• Nenhum resultado encontrado

The ultra-precision -lap grinding U of flat advanced ceramics Journal of Materials Processing Technology

N/A
N/A
Protected

Academic year: 2022

Share "The ultra-precision -lap grinding U of flat advanced ceramics Journal of Materials Processing Technology"

Copied!
21
0
0

Texto

(1)

ContentslistsavailableatScienceDirect

Journal of Materials Processing Technology

j ou rn a l h o m epa ge :w w w . e l s e v i e r . c o m / l o c a t e / j m a t p r o t e c

The ultra-precision U d -lap grinding of flat advanced ceramics

Arthur Alves Fiocchi

a,∗

, Luiz Eduardo de Angelo Sanchez

b

, Paulo Noronha Lisboa-Filho

c

, Carlos Alberto Fortulan

a

aUniversityofSãoPaulo—USP,DepartmentofMechanicalEngineering,Av.TrabalhadorSaocarlense400,13566-590SãoCarlos,SP,Brazil

bSãoPauloStateUniversity—Unesp,DepartmentofMechanicalEngineering,Av.LuizEdmundoCarrijoCoube,14-01,17033-360,Bauru,SP,Brazil

cSãoPauloStateUniversity—Unesp,DepartmentofPhysics,Av.LuizEdmundoCarrijoCoube14-01,17033-360,Bauru,SP,Brazil

a r t i c l e i n f o

Articlehistory:

Received10April2015

Receivedinrevisedform1October2015 Accepted4October2015

Availableonline23October2015

Keywords:

Ud-lapgrinding Advancedceramics Ductilegrinding Surfacecharacterization Nanotechnology

a b s t r a c t

ThepresentstudyfocusesontheUd-lapgrindingprocessanditsmachine-tooldesignaimingatultra- precision(UP)manufacturingofadvancedceramics.Impactsofthreedifferentoverlappingfactorsona dressing(Ud)andthreeabrasivegritsizesofconventionalSiCgrindingwheelswereanalyzedonflatnano- metricsurfacefinishingofdensediscsof3Y-TZPinaductileregimeofmaterialremoval.Microhardness, contactandopticalprofilometry,SEM-FEG,Ramanspectroscopy,XRD,andconfocalepi-fluorescence wereappliedtocharacterizetheceramicmaterial.Mechanicalandelectric-electronicdesignsofthe machine-toolweredevelopedtowardtheUPceramicgrinding.Thedesignmethodologywassuccessful forsupportingtheachievementofalldesignrequirementsoftheCNCFiocchiLapGrinder.Themachine- toolandtheUd-lapgrindingprocesswerecapableofmanufacturingflat3Y-TZPsurfaceswithnanometric finishingwithoutintroducingcriticaldefects.Thebestfinishing,Ra=60.63nm,camefromthe#300grind- ingwheeldressedwithUd=5.Aflatnessdeviationof0.308␮mwasobtainedthroughthe#800grinding wheelandUd=3.Differencesbetweentheoreticalandrealmacroandmicroeffectsoverthegrinding wheelsaftersingle-pointdiamonddressing,epoxybondstrength,abrasiveprotrusion,abrasivegritsize andabrasivefriabilityplayakeyroleintheUd-lapgrinding.Thereisnoreportofanabrasiveprocess capableofachievingsimilarnanometricfinishingwithoutintroducingcriticaldefectswiththesame micrometricgritsizeandtypeofabrasive.TheUd-lapgrindingcanreplacetheengagementofprocesses suchasgrinding,lapping,andpolishingofadvancedceramics.

©2016PublishedbyElsevierB.V.

1. Introduction

Compared with metals, advanced ceramics are materials of higherhardnessandspecialpropertiesthatpresentperformance, reliability, and safety at elevated temperatures; theseare out- comesoftheirexceptionalchemicalstabilityduetostrongionic and/or covalent atomic bonds (Boch and Nièpce, 2007). How- ever,advancedceramicssufferfromtheirrelativelylowfracture toughness,whichalsosignificantlyaffectstheircostasaresultof post-sinteringprocessingnormallydonebyexpensivesuperabra- sivetoolsandmanufacturingprocesses(Janssenetal.,2008).

Exceptforglass,plasticmoldingandcastingofadvancedceram- icsare exceptionallyhighcost, mainlybecauseofthe chemical bondsthatraisethemeltingpoint.Brittlenessisalsoasignificant

Correspondingauthor.

E-mailaddresses:arthuraf@eesc.usp.br

(A.A.Fiocchi),sanchez@feb.unesp.br(L.E.deAngeloSanchez),plisboa@fc.unesp.br (P.N.Lisboa-Filho),cfortula@sc.usp.br(C.A.Fortulan).

characteristicoftheseceramicsbelow1000C.Inadditiontolim- itingtheiruse,thebrittlenessimposesrestrictionsonproduction techniques.Alltheselimitationsofceramicmanufacturingexplain whythebasictechniqueofdensificationisthesinteringofcom- pactedpowder(BochandNièpce,2007).Thisoppositionisquite evidentbetweenmetals(ductile)andceramics(fragile).

Asceramicsaremorebrittle,themainmechanicalfailureusu- allyoccursbycrackpropagation(ArgawalandRao,2008).Initial criticalcracks,precursorsofthefailure,aredefectsusuallyorigi- natinginirregularitiesacquiredintheearlystagesofobtainingthe piece(frompowdertosintering)(Bukvicetal.,2012)and/orduring itspost-sinteringprocessing(hardceramicmachining)(Marinescu etal.,2014),hencethedesignofaproductofadvancedceramic needsspecialcareineachphaseofitsmanufacturing.

Ingenerallines,intheproductionofaceramicartifact,theraw materials (powders) are milledor synthesized, are mixed, and receiveadditivessuchasbindersandplasticizerstoassistthecon- formation,compaction,andsintering(Scheller,1994).Theceramic powderisformedintocompactedpieces,alsocalledgreenpartsdue totheirnotheattreatedstate,whicharesubsequentlysinteredto http://dx.doi.org/10.1016/j.jmatprotec.2015.10.003

0924-0136/©2016PublishedbyElsevierB.V.

(2)

E Elasticmodulus(GPa)

EESC SãoCarlosSchoolofEngineering ELID Electrolytic-inprocessdressing FAPESP SãoPauloResearchFoundation FEB BauruCollegeofEngineering HFW Horizontalfieldwidth(mm) HV Hardnessvickers

HV Highvoltage(kV)

IFSC SãoCarlosInstituteofPhysics

k Parameter formonoclinicphasequantificationby Ramanspectroscopy

KIC Fracturetoughness(MPa.m0.5) LATUS LaboratoryofMachiningTechnology ln Totalmeasuredlength(mm)

LTC LaboratoryofTribologyandComposite lr Singlemeasuredlength(mm)

m Monoclinicphase

MEMS Micro-electromechanicalsystems mf Finalmass(g)

mi Initialmass(g)

MQF Minimumquantityfluid

MRR Materialremovalrate(g/minute) NEMS Nano-electromechanicalsystems ng Grindingwheelrotationalspeed(rpm) nw Workpiecerotationalspeed(rpm) P Ud-lapgrindingpressure(kPa) PABA 4-Aminobenzoicacid

PVB Polyvinyl-butyral PVD Physicalvapordeposition

Ra Arithmeticmeansurfaceroughness(␮m) RMS Rootmeansquare

rp Dressernoseradiusgeometry(mm) Rt Totalheightoftheroughnessprofile(␮m) r Stylustipradius(␮m)

Sa Arithmeticmeanheightoverthecomplete3Dsur- face(␮m)

Sd Dressingfeed

SE Secondaryelectrondetection

SEM-FEG Scanningelectronmicroscopeequippedwithfield emissiongun

SiC Siliconcarbide

Sku Kurtosisofthe3Dsurfacetexture

Sp Maxpeakheightoverthecomplete3Dsurface(␮m) Sq Rootmeansquareheightevaluatedoverthecom-

plete3Dsurface(␮m)

Ssk Skewnessofthe3Dsurfacetexture

Sv Maximumvalleydepthoverthecomplete3Dsur- face(␮m)

Sz Maximum height over the complete 3D surface (␮m)

T Tetragonalphase TZ-3Y-E Tosohzirconiagrade t→m Martensitictransformation

Ud Overlapfactororoverlapratioondressing UNESP UniversidadeEstadualPaulista

W Realwaviness(␮m)

Y-TZP Tetragonal zirconia polycrystals stabilized with yttria

ZrO2 Zirconiumoxideorzirconia

3Y-TZP Tetragonalzirconiapolycrystals stabilizedwith 3 mol%ofyttria

␣ Singleabrasivereliefangle()

␤ Singleabrasivewedgeangle()

␥ Singleabrasiverakeangle()

␦ Parameterformonoclinicphasequantification by Ramanspectroscopy

␪ Diffractionangle()

c Cutoff(mm)

␳ Density(g/cm3)

␴ Bendingstrength(MPa) Ø Diameter(mm)

# Meshnumberofthegrindingwheel(abrasivegrit size)

achieve the desired hardness and density (Richerson, 2006).

Eachstephasthepotentialtoproduceundesirablemicrostructural flaws in the piece, which in itself may limit their properties, reliability, and machinability. Scientific investigations are not exclusivelyfocusedonthematerialitself,butalsoinitsprocessing intopowder,granules,andinthesubsequentstagesofobtaininga partaswellasitsmachining.

For less demanding applications regarding form and finish, thesinteringisthelast stageoftheceramicprocessing.Never- theless, several areas of scienceand technological applications requirebetterfinishing,tighterdimensionalandgeometrictoler- ances,andsurfacescompletelyfreeofdamage(Brinksmeieretal., 2010).Suchpartsmust undergo,consequently,furtherprocess- ing stepsaftersintering, especially toconfer onthem complex geometries and finer finishing. Thedemands for bettersurface finishingandhighaccuracyofoptical,electrical,andmechanical componentsaregrowingalongwiththeminiaturizationofhigh performanceproducts, suchaslenses, electroniccomponentsof computersand smartphones,andsensors,as wellas microand nano-electromechanicalsystems(MEMSandNEMS).

Amongtheavailablemethodsofmaterialremoval,theabrasive machiningisthemostwidespreadinindustry.Grindingreconciles stock removal, form correction, and surface quality for large- scaleproductioninaneconomicallyfeasibleway(Malkin,1989).

Although widespread,hard ceramic machiningis an extremely difficulttaskduetotheexceptionalceramicpropertiesinthesin- teredstate,especially:highhardness,highelasticmodulus,high meltingpoint,andfragility(CallisterandRethwisch,2013).These requiresuperabrasivecuttingtools,stiffmachine-toolsandclamp- ingdevices,microandnanometerdepthsofcut,andahighlyskilled workforce(Brinksmeieretal.,2010).

Intraditionalabrasivetechnology,thebrittlemodeisthepri- mary removal mechanism (Marinescu et al., 2004) caused by thepropagationoflateralcracksthatpromotechipping(Xuand Jahanmir,1995).Higherremovalratescanbeachievedinthebrittle modeinoppositiontoductileabrasivecutting(Xuetal.,1995).On

(3)

theotherhand,grindingmayalsogeneratelongitudinalcracksthat canremainintheproductaftergrinding(Marinescuetal.,2007) andact,forexample,asstressconcentratorsthatunderspecific conditionsstimulatesubcriticalcrackpropagation(Devilleetal., 2006).

Themain technologicalchallengeofhardceramicmachining is,thus,topromotematerialremovalwithoutintroducingcritical defectsthatwillcompromisetheperformanceofthecomponent (Marinescuetal.,2014).Therefore,thebrittleremovalmodeshould beavoidedinultra-precision(UP)grindingorminimizedinorder toavoidmainlytheformationandpropagationofsurfaceandsub- surfacecracks,microstructuralchanges,andresidualtensilestress (Brinksmeieretal.,2010).

Tominimizetheemergenceandevolutionofcracks,machining shouldbeconductedwithdiscretion,soastoavoidtheintroduction ofresidualtensilestressonsurfacesofmaterialssubjecttostress- inducedphasetransformation(Devilleetal.,2006).Thesearchfor high-techceramicpartswithfunctionalsurfacesrequires,hence, themostsophisticatedmanufacturingprocesses.

Over the last three decades,researchers have respondedto demandsofindustryformachine-toolsandfabricationprocesses thataimtoattainhigh-precisionpartsinananometer(109m)and angstrom(1010m)forbothdimensionalandfinishrange(Ohmori, 2011).Nevertheless,thesequalitycharacteristicsareverydifficult, oreconomicallyunviable,toachievebyconcatenatingthetradi- tionalabrasiveprocessessuchasgrinding,lapping,andpolishing.

For finishing flat surfaces, researchers have been putting togetherthemainadvantagesofthetraditionalabrasiveprocesses suchas:facegrindingwithconstantpressure,fixedabrasivesfor a two-bodyremoval mechanism,total contactof thepartwith thetool,and lappingkinematicsaswellas somespecificoper- ationstokeep grindingwheelsharpnessandform.In thisway, differentdressingandconditioningtechniqueshavebeenapplied tothetoolbeforeand/orduringgrindingprocesses.Accordingto Sanchezetal.(2011),thesedifferentdressingproceduresyielddis- tinguishedresultsingrindingperformance.Nanogrinding(Gatzen andMaetzig,1997),facegrindingwithlappingkinematicsorgrind- ingonalappingmachine(TomitaandEda,1996),ELID-lapgrinding (Ohmorietal.,2011),flathoning(BeyerandRavenzwaaij,2005), andUd-lapgrinding(Fiocchietal.,2015)aresomeexamplesof thesehigh-endgrindingprocessesforfinishingflatsurfaces.

TheBrazilianprocess,namedUd-lapgrinding,wascreatedto meetthedemandsfora gold-standardabrasiveprocessesasan economicalandaccessibletechniqueforindustry,especiallythe Brazilianone,whichuntilnowemployedonlyforeigntechnologies forUPfinishingofflatadvancedmaterials(Fiocchietal.,2015).

Ud-lapgrindinginvestigationsstartedwithanadaptedlapping machine(Sanchezetal.,2011).FiocchiLapGrinderI(Fiocchi,2010 Fiocchi et al., 2015)came next toestablish the first fully CNC machine-tooltogrindflatmetallicsurfacesusingepoxybonded SiCgrindingwheelsdressedaccordingtotheoverlapfactorthe- ory(Ud)proposedbyKönigandMesser(1980).Thesecondversion (Fiocchi,2014),presentedinthispaper,representsacutting-edge CNCmachine-tooldesign,improvedforgrindingmaterialsthatare notonlyductile,butalsohardanddifficulttomachine.

Products made of tetragonal zirconia polycrystals stabilized withyttria(Y-TZP)haveanoticeableinterestinfunctionalsurfaces thatrequiretop-qualitymechanical(Zhuangetal.,2015),electri- cal(Flegleretal.,2014),andtribologicalproperties(Chevalieretal., 2007).

The3Y-TZPusedundersomespecificconditionspresentsacom- plexphenomenonofsubcriticalcrackgrowthdefined,amongother names, asstresscorrosion (Sikalidis,2011)or low temperature degradation(LTD)(Chevalieret al.,2007).Thismechanismpro- motesthechangeofthecrystallinestructurefromtetragonalto monoclinic(t→m)whentheceramicissubjectedtostressand/or

temperature(KellyandRose,2002),especiallyinthepresenceof water(Yoshimuraetal.,1987).Thisstructuralchangeofamarten- sitic nature generates volumetric expansion of the monoclinic phase(KellyandRose,2002)thatgeneratescracksandstimulates theirpropagation(SchmauderandSchubert,1986),thereforeinflu- encingthemechanismofmaterialremoval.

Thereisconcernthat3Y-TZPproductssufferunwantedphase destabilizationunderundersizedconditions,evenwithoutapply- ingstress(Lietal.,2001).Undercircumstancesofexternalloading duringeitherthemanufactureoruseoftheceramicproduct,this destabilizationcanhappen.Inthiscontext,thetransformationcan betriggeredbytemperaturesandstressesfromthemachiningpro- cess(Devilleetal.,2006).Thepresenceofaqueous-basedcutting fluidcanaccentuatethisphenomenonofphasedestabilization.

Themicrostructuralcontrolcapabilityofthiseventmakesthe 3Y-TZP a key material in the evaluation of UP machiningpro- cesses,allowingitsscientificstudybyqualifyingandquantifying themartensitictransformationinducedbychipremoval.

The present work hasas its hypothesis the possibility that the achievement of flat surfaces with nanometric finishing (Ra<100nm) in advanced ceramics without critical defects is possiblebycombiningappropriatemachine-toolkinematicsand grinding pressure, proper cutting tool sharpening, and specific lubricationandcooling.

Toprovethishypothesis,theFiocchiLapGrinder’sdesignwas evolved,aimingatUP facegrindingwithlappingkinematicsof advancedceramics,meetingtheconditionsofremovalandfinish- inginasingleautomatedprocessandequipment.

The 3Y-TZP was chosen as theworkpieces’ material due to its phase transformation under severe removal conditions, for instance,excessivespecificcuttingpressureandhightemperatures inahumidatmosphere.Grindingwheelsmadeofsiliconcarbide (SiC)gritsandepoxybondwereappliedunderdifferentUddressing parametersandwaterasadressingfluid.

2. Experimentalset-up 2.1. Machine-tool

AsecondversionoftheFiocchiLapGrinderwasusedinthis work,namedFiocchiCNC LapGrinderII.Themachine-toolwas conceptualizedanddesignedinapartnershipbetweentheLabo- ratoryofTribologyandComposite(LTC)atEscoladeEngenharia deSãoCarlos(EESC)ofUniversidadedeSãoPaulo(USP),SãoCar- losCampus,andtheLaboratoryofMachiningTechnology(LATUS) atFaculdadedeEngenhariadeBauru(FEB)ofUniversidadeEstad- ualPaulista(UNESP),Bauru campus.Themachine-tool wasput togetherandvalidatedatLATUS.Aswellasbeingappliedtobuild thefirstFiocchi’smachine-toolduringhisMaster’scourse(Fiocchi, 2010),designmethodology(Pahletal.,2007)wasusedtoevolve LapGrinder I,aimingtheultra-precision (UP)manufacturingof advanced ceramicsbyimprovingcriticalaspects pointedout in earlystudiesofSanchezetal.(2011),Fiocchi(2010),andFiocchi etal.(2015).

Fig. 1 compares some of the modifications implemented in FiocchiLapGrinders.ThemainimprovementsfromLapGrinder I(Fig.1a) (Fiocchi, 2010; Fiocchiet al.,2011)toLap GrinderII (Fig.1b)(Fiocchi,2014)were:(1)substitutionoftheflexibleshaft byaservomotortocontrolaworkpiece’sangularvelocity;(2)anew spindlehavingtwotaperedrollerbearingsmountedinaback-to- backconfiguration,50Npre-load,andahollowshaftforsupporting avacuumchuck,which hasa universaljointwithtwo degrees- of-freedomtokeepworkpiece’s surfaceparallel tothegrinding wheel;(3)newsteppingmotorsanddrivestopreventlinearaxis resonating;(4)newphysicalconfigurationofthenovelelectrical

(4)

Fig.1.(a)FiocchiLapGrinderI,(b)FiocchiLapGrinderII.

componentsinsidethemainelectricalpanel;and(5)updatedsoft- wareandgrindingstrategytoimprovelinearaxisdynamics.

2.2. Workpiecematerial

Zirconiapowderwith3mol%ofyttria(Y2O3)fromTosoh(TZ- 3Y-E)waschosenastherawceramicmaterialduetoitsattractive surfaceandmechanical(strengthandtoughness)propertiesafter sinteringinthepresentstudy.Moreover,tetragonalzirconiapoli- cristalstabilizedwith3mol%ofyttria(3Y-TZP)isakeymaterial for medical and dental implants as well as the electronic and mechanicalindustries.AccordingtoTosoh,aftersintering,3Y-TZP canachievehardnessof1250HV,elasticmodulus(E)of210GPa, bending strength (␴) of 1200MPa, fracture toughness (KIC) of 5.0MPa.m0.5,and microstructureendowed byfinegrainsunder 0.3␮maveragediameter.

Thematerial alsopresentsinteresting surfacefinishing apti- tudeaswellasthepossibilitythatphasetransformation canbe investigated in outside layers induced by grinding conditions.

Theoccurrenceandmagnitudeofthemartensitictransformation (tetragonal→monoclinic)can,therefore,beusedasascientifictool forevaluatingthequalityofthematerialremovalprocess.

InUPmanufacturingofadvancedmaterials,allproductionsteps areimportantandhaveinfluencenotonlyonthefinalmechanical

andelectricalpropertiesforexample,butalsoonsurfacefinish- ing.Consideringthatfiredadvancedceramicsarehardmaterials that aredifficulttomachine, andtheirmachiningis expensive, an optimized manufacturing routeshould consider a near-net- shapeproduction,inwhichmaterialremovalprocessesareused forimprovingmainlysurfacefinishing.

For thebest machining results, the material properties and geometry must be well-defined and controlled.In the present research,hence,itwaspre-definedthatworkpieces’manufactur- ingrouteshouldstartwithTZ-3Y-Epowderandaimtoaccomplish sinteredflatceramicdiscswith:(1)quasi-zerodefects,(2)homoge- nous grainsize up to 0.4␮m averagediameter, and (3) dense bodies,sothatUd-lapgrindingcouldfocuson(4)achievingsurface roughness(Ra)under100nm,removinglittlematerialusingsili- concarbide(SiC)grindingwheels.Fig.2illustratesthemainsteps andtheirparametersformanufacturingandcharacterizationofthe ceramicmaterialfrompowdertoUd-lapgrounddiscs.

2.2.1. Mixing

The powderwasprepared in a ballmillwitha nylonjar of 100ml internal volume(Carvalhoand Fortulan,2006 Carvalho, 2007)rotating at104rpm,corresponding to65% of thecritical velocity(Vcrit=1338.2/D0.5)for a70-mminnerdiameterjar(D).

30vol%oftheusefulvolumeofthejarwerefilledupwithTZ-3Y-

(5)

Fig.2. MainstepsandcharacteristicsofTZ-3Y-Emanufacturingandcharacterization.

E(=6.05g/cm3),with0.5kgofmillingmedia(TosohYTZ10mm diameter),and70vol%oftheusefulvolumefilledupwithisopropyl alcohol(=0.786g/cm3)asaslurrysolvent.4-Aminobenzoicacid (PABA; =1.37g/cm3) was added in as a deflocculant in the ratioof 0.5%of theTZ-3Y-E mass. Then, theball milloperated for12h.Afterdeflocculatingcompleted,Polyvinyl-butyral(PVB- butvarB98;=1.1g/cm3)wasaddedtotheslurryasabinderin theratioof2.0%oftheTZ-3Y-Emass,andthemixingcontinuedfor 2morehoursforhomogenizingthecolloidalsuspension.

2.2.2. Drying

Afterthedeagglomerationandmixingsteps,dryingwasmanu- allydonebyagitatingtheslurryinahotaircurrentinsideaclean chamber.Gloves withouttalcwereusedand extremecarewas takeninordertoavoidanycontamination.

2.2.3. Granulationandselection

Manualgranulationandselectionappliedan80-mesh(180-␮m aperture)stainlesssteelsievewiththepurposeofstandardizingthe agglomeratesizeinpreparationforpressing.Granulationandselec- tionwerealsoperformedinsidethecleanchamberwhilewearing gloves.

2.2.4. Forming

Formingtookplaceintwosteps;13.6gofthezirconiagranules weremanuallyinsertedintothecylindricalmoldwiththehelpofa plasticcontainer,leveledwithastainlesssteelspatula,andgently agitatedbyavibratingtableLabortechnikGmbH(THYR2model), seekinghomogenizationofthefilling.Next,single-actinguniaxial presshappenedat83.7MPaduring10sina37-mmdiametermold toshapediscs with4.72±0.03-mmthickness.Subsequentwet- bagisostaticpressingoccurredinanAIPwet-bagpresser(CP360 model)at200MPafor30stohomogenizethegreenceramicbodies, reducingdensitygradientsandincreasingdensification.

2.2.5. Sintering

SinteringoccurredinanambientatmosphereLindbergfurnace (BlueMmodel)with3.89C/minheatingandcoolingratesand2h at1400C,resultinginceramicdiscsof3.77±0.02-mmthickness and29.8±0.15-mmdiameter.

2.3. Ceramiccharacterization

AccordingtoFig.3,severalcharacterizationtechniqueswere applied in this research. Apparent density and porosity were determined by ASTM C73. TZ-3Y-E powder, granules, and greenandsinteredworkpieces’microstructurewereinvestigated by SEM-FEG. Green, sintered, and Ud-lap ground workpieces wereinvestigated by confocal microscopy. Opticalfluorescence microscopywasappliedtothepowder,greenbody, andUd-lap groundworkpieces.HardnessVickersweremeasuredinsintered pieces.XRDwasmeasuredinpowder,sinteredandUd-lapground workpieces.Ramanspectroscopyandprofilometrywereusedwith sinteredandUd-lapgroundpartsaswellasopticalmicroscopyin grindingwheels’activesurfaceafterdressingandUd-lapgrinding.

SEMappliedanFEIelectronicmicroscopy(InspectF50model) equippedwitha fieldemission gun(FEG), 10kVelectronaccel- erationvoltage,workingdistancebetween6.7and10.3mm,and secondaryelectrondetection(SE).AboveeachSEMimagethereis alegendwiththemaininformationandconditionsapplied,from thelefttorightside:dateandtime,dwelltime,highvoltage(HV), horizontalfieldwidth(HFW),workingdistance(WD),detection mode,andscale.Thespecimensreceiveda thinlayerofgoldby physicalvapordeposition(PVD).

ANikonmicroscope(modelEclipseTS100)andNISElements softwarewereusedtoanalyzeopticalfluorescence.Powdervisu- alizationapplied10×lensesandsinteredpieceswith4×objective and10×ocularlenses.

XRDwasstudiedthroughaPANalyticalDiffractometer(X’Pert ProMRDmodel)ofCoK␣radiation(1.78Å,40kV,and40mA),con- tinuedscanning,0.02angularresolution,6.667×103degrees,and 2rangingfrom20to65.Quantificationofthevolumepercentage ofthemonoclinicphasewasaccomplishedaccordingtothemodels describedbyTorayaetal.(1984a,b)andSatoetal.(2008).

VickershardnesswasevaluatedbyaMitutoyoMicrohardness tester(HM-200model)withpyramidaldiamondindentatorwith 1kgloadfor10s,andequalloadingandunloadingtimeof4sin sinteredworkpieces.

RamanspectrummeasurementsweredoneinaWitecRaman ConfocalMicroscope(Alpha300A/Rmodel),WitecControlsoft- ware,He-Nelaser(632nm),andUHTSspectrumanalyzer.Different pointsofeachworkpiecesurfacewerefocusedonwitha 1-␮m diameterlaserbeamduring3sofintegratingtimeandtheaverage resultof10integrationswastaken.Quantificationofthevolume

(6)

Fig.3. Workpiecefixturingtechnique.

percentageof monoclinicphase wasaccomplishedaccordingto themodelsdescribedbyKatagirietal.(1988),Limetal.(1992), Kimetal.(1997),Casellasetal.(2001),andTabaresandAnglada (2010).

Roughness(Ra,Rt)andflatnessdeviationofsinteredworkpieces weremeasuredbyaForm TalysurfIntrai60contactprofileme- terwitha2-␮mstylustipradius(rtip),TaylorHobson␮ltraand Talymap Gold software. Dueto green bodies’ fragilityand low strength,anon-contactconfocalLeicamicroscope(DCM3Dmodel) wasusedtomeasuresurfaceandroughnessparameters,controlled byLeicaScan softwareandthedataanalyzed byTalymapGold software.Inputparameterswerea0.25-mmcutoff(␭c),0.25-mm singlemeasuredlength(lr),1.25-mmtotalmeasuredlength(ln), andGaussianfiltertodistinguishroughness,waviness,andshape erroraccordingtoISO25,178andISO4287.

Indirectmaterialremovalrate(MRR)quantificationofthework- pieceswasusedbymeansofananalyticalShimadzuscale(model AUW-D)aftersintering(initialmass,mi)andUd-lapgrinding(final mass,mf).MRRwascalculatedas(mi−mf)/time.

2.4. Ud-lapgrinding

For thedressingoperation, brand-newsingle-point diamond dressers,witha60taperand0.5-mmnoseradiusgeometry(rp), wereusedforeachtest.Thedresserwaskeptperpendiculartothe grindingwheel.

ForUd-lapgrindingexperiments,threedifferentconventional grindingwheelsmadeofdifferentsiliconcarbidegritsizes(#300,

#600,and#800)andhotpressedwithanepoxybondof350-mm diameterand25-mmthicknesswereused.

Filteredwateratan8.33×105m3/sflowratewasappliedby floodduringdressing.A1:40emulsionofsemi-syntheticsoluble oil(RocolUltracut370)wassprayedusingtheminimumquantity fluid(MQF)methodat2.78×107m3/sduringUd-lapgrinding.The locationofthecuttingfluidnozzleswasthesameasthatusedby Fiocchi(2010,2014),Fiocchietal.(2011,2015),andFiocchiand Sanchez(2011).Ahighercuttingfluidflowratewasnecessarydur- ingdressingtoremovetheslurryformedonthegrindingwheel, thuscleaningitupbeforeUd-lapgrinding.

Ud-lapgrindingtestswereconductedtoanalyzetheinfluences ofthreeUdsandthreeSiCabrasivesizesonsurfacefinishingof3Y- TZPspecimens(3×32experiments).Themainparameterswere:

Ud=1,3,and5;#300,#600,and#800SiC;0.1-mmdepthofdress- ing(ad);100-kPalapgrindingpressure(P);and100-rpmworkpiece rotationalspeed (nw)andgrindingwheelrotational speed (ng).

Thesefinalparameters,summarizedinTable1,weredefinedaftera longseriesofpretestsaimingtoproducethebestsurfacefinishing.

TheUd-lapgrindingwasinterruptedat5,10,15,20,25,30,40, and60minand theworkpieces wereremovedfromtheholder, cleanedwithacetoneinanultra-soundbath,dried,weighedfor MRRquantification,andtheirsurfaceroughness(Ra,Rt)andflat- nessdeviationweremeasured.

Bothdressing and Ud-lapgrindingstepsweremonitoredby measuringthemainmotorelectricpower(currentsensor),apply- inga Ciber PowerAnalyzer(CVM NGR96model),described in Fiocchi(2010),whichisintegratedintothemachine-tool.

ANikonandaCarlZeissstereomicroscope(SMZ800andCitoval 2models,respectively),bothlinkedtoaSamtekdigitalcamera(STK 3520model),aswellasPixelViewsoftware,wereusedforgrinding wheelopticalmicroscopy.

The clamping technique was brought from the electronic ceramic industry by melting of Thermoplastic Quartz Cement (model70CfromLakeside)onaglasssubstrateetchedbyhydroflu- oricacidtoenhanceadhesion.Oneworkpieceattime wasfirst gluedontothecenterofthediscusingatemplatetocentralize theworkpieceontheglass.AQUIMIShotplateat100C(Q.261.2 model)wasusedtomelttheglue.Duringcoolingoftheglue,asec- onddiscanda2-kgmasswereplacedovertheworkpieceinorder tohomogeneouslydistributethemoltenglueandkeepthework- piece’ssurfaceparalleltotheglass.Next,thesmoothfaceofthe discwasputincontactwiththevacuumchuck,thevacuumpump turnedon,andthevacuumvalveclosedtoholdthediscinposition onthechuck.Fig.3showstheworkpiecefixturesequence.

A2400-Wvacuumcleanerwasusedtocleanupthegrinding wheel’s surface duringdressingand Ud-lapgrindingoperations inordertoavoidanylooseparticlesonthegrindingwheelthat mayhavedamagedthefinishingand/ortheworkpiece’ssurface integrity.Therefore,onlytwo-bodyabrasionisexpected.Suction nozzles are strategically located before and after grinding and dressingareasaccordingtoFiocchi(2010,2014)andFiocchietal.

(2015),ascanbeseeninFig.4.

3. Resultsanddiscussion 3.1. LapGrinderII

Fig.4showsthegrindingareaofLapGrinderIIanditskeyele- ments,suchascontrolpanel,servomotor,grindingwheel,vacuum chuckandvalve,aswellascuttingfluidsandsuctionnozzles.

AccordingtoPahletal.(2007),theevolutionofthemachine-tool regardingdesignmethodologyisconsideredanadaptivedesign, since it kept toknownand establishedsolution principlesand adaptedtheembodimenttochangedrequirements.Ontheother hand,theoriginalmachine-toolbasedonamodifieddesignthat confersonitnewfeaturesandcapabilitiesisconsideredadevelop- mentaldesignbyKerala(2002).

Thevacuumchuckandglued workpieceshowedthemselves efficientandpractical,withgoodresistancetocuttingfluidsand grindingforcesinadditiontoeasycleaning.

Thedevelopmentaldesignofthespindlewasnecessarytoclamp theworkpiecebyavacuumandtoaddtwodegrees-of-freedom tothejointconnectingthevacuumchuckandthespindle’sshaft.

Thisarticulationdynamicallycompensatesforanymisalignment, keeping theworkpiece’ssurface alwaysparalleltothegrinding

(7)

Table1

InputandoutputparametersstudiedthroughUd-lapgrinding.

Inputparameters Outputparameters

Dressing Ud-lapgrinding Workpiece Grindingwheel

ad=0.1mm P=100kPa 3Y-TZPTosoh SiliconCarbide(SiC) Workpieceroughness(Ra,Rt) 60taperand0.5mm

dressernoseradius(rp)

ng=100rpm 1400Csintering temperature

Hotpressedepoxybond Workpieceflatnessdeviation(␮m) nw=100rpm nw=100rpm 29.8±0.15mmdiameter Ø350×25mm DressingandUd-lapgrindingpower(W) 8.33×10−5m3/s

filteredwater

2.78×10−7m3/s 1:40

RocolUltracut370

3.77±0.02mmthickness #300 Grindingwheelsurfacemicroscopy

Flood MQF 0.45␮mRa

7.47␮mRt

#600 Workpiece3Dprofilometry

Ud=1,3,and5 60min 29.3±1.7␮mflatness #800 MRR(g/minute)

Fig.4.CNCFiocchiLapGrinderII.

wheel.Thesefeaturesresultedinastifferandmorestablemechan- icalsystem.Theservomotorofferedconstantangularvelocityto thevacuumchuck,andthenewsteppingmotorsanddrivespre- ventedthelinearaxisfromresonating.Thus,alltheirregularities pointedoutwereovercomeandarobustUPmachine-tooldesign wasachievedandvalidated.

3.2. Ud-lapgrinding

3.2.1. Grindingwheeltopography

Figs.5and6showtheopticalmicroscopiesof4×4mmareasof thegrindingwheels’active/functionalsurfacesafterdressingand 60minofUd-lapgrinding,respectively.Udvaluesarearrangedin rowsandabrasivegritsizeincolumns.Darkerregionsrepresent deeperareasofthegrindingwheel.

TheimagesshowedinFig.5aresimilartothosefoundinFiocchi (2010)andFiocchietal.(2011,2015).Themacroeffectwasevident forallUd dressingsappliedonboth#800and #600andonthe

#300grindingwheeldressedwithUd=1.AfterUd-lapgrinding, allgrindingwheels(Fig.6)kepttheirstructuresopenedwithout adhesionofmaterialfromtheworkpiece,asopposedtothatseen inthemachiningofAISI420stainlesssteelinFiocchi(2010)and Fiocchietal.(2011,2015).Thelowerductility,higherhardness, andlowerfrictioncoefficientofthe3Y-TZPaswellasthegrinding conditionsmaysustainthesefindings.

3.2.2. RMSelectricpower

Forthosegrindingprocess,i.e.,peripherallongitudinalgrinding (Fig.6a),inwhichtheactivesurfaceisthecylindricalperipheryof

thegrindingwheel,ataconstantdistancefromitsrotationalcenter, thegrindingpower(Pc)canbecalculatedas:

Pc=Ft×Vc (1) whereFtisthetangentialgrindingforceandVcthetangentialveloc- ityofthetool(Tönshoffetal.,2002).Insuchprocess,forinstance, Ftcanbeaccuratelyobtainedbycommercialpiezo-dynamometers and,hence,Pccanbecalculated.

Ontheotherhand,theabrasiveprocessstudieddoesnothavea constantrelativetangentialvelocityduetotherelativekinematics oftheworkpieceandgrindingwheel(Fig.6b).Thus,everysingle point(i)ofthetool/workpieceinterfacehasitsownVc,i.In this particularfacegrindingwithconstantpressureandlappingkine- maticsprocess,theuseofahalleffectsensorisrecommendedto determinePc,consideringthatadynamometerwouldnotbeable todeterminelocaltangentialgrindingforce(Ft,i).

Commercialpiezo-dynamometers providethemost accurate measurementofcuttingforces.Thesetransducershavehighstiff- ness,butsincetheyusepiezoelectricceramics,themeasurementof staticforcesoveralongperiodoftimemayresultinsignificantdrift (Byrneetal.,1995),andbecauseoftheirdifficultiesforinstallation intheforceloop,theapplicationofpiezo-dynametersisrestricted (Jemielniak,1999).Forcetransducersarealsosubjectedtodynamic influencesfromthemachine/tool/workpiecesystem(Oliveiraand Valente,2004).

AccordingtoOliveiraandValente(2004),currentsensorsfor powermeasurementarethesimplestandlowest-costalternative forprocessmonitoring;inaddition,theirinstallationissimple,with

(8)

Fig.5. Opticalmicroscopyofthegrindingwheel’sactivesurfacesafterdressing.(ad=0.1mm,0.5-mmnoseradius,ng=100rpm,andfilteredwaterappliedbyflood).

Fig.6. Opticalmicroscopyoftheworngrindingwheel’sactivesurfacesafter60minofUd-lapgrinding.(P=100kPa,ng=nw=100rpm,andemulsionappliedbyMQF).

noinfluenceonthesystemstiffness.Thepowerisusuallyobtained bymeasurementofthevoltageandelectriccurrentatthegrind- ingwheelspindlemotor(MalkinandKoren,1980).Byrne etal.

(1995),Jemielniak(1999),and OliveiraandValente(2004)con- tendthatthismethod(halleffect/currentsensor)ispreciseandit haslowintrusiveness,but itisdampeddue tothesysteminer- tia,i.e.,grindingwheel,andthesignalprocessingcharacteristics needed.Thisleadstoadelayedresponsethatisnotimportantfor

thisresearch,sincetheworkisnotfocusedonshortevents.Thus onlytherelativemagnitudeofUd-lapgrindingpoweramongthe grindingconditionsisimportantinordertorevealtheinfluenceof Udandgrindingwheelgritsizesalong60minofUd-lapgrinding.

3.2.3. Dressingpower

The updated machine-tool permits deactivation of the workpiece-holder rotation during dressing. In the older ver-

(9)

Fig.7.(a)Discretizationofforcecomponentsinperipherallongitudinalgrinding(Tönshoffetal.,2002).(b)GeometricalmodelforUd-lapgrinding.

sion this was not possible due to the permanent mechanical connection between the workpiece-holder and the main shaft throughtheflexibleshaft(Fiocchi,2010;Fiocchietal.,2011).In suchcase, LapGrinder IIdemands less energy by stoppingthe servomotorand,consequently,thevacuumchuckspindle,during dressing.

Asforthereductionintheamountofenergyrequiredfordress- ing,somecorrelationwasexpectedbetweenthepresentandearlier results(Fiocchi et al., 2015) regarding dressing power (Fig.8).

Althoughthecurrentresearchhasusedpreciselythesamegrinding wheels,thesamedresser’sgeometry,andexactlythesamedress- ingparametersappliedbyFiocchietal.(2015),therewasachange intheRMSdressing’selectricpower.The#600and#800grind- ingwheelsintegritiesmayhavebeendamagedbytheageofthe toolspurchasedin2004.Asofteningofthesegrindingwheels,by looseningabrasivesparticlesmoreeasily,wasnoticed;thiscaused theirmacrogeometriestodeterioratemorerapidly,especiallyfor the#800grindingwheel.

3.2.4. Ud-lapgrindingpower

Fig. 9 indicates that, on average, the higher the Ud value, thelowertheRMSUd-lapgrindingpower: 176.71W forUd=1, 173.91WforUd=3,and172.34WforUd=5.Thisgraphicsuggests, therefore,that,onaverage,higherUdshaveatendencytodemand lesspowerinzirconiaUd-lapgrindingusingSiCgrindingwheels.

Thedifferencebetweentheoreticalandrealmacroandmicro effectsoverthegrindingwheelandtheweakeningoftheepoxy bondcanexplainthosevariances.Fig.10illustratessomeimportant aspectschangedingrindingwheeltopographybyasingle-point diamonddressing,suchasthenumberofactiveabrasives(cutting edgesinteractingwiththeworkpiece),aswellasthewedgeangle (␤),rakeangle(␥),reliefangle(␣),andmacroporosity.

Suchanalysisisbynomeansatrivialtask.Itdependsonabra- sivesizeandfriabilityaswellasthe“health”ofthebondholding theseabrasiveparticlesandagglomeratesinplace,whichinareal dressingsituationinvolvediversemacroandmicroeffects,besides structuraldamageingrindingwheel,whichcanbeextremelyhard topredict,quantify,andqualify.Thepredominanceofoneormore effects canthusjustify thedispersal ofUd-lapgrindingpowers depictedinFig.9.

3.2.5. Materialremovalrate

A very important consequence of applying different Uds is showninFig.11,whichpresentsdifferentconditionsofmaterial removalasafunctionofallUd-lapgrindingsituations.Therange

oftheresultsusingthesamegrindingwheelreinforcesthecapa- bilityoftheUddressingtomodifygrindingwheels’aggressiveness.

ThelowestMRRwas1×106g/minuteforUd=1and#800.The highest,19timesbigger,happenedwithUd=5and#300.

3.3. 3Y-TZPcharacterization

Unfortunately,itwasnotpossibletostudy/focusonexactlythe sameregionsduetolackofacommonreferencesystemamong theconfocalmicroscopy,contactprofilometer,SEM-FEG,XRD,and Ramanspectrometer.Allanalysishappenedinthesamevicinity, whichalsorepresentsthecharacteristicsoftheentiresurface.

3.3.1. 3Y-TZPpreparation

TwodifferentmagnificationsofSEM-FEG ofTZ-3Y-Epowder areshowninFig.12aandb.Therawmaterialpresentedthepre- dominanceofaroundedagglomerateof0.06±0.03-mm(Fig.12a) andparticlesizeof0.08±0.02-␮maveragediameter(Fig.12b).

Someagglomerateshapeswereobtainedafterdrying(Fig.12c)and granulation/selection(Fig.12d).Fig.12epresentsthecompacted ceramicat200MPa,representingthenewbiggeragglomeratesand theresidual interagglomerateporosity as described byGerman (1987).Fig.12fdemonstrates zirconia’s 0.35±0.07-␮maverage grainsizeachievedby2hofsinteringat1400C.Table2displays thefinalzirconia’sproperties.

3.3.2. Profilometry

Uniaxialpressinginducteddensitygradientsinthegreenbodies, compactedaswellexplainedinliterature.Thosegradientsgen- eratedplasticdeformationsand,therefore,formerrorinthethin ceramicdiscs.Aftertheisostaticpressingstepthosedensitygradi- entswereconsiderablyreduced,buttheformerrorremainedand producedconcaveandconvexoppositesurfacesinthesamework- piece.Theformerrorincreasedordecreaseddependingontheside oftheworkpieceexposedtothehighertemperatureinsidethefur- naceduringsintering.Anamplificationofformerrorwasnoticed whentheconcavesurfaceswereexposedtohighertemperatures.

Ontheotherhand,betterformcorrectionhappenedwhenconvex surfacesweresubjectedtohighertemperatures.Alsoobservedwas animportantinfluenceofgravityaccelerationduringsinteringas consequenceofdeformationofthinceramicdiscsthathadrested onirregularsurfaces.

IntheuniverseofproducingnearnetshapesinUPceramicman- ufacturing,thefollowingshouldbeconsidered:thetypeofpressing anditsassociations;theadoptionofgreenceramicmachiningto

(10)

Fig.8.RMSdressingelectricpowerasfunctionofoverlapfactorandSiCgritsize.(ad=0.1mm,0.5-mmnoseradius,ng=100rpm,andfilteredwaterappliedbyflood).

Fig.9.RMSUd-lapgrindingelectricpowerasfunctionofoverlapfactorandSiCgritsize.(P=100kPa,ng=nw=100rpm,andemulsionappliedbyMQF).

Fig.10.Schematicrepresentationoftherealandtheoreticalmacrogeometry(dashedlines)ofthegrindingwheelsdependingontheabrasivesizeandoverlapfactor(Ud).

Table2

Finalzirconiadiscspropertiesaftersintering.

TZ-3Y-E

Microhardness(HV) Openedporosity(%) Density(g/cm3) Averagegrainsize(m)

1414.10±339.28 0.002±0.001 6.0±0.06 0.35±0.07

(11)

Fig.11.MRR(g/minute)asfunctionofoverlapfactorandSiCgritsize.(P=100kPa,ng=nw=100rpm,andemulsionappliedbyMQF).

shapethepartandalsoremovethehighest-densitygradientsin outsidelayers;theflatnessofthebase,andthetemperaturedis- tributioninsidethefurnace.Alltheseconsiderationwillminimize theeffectsofheterogeneouspressingandsinteringshrinkage.The judiciouschoiceoftheseparameters,therefore,shallresultinless machiningtimeforcorrectionofform,whichwillfocusmainlyon finishing.

3.3.3. Surfaceroughness

Foralltestedconditions,therewasasharpdiminishinginthe valuesofsurfaceroughnessinthefirst5minandaslightvariation ofsurfacequalityuntiltheendof60minofUd-lapgrinding.Ascan beseeninFig.13,allthreegraphicshadthesamescaling.

Forthe#800dressedwithUd=1and3,theroughnessdecreas- ingrateispracticallythesameandseveraltimeshigherthanfor Ud=5inthefirst5min.Theworstperformancecanbeexplainedby theprematurewearofthe#800grindingwheeldressedwithUd=5, whichlostitsmacroeffectsooner,ascanbenoticedcomparingFigs.

5-7and6-7.

The#600grindingwheeldidnotshowavisiblemacroeffect forUd=3and5(Fig.5-5and5-8).However,themicroeffectmay havepositivelyinfluencedsurfaceroughnessbyofferingnewcut- tingedgeswithdifferentanglesandlowerdepthofcutaswell assufficientgrindingwheelporositytocarrydebrisandcutting fluid,thereforeshowingthattheUd throughasingle-pointdia- monddressingchangesgrindingwheelaggressiveness.Fig.6-5and 6-8showsimilartopographiesandmayexplainthesimilarityin roughnessproducedbythe#600dressedwithUd=1and5during thelast20min.

TheinfluenceofUdwasmoreevidentoverthe#300grinding wheel.Duringthefirst5minallUdsproducedquitesimilarrough- nessdecreasingrates.ThebestroughnesscamefromUd=5andthe worstonefromUd=3.

Figs.14and15presentRa(nm)andRt(␮m)roughnessparam- etersvaluesattheendof60minofUd-lapgrinding,respectively.

TheideabehindFig.10canjustifytheroughnessdepreciation withthesubstantialincreaseofdeeperrisksproducedbythe#600 and#800grindingwheels;higherRtvaluesareshowninFig.15.

TZ-3Y-EisharderthantheAISI420stainsteelstudiedbyFiocchi etal.(2015)andthusnomaterialisimpregnatedintothegrinding wheel’ssurface,resultinginhigherabrasiveprotrusionthatmay depreciatefinishing,butkeepsthetoolcleanerandwithopened porosity.Ontheotherhand,theconstantabrasiveexposureassoci- atedwiththehigherhardnessofzirconiawearsthegrindingwheel easily.

3.3.4. Topography

Fig.16and17illustrateconfocalmicroscopyofagreenanda firedpiece,respectively.Surfacefinishingimproved36%inarith- metic mean height (Sa) from green (0.71␮m) to the sintered (0.52␮m)stage,aswellasproducingflatterandsmootherareas.

Fig. 18 represents workpieces’ surface obtainedby confocal microscopyafter60minofUd-lapgrinding.Inthisfigurecanbe seenapredominanceofductilecutting,absenceofcracks,andpat- ternofaleatoricscratches.Atleasttwolevelsofscratchingcanbe visualized:oneseverewithlargerwidthanddepth,andanother gentlerwithshallowmarks.Inbothcasesthemicrocuttingseemsto bethemajorductileremovalmechanisminallconditionsstudied.

Insomeworkpiecestherearestillremnantsofareasfromthe sinteredgenitor surface,suchas thepresenceof openedpores, depressionswithsphericalgeometry,andnon-machinedareas.In thecontextofthiswork,thosezonesareclassifiedasirregularities andmightberemovedbyincreasingtheUd-lapgrindingtime,since thestudiedcuttingconditionsprovidedlowMRR.

TheconditionofUd=1andthe#300grindingwheelraiseda suspicionthatbesidesductilecutting,therewasalsomicroplowing andplasticdisplacementofmaterial,asdemonstratedinFig.19.

The white arrows point to microcutting marks. The white dashedcirclesdelimitareasofplasticdeformationattheintersec- tionofthescratches,whichrevealsthechronologicalsequenceof events.Thenewestscratcheshaveovershadowedthepredecessor ones.

Theblackarrowspointoutpossiblemicrocracks.However,SEM provedthatthoseregionscamefromplasticdislocationofmaterial towardthecenterofthescratch.Anotherfactthatsupportsthis thesisisthatthewhiteleftarrowpointstoadeeperscratchwithout cracking.Howcouldalessaggressivecondition,pointedoutbythe blackarrows,generatesuchfailure?

Thepresenceofplasticdeformationputforwardtheideathat strainhardeningmaybetakingplace.Insuchcase,bythetimeits limitisreached,thematerialsuffersbrittlefracture.Thisdiscussion willberesumedbySEManalysis.

3.3.5. Flatnessdeviation

Fig.20displaysawidevarietyofflatnessdeviationandrein- forcestheability thatthesingle-pointdiamond dressinghasto modifygrindingwheels’characteristicsinUd-lapgrinding.

3.3.6. SEM-FEG

ThequalityoftheimagesacquiredbytheSEM-FEGismuchbet- terthanthoseobtainedbyconfocalmicroscopy;however,theSEM employedwasnotcapableofmeasuringroughness.Figs.18and21

(12)

Fig.12.SEM-FEGofTosohTZ-3Y-E:(aandb)rawmaterial;(c)driedand(d)granulated/selectedby#80(180␮m)sieve;(e)isostaticallypressedat200MPa;(f)sinteredby 2hat1400Candaheating/coolingrateof3.89C/minute.

representthesamesurfacesshownbydifferenttechniques.Such comparisonisimportanttodisplaythedifferencesininterpretation ofthesurfacesdependingonthemicroscopytechniqueused.

The confocal and electron microscopy images demonstrate removal mechanismssuch as microcutting,microplowing, and microgrooving. During both microscopy procedures the work- pieceshadalloftheirsurfacesscanned,andnocrackintroduced byUd-lapgrindingwasdetected.

Themicrocuttingwasmoreactiveatthebeginningofthepro- cess,whentheabrasiveswerestillsharp.Theothermechanisms happenedpracticallyduringtheentiremachiningtime,becoming moreevidentwiththewearofthecuttingedgesthatincreasedthe tip’sradius,exactlyasdescribedbyMarsh(1964).

Fig.22ashowsthemicroplowing,whichplasticallydislocates materialtotheborder ofthefurrow.Themicrogroovingcanbe seen inFig.22b;zirconia’sgrains aresmashed, minimizing the grainboundariesbymeansoflongitudinalplasticdeformationof theceramicmaterial.

Insomecircumstancestheabrasivepenetrationintothework- pieceprovidedacuttingcrosssectionthatpromotedductilecutting withlowresidualplasticdeformation.Thisspecificcuttingcondi- tionsdidnotraisematerialattheborderofthescratchandleft parallelabrasivemarksinthecuttingdirectionatthebottomof thescratch.Fig.22cillustratesthismicrocuttingmechanism.

According to Schinker and Döll (1987), microplowing and microgrooving mechanisms promote not only lateral deforma- tion,butalsothecompressionofthematerialinthedirectionof

(13)

Fig.13.Roughness(Ra)versustimefordifferentSiCsizesandUds.(P=100kPa,ng=nw=100rpm,andemulsionappliedbyMQF).

Fig.14. Finalroughness(Ra)after10minofUd-lapgrindingfordifferentSiCgritsizesandUds.(P=100kPa,ng=nw=100rpm,andemulsionappliedbyMQF).

Fig.15.Finalroughness(Rt)after10minofUd-lapgrindingfordifferentSiCgritsizesandUds.(P=100kPa,ng=nw=100rpm,andemulsionappliedbyMQF).

machining.Byincreasingdepthofcutthemicrocuttingbecomes predominantagain.Thiswasoneoftheexplanationsfoundtojus- tifythebestroughnessproducedbythelargestabrasivegrit(#300 SiC),whichalsohashigherfriability.

TheSEMimagesshowthatUd-lapgrindinghasachievedplas- ticityconditionsof 3Y-TZP,which leadus awaytobelieve that thenanometricsurfacefinishingwasproducedmainlybyductile

microcuttingandpulverizationremovalmechanisms.Inthefirst mechanism,thevolumeofremovedmaterialisequaltothescratch volumeleft.Itisbelievedthatthesecondmechanismhappensdue tosuccessiveplasticdeformationaccumulatedby3Y-TZPindis- tinctdirections(defects)thathasachievedacriticalstrainstate, inwhichmicrocrackswerenotonlynucleated,butalsoformeda recrystallizedlayerofnanometricthicknessaccordingtoTabares

(14)

Fig.16.Confocalmicroscopy(LeicaDCM3D)ofagreenpiece.20×objective,darkfield,andmonochromaticlight.

Fig.17.Confocalmicroscopy(LeicaDCM3D)ofasinteredpiece.100×objective,brightfield,andwhitelight.

etal.(2011).Thisspecificremovalconditioncanbeassociatedwith metalhardeningasaconsequenceofcoldworking(strainharden- ing),whichproducesverysmallfragmentsofmaterial(debris),as wellillustratedinFig.22d.Itisalsobelievedthatthelastremoval mechanism is a consequenceof low-cycle fatigue,producing a smoothandpracticallynotscratchedsurface,asshowninsome areasofFig.22c.

3.3.7. Ramanspectroscopy

Fig.23presentsRamanspectraoftheworkpiecessubmittedto thenineUd-lapgrindingconditions.Thisfigurealsocontainsthe spectraofaworkpieceassinteredaswellasacompactedmono- clinicphasezirconia.Verticalblackandredlinesaresuperimposed onthegraphicinordertohelpdifferentiatetetragonal(t)andmon- oclinic(m)Ramanintensities.

Thevolumepercentageofthemonoclinicphasewaslowerthan 1%,accordingtoTable3,whichpresentsdifferentapproachesfor determiningthemartensitictransformation.Theresultswerequite similarineachappliedtheory.Consideringthosevalues,themon- oclinicphasemaybenegligible.

3.3.8. XRD

Fig.24presentsx-raydiffractionpatternsfor(1)rawmaterial (TosohTZ-3Y-E)assupplied,(2)sintered3Y-TZP,(3)sintered3Y-

TZPafter60minofUd-lapgrindingusing#300SiCdressedwith Ud=5,P=100kPa, ng=nw=100rpm, and 1:40emulsion applied byMQF,and (4)compacted monoclinicpowderwithoutbinder pressedat80MPa.

Therawmaterialhastracesofthemonoclinicphase,asindicated bythepurplearrowsinthevicinityof2␪=33and36.5(Fig.24-1).

Sinteringresultedinafullytetragonalmaterial(Fig.24-2).

Fig.24showstheXRDpatternsforallnineUd-lapgrindingcon- ditionsstudied.

The#600and#800grindingwheelspracticallydidnotalter thespectrainrelationtothesinteredworkpiece,onlytenuous, asymmetric,leftbroadeningaround35.Thesingularitywasmore pronouncedusingthe#600grindingwheel.

Two characteristic phenomena arising from Ud-lap grinding usingthelargestabrasivegritwereobserved.Theseweretheasym- metric broadening of tetragonalpeaks near35 and 59 (black arrows),andreversaloftheintensityofthetetragonalpeaksat40 and41(Figs.24-3and25).VirkarandMatsumoto(1986),Mehta etal.(1990),andTabaresetal.(2011)affirmthatreversaloccursdue toferroelasticdomainswitching,whichcanbeexplainedasareori- entationofthecrystallographicplanesatthelevelofcrystallite.The asymmetricbroadeningmayalsoindicateasuperficialgrainrefine- ment.ThisfindingisinagreementwithTabaresetal.(2011)that

(15)

Fig.18. Confocalmicroscopy(LeicaDCM3D)after60minofUd-lapgrindingwith9differentconditions.100×objective,brightfield,andwhitelight.

Fig.19.Confocalmicroscopy(LeicaDCM3D)ofasurfaceproducedbya#300SiCgrindingwheeldressedwithUd=1.100xobjective,brightfield,andwhitelight.

Fig.20.Flatnessdeviationafter60minofUd-lapgrinding,asafunctionoftheoverlapfactorfordifferentSiCgritsizes(P=100kPa,ng=nw=100rpm,andemulsionapplied byMQF).

(16)

Fig.21.SEM-FEGafter60minofUd-lapgrindingwith9differentconditions.

Fig.22.SEM-FEGof(a)microplowing,(b)microgrooving,(c)microcutting/scratching,and(d)chipsproducedbyUd-lapgrinding.

(17)

Fig.23.RamanintensityversusRamanshift(cm−1)ofthenineUd-lapgrindingconditionsappliedin3Y-TZP,3Y-TZPassintered,andacompactedmonoclinicphasezirconia.

VerticalblackandredlinesaddedtothegraphictoindicateRamanshiftoftetragonalandmonoclinicphases,respectively,accordingtoPezzottiandPorporati(2004).

Table3

Volumepercentageofmonoclinicphase(Vm)measuredbyRamanspectroscopyinUd-lapgroundworkpieces.

Author/equation Parameter Grindingwheelmesh Overlapfactor(Ud) Vm(%) Vmmax.[%] Vmmin.[%]

k

Clarke and Adar (1982)

0.97 1 800 1 0.50

600 1 0.47

300 1 0.50

800 3 0.50

600 3 0.50

300 3 0.50

800 5 0.50

600 5 0.50

300 5 0.50

Katagiri etal.

(1988)

0 2.2±0.2 800 1 0.48 0.50 0.46

600 1 0.49 0.51 0.47

300 1 0.48 0.50 0.46

800 3 0.48 0.50 0.46

600 3 0.48 0.50 0.46

300 3 0.48 0.50 0.46

800 5 0.48 0.50 0.46

600 5 0.48 0.50 0.46

300 5 0.48 0.50 0.46

(1992) 1 0.33±0.33 800 1 0.75 1.0 0.60

600 1 0.73 1.0 0.57

300 1 0.75 1.0 0.60

800 3 0.75 1.0 0.60

600 3 0.75 1.0 0.60

300 3 0.75 1.0 0.60

800 5 0.75 1.0 0.59

600 5 0.75 1.0 0.60

300 5 0.75 1.0 0.59

(1997) 800 1 0.20

600 1 0.20

300 1 0.20

800 3 0.20

600 3 0.20

300 3 0.20

800 5 0.20

600 5 0.20

300 5 0.20

Casellas etal.

(2001)

800 1 0.53

600 1 0.52

300 1 0.53

800 3 0.53

600 3 0.53

300 3 0.53

800 5 0.53

600 5 0.53

300 5 0.53

(18)

Fig.24. X-raydiffractionpatternsfor(1)TosohTZ-3Y-Epowderassupplied,(2)sintered3Y-TZP(2hat1400C),(3)sintered3Y-TZPafter60minofUd-lapgrindingusing

#300SiCdressedwithUd=5,P=100kPa,ng=nw=100rpm,and1:40emulsionappliedbyMQF,and(4)compactedmonoclinicpowderwithoutbinderpressedat80MPa.

Fig.25.X-raydiffractionpatternsfor(1)compactedmonoclinicpowder,(2)TosohTZ-3Y-Epowderassupplied,(3)sintered3Y-TZP,(4-12)sintered3Y-TZPafter60minof Ud-lapgrindingusingP=100kPa,ng=nw=100rpm,and1:40emulsionappliedbyMQF.

foundarecrystallizedzoneofgrainsof∼20nmdiameteringround 3Y-TZP.

Tabaresetal.(2011)alsoshowedthattheasymmetryobserved inthespectrumofgroundzirconiaisrelatedtotheoverlapping of thetetragonal and rhombohedral phases.The rhombohedral

phase hasbeenassociatedwithdistortion ofthecubic phasein the[111]directionaccordingtomanyauthors,suchasHasegawa (1983),Kitanoetal.(1988),Mitraetal.(1993),andTabaresetal.

(2011).Therefore,theprobabilityoffindingarhombohedralgrainis equaltotheprobabilityoffindingacubicgrain.Ontheotherhand,

Referências

Documentos relacionados

Não é suficiente afirmar que as imagens de Cláudio Assis em Amarelo Manga e Baixio das Bestas querem o desconforto, porque o desconforto ali está sempre associado a uma

The probability of attending school four our group of interest in this region increased by 6.5 percentage points after the expansion of the Bolsa Família program in 2007 and

Ousasse apontar algumas hipóteses para a solução desse problema público a partir do exposto dos autores usados como base para fundamentação teórica, da análise dos dados

didático e resolva as ​listas de exercícios (disponíveis no ​Classroom​) referentes às obras de Carlos Drummond de Andrade, João Guimarães Rosa, Machado de Assis,

The contribution of blood cultures to the clinical management of adult patients admitted to the hospital with community-acquired pneumonia: a prospective observational study. Causes

This study identified an increased length of hospital stay, a higher mortality rate, and an excess of direct costs related to.. Table 2 – Length of stay and exposure to

negative); Lanes 1, 2 and 4 (isolates CAMP test positive); Lane 5 reference strain Rhodococcus equi ATCC 6939; M = DNA size marker (100 bp).. five isolates CAMP test-positive

i) A condutividade da matriz vítrea diminui com o aumento do tempo de tratamento térmico (Fig.. 241 pequena quantidade de cristais existentes na amostra já provoca um efeito