinjudgingor char-acterizingsurfacefinishoritstexture,asforfabricsandcosmetics, butalsoinidentifyingshapeandweightofanobject.Inapplications ofhaptictechnologiestograspandliftortomoveverylight-weight orfragileobjectsmechanically,tactilesensinghasaprimaryrole in theregulation oftheappliedpower.Inallsuchapplications, thereisaneedtorelatethephysicalcharacteristicsofsurface qual-itiesandtexturetosomaticsensationandperception.However, thequestionremainsopenwhetherornotthesubjective evalua-tionssuchasrough/smooth,sticky/slippery,andhard/softcanbe identifiedreliablyusingobjectivemeasures.Theextremesurface characteristics,ofcourse,canbeidentifiedusinganumberof differ-entmethods.Forsurfaceswithmixedorlesspronouncedphysical characteristics,however,objectivemeasureshavenotyetreplaced humanperceptions,andtherecontinuestobeaneedtodevelop measurementtechniquestoemulatesomaticsensationthatalso correlatewellwithperception.

Thepurposeofthispaperistodescribeamethodtomeasure sig-nalssimilartothetactilesignalsthataﬁngersenseswhenitlightly touchesasoftsurfacewhilesimultaneouslycapturingtheair-borne acousticsignalsgeneratedduetothefrictionforcesthatdevelop betweenthem.Themethodalsoincludesa uniqueapproach to measuringfrictionandadhesionbetweentwo surfacesthatcan pavethewaytodiscernbetweeneachpairofthesubjective quali-tiesdescribedabove.

(2)

Fig.1.Schematicdescriptionofapparatustomeasuretribologicalandacousticsignalsfromfrictionbetweensoftmaterialswiththecrosssectionalviewofcontactbetween ﬁngerprobeandarmduringcontactindicatingthepositionsofaccelerometersinsidetheﬁngerprobeandforcetransducersinthearm.

2. Contactparametersbetweensoftsurfacesandtheir measurement

Themechanicalstimulithatformthebasisofsomaticsensory perceptionoriginatefromcontactforcesthatdevelopduringthe rubbingprocess,whichcontainsinformation aboutsmoothness, roughness,stickiness,slipperiness,hardness,andsoftnessofthe surfaces.

Atroomtemperature,thephysicalparametersthatareclosely related to and influence subjective measures are the material properties,surfacetextureandshape,andadhesionbetweenthe surfaces.Theslidingspeedandcontactpressuredefinethedynamic effects of these parameters. As a by-product of light rubbing, structure-andair-bornesoundsprovideadditionalsignalsthatcan beusedinevaluatingthesurfacecharacteristicsandcorrelating themeasurementswithperception[6].Effectsofsomeofthese materialandtexturepropertiescanalsobemeasuredbymounting accelerometersonafingertodetectsurfaceroughnessand slipper-iness.

Similarly, accelerometers placed on a probe can provide information regarding surface roughness/smoothness and even hardness/softnessof a surface throughthe transientoscillatory responseof theprobetipwhen theyﬁrstcomeintocontact as discussedbelow.Forexample,surfacesthat areoftendescribed asstickycorrelatewiththetangentialforce[viz.,7–9],whichon anotherwisesmoothsurfacestemsfromadhesiveforcesthatare knownto be dissipative and, consequently,those described as smoothandsilkyproducealowtangentialdissipativeforce.The magnitudesandtime(orfrequency)dependenceofthesesignals inﬂuencetheperceptionaboutasurface.Inaddition,thirdbodies suchaslubricantsbetweenthesurfacesalsoaffectthemagnitude ofthetangentialdissipativeforce[10].

Theapparatusdescribedinthispapercanmeasuretribological anddynamicalparametersthatresultfromfrictionbetweensoft materialsunderverylightloads,suchthattheycanbeusedtorelate dynamicaloutputtomaterialandtextureproperties.Thesimple apparatusdescribedheremimicsverylightpassivetouchofa ﬁn-geroveranothersurface.Thedevicehasthecapabilitytomeasure air-andstructure-borneacousticsignalsaswellasthetangential contactforcesandnormalloadsduringthecontactprocess.

3. Testapparatus

3.1. Designcriteria

Thefunctionalrequirementsonwhichthedesigncriteriaare basedinclude:(i)aﬁnger-likemechanismthatcanlightlytoucha movingsurface,(ii)signalsgeneratedduringitsoperationlargely resultfromthecontactofthe“ﬁngerprobe”withasample,with negligible contributionsof ancillary signals from othersources, and(iii) theapparatusallowsaccess tosensorsandaccessories formeasurementsduringitsoperation.Inadditiontothese func-tionalrequirements,furtherdesignrulesweredevelopedforease ofoperationandmeasurementswiththeapparatus.

Likeafinger,theprobeshouldbeabletoretractfromorpress onamovingsurface,withadesirednormalcontactloadsimilar toalight touchthatanactualfingercanexert.Asafingerdoes whileprobingasurface,therelativemotionbetweentheprobeand sampleshouldbesteady,controlled,andpreferablywithconstant velocity.Also, thedevice shouldenableeasy changeof materi-alsoneithersurfaceinordertobeabletotestdifferentmaterial combinations.

Signalsgeneratedbyalightcontactarenormallylocalizedand canbedetectedbycollocated sensors.However,theyareoften

(3)

A.Akayetal./Wear276–277 (2012) 61–69 63

Fig.2. (a)Contactforcecomponentsbetweentheprobeandarm.(b)Schematic descriptionofthecontactforceasittravelsonthesampleandthemomentarmh.

contaminatedbyancillarysignals,suchasthoseassociatedwith thepowerand transmissioncomponentsofthesystem,as well assourcesexternal tothetest apparatus,effectsof whichneed tobemitigated. Justas a ﬁngersenses surface properties with themechanoreceptorsinthedermisnearestthecontactarea,the apparatusshouldbeabletosensethesurfacetopography,suchas itsroughnessandcontour,asslidingtakesplace.Itisdesirablethen thatthenoiseandvibrationproducedbyotherpartsofthesystem thanthecontactareashouldbenegligiblecomparedwiththe sig-nalstobemeasured.Theapparatusshouldalsohavetheability

tomonitorsoundradiationandvibrationresponseduetorubbing and,asimportantly,havetheabilitytomeasurefrictionforceand adhesionduringcontact.

3.2. Designdescription

The test apparatus,depicted in Fig. 1, consists of four sub-systems.Theseare:

i.An arm assembly that consists of a hollow box beam, an 203.2mmdiameterdrivepulley,motionsensorﬂags,bearing housing,andaﬂoatingcontactplatform.

ii.Afingerprobeassemblythatconsistsofafingeractuator,fine andcoarseheightadjustments,andaparallelposition adjust-mentmechanismforthefinger.

iii.Thedrivetrainconsistsofadrivemotor,beltsandpulleys. iv. Themotioncontrollerconsistsofthecircuitsthatcontrolthe

armmotion.

Eachofthemechanicalpartsisdescribedinthefollowing sub-sections.

3.3. Thearmassembly

Thecantileverarm,constructedfromahollowboxaluminum beamwithrectangularcrosssection,rotatesaboutaverticalpost towhichitisattachedwithhighprecisionbearings.Thearmcan swinginahorizontalplane,sweepinga90◦arctoslideunderthe stationaryﬁngerprobe.Powertorotatethearmisprovidedbya DC-drivebrushmotor poweredbyasinglevoltage.Atwostage beltsystemisused:a12.7mmtimingbelttransmitstorquefrom themotortotheintermediatepulleys.Thebeltisarrangedsuch thatitsthreadsfaceawayfromthecenterofthepulleyinorder

Fig.3.Schematicdescriptionoftherelationshipbetweenthemeasurednormalforcesandcontactforce.Thedistancetothecontactsurface,h,providesthemomentarmthat translatestangentialforcetonormalforces.Thediagonaldottedlinesshowtherelationshipbetweentheforcecomponentsmeasuredbyeachforcetransducerbelowthe contactplatformwiththeirsumshownasahorizontalline.Thesolidlinesshowtheinﬂuenceoffriction,throughamomentarmthatchangesthenormalforcesmeasured byeachtransducerandreducestheoverallcontactforce.Forcesfromthetrailingtransducersstarttosensenormalloadwithalinearlyincreasingmagnitudeastheforce travelstowardit.Theleadingtransducersensesthefullbearingoftheforceatthestartofthecontactandlinearlydecreases.N=W/(1+cot).

(4)

Fig.4.Asamplemeasurementofnormalforcesbythetwoforcetransducersand theirsum.ArtiﬁcialskinisusedontheprobewithPDMSasthearmmaterialwitha lubricantoiltoreducefriction.Theoscillationsandthechangeinthesumthroughout thecontactdurationresultfromdynamiceffectsofthearm.

toavoidunnecessaryvibrations.Themotorbeltridesonapulley thatcanbeattachedtothemotor.Themotorthendrivestheshaft throughacoupling.

A203.2mmdiameterpulleyisusedtotransmitthedrivetorque tothearm.Placedunderneaththebearinghousing,thispulleyis concealedbyanenclosuretoreducesoundtransmissiontothetest area.Thedrive-belttransmitstorquefromtheintermediate pul-leystothearmpulley.Aswiththemotorbelt,thethreadsofthis 12.7mmwidetimingbeltalsofaceawayfromthepulleysurface. 3.4. Floatingcontactplatform

A ﬂoating contact platform carries the test sample on the arm and permits direct normal force measurements resulting fromﬁngercontactonthetestsample.Theplatformconsistsof two50.8mm×101.6mmaluminumplatesmatedtogetherwith

Fig.5.MeasurementswiththesamematerialsasinFig.4butwithoutoil,resulting inhighfrictionandmuchlowercontactforce.Theadhesionbetweenthesurfaces hasaliftingeffect.

Fig.6. Ademonstrationoffrictioneffectisachievedbycoveringone-halfofthe contactsurfacewithscotchtape,thenon-adhesivesidefacingcontact.This mea-surementdelineatesthedynamiceffectscausedbyfriction.Astheprobeslidesover thelow-frictionsurfacetoamoreadhesiveone,itliftsoffthesurfaceandstartsto reboundasindicatedbytheoscillatorybehavioroftheforcesignals.

machine screws at their corners. The bottom plateis attached directlytotwoforcetransducersatitsends,illustrated schemati-callyinFig.1.Thetopplateservesasthesmootharmsurfaceon whichthetestsampleismounted,coveringthemountingholeson thebottomplate.Thesamplematerials,25.4mmwideand100mm longstrips,areadhesively mountedonthetopplate.Withthis arrangement,theplatesandthesamplearesupportedonlybythe twoforcetransducersattachedtothebottomwallofthehollow boxbeamandareotherwiseisolatedfromthearm.

Theﬂoatingplatformisplacedacrossthewidthandnearthefree endofthecantileverarm,nominally508mmfromitsaxisof rota-tion,subtendinganangleof10◦,andcanaccommodatedifferent materialsandsurfacetreatments.

Fig.7. NormalforcemeasurementsonaPDMSwithdrysmoothlandingsbetween parallelgrooves.Thismeasurementshowstherubbingofthewaferwiththedry surface.Becauseofthesmoothnessofthewafer,thetangentialdissipativeforceis large.

(5)

A.Akayetal./Wear276–277 (2012) 61–69 65

Fig.8.NormalforcemeasurementsonaPDMSwithmicro-roughnessinducedon thesurfacesbetweenparallelgroovesthatreducesadhesionandincreasesnormal contactforce.Theaddedmicro-roughnessreducesdissipativeforces.

3.5. Fingerprobeassembly

Likeapendulum,thefingerprobeisdesignedtorotatefreely aboutastationaryhorizontalpinjointwithanultra-lowfriction bearing.Therelativeheightofthepinjointisadjustedsuchthat theaxisofthefingerprobemakesa45◦anglewiththearmduring contact.Toassureasmoothand continuousmotion,thesliding contacttakesplaceonlywhenthearmmovesinthesamedirection astheanglethependulummakeswiththenormaltothearm,thus resultingintensionintheaxisofthefingerprobeandpreventing developmentofacompressiveforceinitsaxisthatwouldleadto ratchetingofthefingerprobe.Toaccomplishthis,anembedded electromagneticactuatoratthepinjointofthefingerliftsthefinger probeduringthearmretractionswingtoavoidacollisionwhenthe directionofthearmmotionisreversed.Inthisconfiguration,the fingerprobeisabletofollowsurfacecontourswithradiilargerthan thatofprobetiptoassurecontinuouscontactduringtherubbing event.

Theelectromagneticactuatoratthepinjointisalsousedto pro-videatorquetothefingerprobeinitsplaneofmotioninorderto applyadesiredforceonthearmsurface,ontheorderof0.2–0.8N, similartothatexertedduringactualhumantouch.Therelative slid-ingspeedbetweenthefingerprobeand thesampleonthearm isnominally125mm/s,representativeoftheprobingspeedofa humanfingeronasurface.

Thetippartofthefingerprobecanbecoveredwithasleeve madeofanysofttestmaterial.Boththefingermaterialandthearm materialcanbechangedtotestdifferentmaterialcombinations andsurfacetreatments.Useofsoftmaterialatthetipofthefinger probeprovidesbothradialandtangentialcompliance.Soft mate-rialonthearmsurfacehascomplianceonnormalandtangential directions.

Thefingerassemblyismountedonarigidpostthatalsohasa combinationofcoarseandfineadjustmentmechanismstoposition itsheight.Thefineadjustmentisusedtobringthefingertothe desiredheightforcontactwiththearmmaterialundertestsuch thattheflatmountforthenormalaccelerometerisparalleltothe armsurface.Thisalsoassuresthatthefingerisata45◦angleduring slidingcontact.ThefineadjustmentstageismadeusingaDel-Tron lineartranslationstagewithanattachedmicrometer.

3.6. Motioncontroller

Acustomelectronicspackageisusedtocontrolthemotionofthe apparatus.Theoperatingparameters,includingthecontactforce and rubbingspeed, are enteredinto themotioncontroller.The apparatuscanthenruncontinuously.Themotioncontrollerlifts theﬁngerprobefromthepathofthearmonthereturnstroke. 3.7. Systemdynamics

Acriticalaspectofsuchanapparatusistoavoidancillary vibra-tionsandnoiseduringtheoperation.Particularlyforsoftmaterials andlightcontactloads,anynoisefromthemotorandthepower transmissionmechanismoftheapparatusmustbemuchlessthan thesoundgeneratedbyslidingtheﬁngeroverthearm.

In thecase ofthe design described here,thetest apparatus produces negligible mechanical noisecompared tothe acoustic emission fromthe contact area. For testsusing very light con-tactforces,tofurthermitigateexternalnoiseproblems,anacoustic enclosurewasdesignedandconstructedtoenclosethetestsection ofthearmandthefingerkeepingthedrivemechanismoutsidethe measurementspace.Notwithstandingthelowinterferencebythe drivetrain,armoscillationsduringmeasurementalsoneedtobe reduced.Reductionofvibrationamplitudeofthefirstcantilever modeofthearmwasaccomplishedbyusinga25.4mmdiameter stainlesssteelrodforthemainrotationalshaftofthearmaswell asstiffeningofboththearmandthebearinghousing.Withthese adjustments,thefirstbendingandtorsionalnaturalfrequenciesof thearmweremeasuredas57Hzand560Hz,respectively.

Thefollowingdynamiceventscanbesensedbytheforce trans-ducers,accelerometers,and microphones:(i)Impactforcesthat developduringtheinitialcontactoftheﬁngerprobewiththe sam-pleonthearm,(ii)normalforceexertedbytheprobeasittravels acrossthesampleonthearm,(iii)impactforcesthatresultfrom thereboundofprobetipfollowingliftoffthesamplesurfacedue toadhesive forces,(iv)stick-slipeventsastheprobeslidesover anadhesivesurface,(v)accelerationsignalsrepresentingsurface roughness,and(vi)torsionaloscillationsofthearmexcitedbythe stopandstartasitchangesdirectionofmotion.Amongthesethe impactforcethatcandevelopasthearmmakesaninitialcontact withthestationaryprobe,case(i),ispreventedbyattachingasmall ﬂexiblemembranewithwhichtheprobemakesitsinitialcontact andslidesoverittothesamplesurface.Followingthislanding,the probetipalignswiththearmsurfacethusavoidingimpact exci-tationoftheprobetipaswellasadditionalexcitationofthearm vibrations.

Instrumentationtodetecttheseeventsandthemeasurement examplesaregiveninthefollowingsections.

3.8. Instrumentation

Twominiatureaccelerometers(PCB352A24)withasensitivity of100mV/g[10.2mV/(m/s2_)]_and_a_frequency_range_of_1–8000_Hz

and0.8gmwereusedtosensetheverysmallchangesinthe accel-erationofthefingerprobetipduetosurfacetextureandadhesion asittravelsoverthearmsample.Theyweremountedonthetwo respectiveflatsurfacesontheinteriorofthefingerprobetip,just abovethecontactarea,withaxesparallelandnormaltothearm surface,respectively,asindicatedinFig.1.Theseaccelerometers areACcoupledandeachconnectedtoeitheranIntegratedCircuit Piezoelectric(ICP)powersupply ortoadataacquisitionsystem thatcanactasanICPdevice.Duetotheirproximitytothe con-tactarea,theseaccelerometerscoulddetectvibrationsinducedby surfaceroughness.

Amicrophone(B&K4165)placedveryclosetothecontactarea, approximately25mmawayfromtheoutersurfaceoftheﬁnger,

(6)

Fig.9.Theseriesofspectrogramsshowtheaccelerationsignalsrecordedbythesensormountedinthenormaldirection.Eachofthesevenspectrarepresentsdifferenttimes duringoneslidingperiodoftheprobeoverthearm.

togetherwiththenecessaryinstrumentation,recordedthesounds generatedinthecontactarea.Theaxisofthemicrophoneisdirected towardcontactarea.

Twopiezoelectricforcetransducers(PCB208C01)thatsupport thefloatingcontactplatformatthefree-endofthearmareused tomeasurethenormalforceexertedbythefingerprobe. Based ontheirconfiguration,thesamenormalforcemeasurementsalso yieldfrictionforceduringtherubbingeventsasdescribedinthe nextsection.Todescribetheoperationofthesystem,theleading transducerreferstothetransducerthatisfirstpassedoverasthe fingerandarmcomeintocontact.Thetrailingtransducerrefersto thetransducerthatispassedoverbeforethefingerterminates con-tactwiththearm.Thetransducersarelocatedinsidebox-channel, eachmountedonanaluminumblockatitsbottomandconnected

Fig.10.Thismeasurementshowsfrictionforcesthatdevelopduringslidingofthe probetipoversilkmadeofsyntheticsilkﬁberswoveninacomplexpattern.

tothecontactplatformatthetop.Eachforcetransduceris con-nectedviaanICPpowersupplytoadataacquisitionsystem.Their signalsareDCcoupled,inordertobeabletomeasureDCloads.

Itis worthnotingthatsince theseareforcetransducers,the voltagefromthetransducerswilldecayovertime(dischargetime constant>50s)afteraDCloadisapplied.Theyprovideaccurate measurementsaslongastheeventisshortcomparedtothetime constantofthetransducer,asinthepresentcase.Useofloadcells presentanalternativesensorchoice,andmightbeusefulformuch slowerrubbingspeeds.Althoughloadcellsgreatlyenhancethe sta-bilityofthestaticloadmeasurement,theyhavealowerdynamic response.

3.9. Contactforcemeasurementmethod

Apairofidenticalforcetransducerswasusedtomeasurethe normalandtangential(friction)forcesduringrubbingevents.As thearm rotates and comesinto contact withthe ﬁngerprobe, theforcetransducersmountedinthearmstarttosensetheforce appliedtothesampleapportionedaccordingtothecontactposition ofthetravelingforce.

ReferringtoFig.2a,thereisakinematicrelationshipbetween thenormalcontactforce,N,andtheweight,orappliedforce,Wto thetipoftheﬁngerprobe:

N=₁ W +cot

whereistheangletheﬁngerprobemakeswiththenormalto thearm,inthiscase45◦,andreferstothecoefﬁcientoffriction duringcontact.Wenotetheliftforcecausedbyfrictionas: W−N=Fcot=Ncot

Thecontactforcesappliedtothesurface,thenetnormalforce, andthemomentduetofrictionforceareapportionedtothetwo

(7)

A.Akayetal./Wear276–277 (2012) 61–69 67

Fig.11.Changeoffrequencieswithspeedofarmduringcontactasdemonstratedbytheslopesofthespectraandtheirharmonics.

supportsneareitherendoftheplatformwheretheforce transduc-ersareplacedtomeasurethenormalforcesas

NL= ₍₁₊W cot)

_L₋_d L − h L

NT=₍₁₊W cot)

_d L+ h L

wherethesubscriptsLandTrefertotheleadingandtrailing posi-tionsoftheforcetransducers.Aspreviously stated,theleading transduceristheonethatisclosertotheinitialcontactlocation. Sincecontactismadeinonlyonedirectionofmotionofthearm, thesedesignationsremainthesame.Theamplitudesofthetwo normalreactionforceschangewithdistance,d,ofthecontact loca-tionfromtheleadtransducerlocation.Presenceoffrictionmodiﬁes themeasurednormalreactionforcesbymeansofamoment pro-ducedbythefrictionforceandthemomentarmh,thedistance fromthetransducersurfacetothesurfaceofthesamplematerial wherecontacttakesplace,asshowninFig.2b.

Fig.3schematicallydemonstrateshowthenormalforcesatthe twotransducerpositionsvarylinearlyastheappliedcontactforce movesacrossthesamplesurface,withandwithoutthepresenceof frictiononthesurface.AsshowninFig.3a,withoutfriction,as con-tactisinitiatedatd=0,theleadingtransducercarriesthefullweight ofthecontactload,whichlinearlydecreasesandvanishesasthe contactpointtravelstowardandreachesthetrailingtransducer, d=L.Simultaneously,thetrailingtransducerstartstosensea nor-malloadasthecontactpointstartstotraveltowardit,andincreases linearlytofullmagnitudewhenitreachesthetrailingtransducer. Theirsum,shownbythehorizontaldottedline,remainsconstant andequaltothecontactloadwhich,intheabsenceoffriction,is equaltotheappliedforce,W.

Inthepresenceoffriction,however,theslopesandtheinitial andﬁnalmagnitudesofthediagonallineschange,asindicatedby thesolidlinesinFig.3b.Theliftforceresultingfromthemoment introducedbythepresenceoffrictionreducestheinitial magni-tudeofNL andincreasesNT,respectively,byanamountequalto

Nh/L.Theirsumatanytimeduring thesliding processequals thetotalnormalcontactforce,N.Thus,thechangeinthenormal contactforceNdirectlyrelatestothetangential,orfriction,force

inaccordancewiththeinverserelationshipbetweenfrictionand normalforce,expressedbythekinematicrelationgivenabove.In summary,frictionandadhesionforcescanbemeasureddirectlyby measuringthenormalcontactforcesusingtwoforcetransducers asthecontactforcetraversesoverthesample.

3.10. Measurementexamples

Typical contact force measurements, using the method describedabovearedisplayedinFigs.4and5forlow-and high-friction contacts, respectively. The low-friction sample used in Fig.4waspreparedbyapplyinglight lubricanttothenormally adhesivePDMSthatwasusedforFig.5.Theforcesmeasuredby theleadingandtrailingtransducersandtheirsum,which repre-sentsthenormalcontactforcethatdirectlyrelatestothefriction (tangentialcontact)force,illustratethemeasurementtechnique describedabove.Thesemeasurementsweremadeusingafinger probecoveredwithBioskinproducedbyBeaulax,askin-like poly-mer,sometimesreferredasartificialskin.Thefingerrubsoverthe flatsurfaceontherotatingarmasdescribedabove,whichiscovered bypolydimethylsiloxane(PDMS).ThePDMSsamplesweremolded fromflatsiliconwafers.Themoldedsampleshadasmooth,sticky andsoftfeeltothetouch.Inthesemeasurements,thetransducers arepositioned60mmapartandthenormalload,10.6g,remained thesame.

Astheprobetraversesacrossthesampleonthearm,themeanof themeasurednormalforcebyeachtransducerchangelinearlywith distanceofthecontactpointfromforcetransducers,asillustrated inFig.3.Theoscillationsaboutthemeanrepresentthetorsional oscillationsof thecantilever arm aboutits axis,excitedby the motor-beltsystem,andarealsoobservedbeforeandaftertheprobe travelsoverthesample.Becausethesenearlyharmonicoscillations areoutofphase,theycanceleachotherwhentheyaresummed anddonotappearinthesumofthereactionforcesasshownin Figs.4and5.

Adhesionbetweentheprobe andthe samplesurface onthe armgivesrisetoadditionaldynamicphenomena.Whentheﬁnger

(8)

probecomesintocontactwithorasitslidesoveranadhesive sur-face,undercertainconditions,thetangentialadhesiveforcelifts theprobe-tipoffthearmsurface.Followingabrieffree-flight,the probebouncesbackonthesurface,followedbyrebounds.An exam-pleofsuchareboundisobservedintheforcemeasurementsshown inFig.6,whichalsogivesacleardemonstrationofthisflutter-like phenomenon.Thetestsampleonthearmmeasuredtogenerate Fig.6hastworegionswithdifferentsurfaceadhesion.Theleading partofthesampleiscoveredwithasmoothmembranewithvery lowfrictionand thesecondhalfofthesurfacehashighly adhe-sivedryPDMS.Asthefingerprobemovesoverthefirsthalfofthe sample,theresponseisthesameasthatinFig.4,showinga rela-tivelyhighnormalcontactforce,representativeoflowfriction.The responseofthetwoindividualtransducerssumsuptoasmooth totalnormalcontactforcedevoidofoscillations.

Astheprobetipcomesintocontactwiththeadhesivepartof thesurface,thependulum-likeprobeliftsoff,itstipmakinganarc aboutitshinge,andreboundsleadingtoadditionaloscillatory sig-nals.Themeanamplitudeofthetotalnormalforceisloweronthis partofthesampleduetotheliftforcecausedbyadhesion.The oscillationsrepresentthependulumreboundsoftheprobe.Both themagnitudeoftheliftandthereboundarerelatedto,andthus containinformationabout,theadhesiveandelasticpropertiesof thesurfacesmaterials.

Thehigherfrequencyoscillationsobservedinthesecondhalfof thesampleinFig.6,aswellasinFig.5,representthelocal stick-slipbetweentheadhesivesurfaceandtheprobe,andareabsent inthemeasurementsthatinvolvelow-frictionsurfaces.Similarly, thereboundsareabsentintheforcemeasurementsshowninFig.4 whereasthemeasurementsfor thecase witha strongadhesion (Fig.5)containstrongoscillationsofthenormalcontactforces.

In cases where the contact forces that develop due to sur-face roughness have much smaller amplitudes compared with thenormalforces,theireffectsarenotaseasilydetectedbythe forcetransducers.Instead,thetwoaccelerometersmountedonthe insideof thefingerprobe-tip areused toobtainspectral infor-mationaboutthesurfacecharacteristics.Todemonstratesurface roughnesseffects,threesiliconwaferswithdifferentsurface prop-ertieswereusedassamplesonthearm.Twoofthewafers had periodicgroovescutintothesurfaceandthethirdwasaperfectly flat(standard)wafer.Thegrooveshadthesametrapezoidalcross sectionandwereperiodicallyspaced250␮mapart.Eachgroove was80␮mwideatthetopand58␮mdeep.Thetwowaferswith thegrooveswereidenticalexceptforthesurfaceroughness.One ofthegroovedwafershadbeenchemicallyetchedsothattheland (surfacebetweenthegrooves)hadamicro-roughness.Theother waferwasperfectlyflatonthelands.Eachwafercouldberubbed indifferentdirections.

Normalforcemeasurementsoftheetchedwafersaregivenin Figs.7and8forthecasewhenthelandingbetweenthegrooves issmoothandhasmicro-roughness,respectively.Neithercasehas discernableforceresponseduetothegrooves.Thesmoothcase, Fig.7,hashighadhesionandthereforelownormalcontactforce, similartoFig.5.Thesurfacemicro-roughnessbetweenthegrooves, however,reducesadhesionand,asaresult,hashighernormal con-tactforce.Inbothofthesemeasurements,thedecayingoscillations shownintheleftsideoftheplotsresultfromtheimpactofthe pen-dulumtipwiththesampleonthearmatinitialcontact,whichdo notappearinFig.4or5sincethependulumwasaffordedasoft landingasdiscussedabove.

Roughnesseffectsfromtheperiodicallyspacedgroovescanbe seenintheaccelerationresponseofthependulumtip.Fig.9,for example,showsaseriesofnormalaccelerationspectrograms rep-resentingnormalaccelerationoftheﬁngerprobetip.

Thesemeasurementsdirectlyrepresentthetextureofthe sur-faces.EachspectrograminFig.9representsthefrequencycontent

ofaccelerationsignalsatdifferenttimes,correspondingtodifferent partsofthearmmaterial,duringasinglesweep.Forthis partic-ularmeasurementtheﬁngerprobe surfacewascovered witha softmaterial andthearmsurfacewasPDMS(10:1)with paral-lelgrooves.Thefrequencypeakscorrespondtothespacingofthe grooveswithafundamentalofapproximately670Hzandits har-monics.Theadditionoftheaccelerometerstotheapparatusallows foradditionalsurfacecharacteristicstobemeasuredthatcannotbe discernedwiththeforcetransducersalone.

Inthosecaseswherethesurfacehaspronouncedtopography, asexpected,thependulum-likefingerprobewillexertsufficient pulsesonthesamplesurfacetobesensedbytheforcetransducers. OnesuchexampleisgiveninFig.10wherethesampleconsistsof asilkmadeofsyntheticsilkfiberswoveninacomplexpattern.

Themicrophoneinthesystemcanalsomeasureperiodicevents duringthe rubbingmotion. It shouldalsobe notedthat asthe armcomesintocontactwiththeﬁngerprobe,dependingonthe appliedforcebytheprobe,thearmmayinitiallyslowdownand thenspeedupagain.Theincreasingslopesofthespectrallinesin Fig.11demonstratethisspeedchange,whichbecomesmoreacute athigherharmonics.Theharmonicsinthespectrapointto non-linearnatureoftheinteractionduringslidingaswellastheslight aperiodicityoftheridgesonthesurface.Asexpected,thesound pressuremeasurementsmadeverynearthecontactareashowthe sametrendsastheaccelerationsignals.

4. Discussionandconcludingremarks

Theconnectionbetweentextureandmaterialofasurfaceand thesubjective dimensionsthat describeperception of itis first madethroughthemechanoreceptorsinthedermisoffingertips. Objective evaluationsof the surfaces bymeasuring parameters that influencethe correspondingsubjective dimensionsrequire accuratephysicalcharacterizationofmaterialsandsurfaces.Such measurementscanultimatelyleadtoestimationofsubjective char-acterizationofsurfacesbasedonobjectivemeasurableparameters. Theapparatusandthemeasurementmethoddescribedinthis paperaredesignedtoemulatethecontactconditionsbetweena fingerandasoftmaterial.Thedevicehasthecapabilitytomake simultaneousmeasurementsofdifferentdynamicsignals gener-atedasthefingerprobelightlyslidesover thesurfaceofa soft material. These include acceleration signals measured in close proximityof thecontact point within theprobe tip,which are causedbyandthusprovideinformationabout,thesurfacetexture. Boththeaccelerometersignalsandthesoundpressurerecorded nearthecontactareacontainfrequenciesfarabovethosethatcan bedetectedbythemechanoreceptorsandprovideadditionaluseful information.

Theapparatusdescribed herealsohasa uniquecapabilityto measureadhesionbetweensurfaces,whichdirectlyrelatestothe “stickiness” ofa surface. Themethoddeveloped herecan mea-sureanddistinguishadhesionbetweensurfacesandovercomes thedifﬁcultiesassociatedwithmeasuringfrictionforcebetween softsurfaces.Useofapairofforcetransducersallowsthe measure-mentofboththenormalcontactforceandthetangentialdissipative force,withoutinterfering withthesliding process.Themethod reliesontheliftforcethatdevelopsthrough amoment-arm,h, whichisthedistancetothesamplesurfaceabovetheforce trans-ducers.Thedesiredmeasurementsensitivitycanbeachievedby adjustingh.

The sensitivity of the present apparatus in making air- and structure-borne acoustic measurements under light loads is enhancedinpart,duetothesmoothandquietmotionthearm. The systemdynamics have negligiblecontribution tothe mea-surements,save forthetorsionaloscillationsthat appearinthe

(9)

A.Akayetal./Wear276–277 (2012) 61–69 69

individualreactionforcemeasurementsbutnotontheirsum,as describedearlier.Samplemeasurementsshowhowtheordinary deﬁnitionofcoefﬁcientoffrictiongiveswaytoconsiderationof frictionforceasapureadhesion,withnormalaswellastangential componentsasillustratedinFig.6,clearlydistinguishingthetwo regimesthathadtwosurfaceproperties.

Thedevicedescribedherecanbeusedtodetermineand corre-latetribologicalparameterstodynamiceventsandeventuallyto subjectiveevaluations.

Acknowledgments

AuthorsacknowledgegeneroussupportofUnileverR&D(USA). Oneoftheauthors(AA)gratefullyacknowledgessupportbythe NationalScienceFoundationwheretheauthorAAwasemployed during part of the period when this research was conducted andgeneroussupportofEUMarieCurieprogramforsubsequent research.

References

[1]S.J.Lederman,Auditorytextureperception,Perception8(1979)93–103. [2]G.Lundborg,B.Rosén,S.Lindberg,Hearingassubstitutionforsensation:anew

principleforartiﬁcialsensibility,J.HandSurgery24A(1999)219–224. [3]C.Kayser,C.I.Petkov,M.Augath,N.K.Lohothetis,Integrationoftouchandsound

inauditorycortex,Neuron48(2005)373–384.

[4]S.Guest,C.Spence,Whatroledoesmultisensoryintegrationplayinthe visuo-tactileperceptionoftexture?Int.J.Psychophysiol.50(2003)63–80. [5]R.Newport,J.V.Hindle,S.R.Jackson,Linksbetweenvisionand

somatosensa-tion:visioncanimprovethefeltpositionoftheunseenhand,Curr.Biol.11 (2001)975–980.

[6] A.Akay,Acousticsoffriction,J.Acoust.Soc.Am.111(2002)1525–1548. [7]T.Yoshioka,S.J.Bensamaïa,J.C.Craig,S.S.Hsiao,Textureperceptionthrough

directandindirecttouch:ananalysisofperceptualspacefortactiletexturesin twomodesofexploration,Somatosens.Mot.Res.24(1–2)(2007)53–70. [8] MarkHollins,AdamSeeger,GabrielePelli,RussellTaylor,Hapticperception

ofvirtualsurfaces:scalingsubjectivequalitiesandinterstimulusdifferences, Perception33(2004)1001–1019.

[9] MichelParé,HeatherCarnahan,M.Allan,Smith,Magnitudeestimationof tan-gentialforceappliedtotheﬁngerpad,Exp.BrainRes.142(2002)342–348. [10] M.J.Adams,B.J.Briscoe,S.A.Johnson,Frictionandlubricationofhumanskin,