Any correspondence concerning this service should be sent to the repository administrator:
[email protected]
To link to this article:
DOI:10.1016/j.ceramint.2012.03.028
Official URL:
http://dx.doi.org/10.1016/j.ceramint.2012.03.028
This is an author-deposited version published in:
http://oatao.univ-toulouse.fr/
Eprints ID: 8725
To cite this version:
Pavia, Anthony and Laurent, Christophe and Weibel, Alicia and Peigney, Alain
and Chevallier, Geoffroy and Estournès, Claude
Hardness and friction behavior
of bulk CoAl2O4 and Co–Al2O3 composite layers formed during Spark Plasma
Sintering of CoAl2O4 powders. (2012) Ceramics International, vol. 38 (n° 6).
pp. 5209-5217. ISSN 0272-8842
O
pen
A
rchive
T
oulouse
A
rchive
O
uverte (
OATAO
)
OATAO is an open access repository that collects the work of Toulouse researchers and
makes it freely available over the web where possible.
Hardness
and
friction
behavior
of
bulk
CoAl
2
O
4
and
Co–Al
2
O
3
composite
layers
formed
during
Spark
Plasma
Sintering
of
CoAl
2
O
4
powders
Anthony
Pavia,
Christophe
Laurent
*
,
Alicia
Weibel,
Alain
Peigney,
Geoffroy
Chevallier,
Claude
Estourne`s
Universite´ deToulouse,InstitutCarnotCIRIMAT,UPSCNRS,Universite´ Paul-Sabatier,31062Toulousecedex9,France
Abstract
MaterialsmadeupofaCo–Al2O3compositecoatingoveraCoAl2O4corearepreparedduringSparkPlasmaSinteringofCoAl2O4powders.
TheCoparticlesareprecipitatedbecauseofacombinationofhightemperatureandlowO2partialpressure.Theprecipitationanddensification
processeshampereachotherandthusthewaytheuniaxialpressureisappliedduringthesinteringcycleisanimportantparametertocontrolthe microstructureofcompositelayeranditsthickness(about100mm)andobtainadensesample(about4g/cm3).Thefrictioncoefficientofthe
Co-Al2O3compositesagainstanAl2O3ballislowerthanthatfoundforanAl2O3specimen,whichcouldrevealthelubricatingroleofsubmicrometer
Coparticles.However,increasingtheloadfrom5to10Nloadcausesmajorchangesinthefrictioncontact,whicharedetrimental.BulkCoAl2O4
wasfoundtohaveaVickersmicrohardnessabout15.5GPaandanaveragefrictioncoefficientlowerthanthatofanAl2O3sample.
Keywords:B.Composites;C.Friction;D.Spinels;SparkPlasmaSintering
1. Introduction
Metal–Al2O3 ceramic matrix nanocomposites and nano/
microhybridcomposites could beinteresting for tribological applications[1–6],although some authors [7] suggestedthat metal particles may be detrimental and that oxide–Al2O3
compositesmaybemoredesirable.Thepreparationofmetal– Al2O3 composites usually involves firstly the synthesis of a
metal–Al2O3compositepowderandthenitsconsolidationby
hot-pressingorSparkPlasmaSintering(SPS).However,itwas shown that materials with a Fe–Al2O3 or Fe/Cr–Al2O3
composite layer at the surface could be directly prepared by SPS of a powder of a reactive oxide solid solution (a-Al1.86Fe0.14O3 or a-Al2ÿ2x(Fe0.8Cr0.2)2xO3, respectively)
[5,8,9]. The core of the material is made up of the spinel FeAl2O4andAl2O3(andthesurfacemaycontainFeAl2O4too
dependingonthe experimentalconditions).Other authors[7]
alsoreportedtheformationofFe-FeAl2O4-Al2O3
nanocompo-sitesbyagingsinteredsolidsolutionsinN2–H2gasatmosphere.
Thefirstaimofthispaperistostudytheinsituformationof Co–Al2O3coatingsduringSPSofCoAl2O4powders.CoAl2O4,
adefined-compoundasopposedtoasolidsolution,isanormal spinel,i.e.theCo2+ionsarelocatedinthetetrahedralsitesof the cubic close-packing of O2ÿ ions whereas the Al3+ ions occupy the octahedral sites. Moreover, a considerable solid-solution range exists on the Al2O3-rich side of the
stoichio-metric spinels [10,11]. The second aim is to investigate the microhardness and friction behavior of the Co–Al2O3
compositelayers andofthe coreof the specimens, madeup ofbulkCoAl2O4.
2. Materialsandmethods 2.1. Powdersynthesis
Three different CoAl2O4 powders were investigated. The
first one, designated COMinthe following, isacommercial powder (Aldrich 633631-25G, <50nm, 99.9%). The other powders were prepared by combustion synthesis, using the *Correspondingauthorat:Universite´ deToulouse,InstitutCarnotCIRIMAT,
UPSCNRS,Universite´ Paul-Sabatier,118routedeNarbonne,31062Toulouse cedex9,France.Tel.:+33561556122;fax:+33561556163.
E-mailaddress:[email protected](Ch.Laurent).
appropriate amountsof Co(NO3)26H2Oand Al(NO3)39H2O
as theoxidizersandeithercitricacidorureaasthe fuel,ina procedure similar tothat described elsewhere[12–14]. Note thatthecombustionwithureaismuchmoreviolent,reachinga higher temperature than that with citric acid (smouldering combustion).Thecombustionproductsweremanuallyground in an agate mortar andcalcined in air(5008C, 2h of dwell time)inordertooxidizeanypossibleresidualcarboninthe as-preparedpowders,producingpowdersdesignatedUandCAin thefollowing.Thepowdersweredividedintoseveralbatchesas required forthestudy.
2.2. Sparkplasma sintering
ThepowderswereconsolidatedbySPS(DrSinter2080,SPS Syntex Inc., Japan). They were loaded into an 8mm inner diameter graphitedie. Asheetof graphitic paper was placed betweenthepunchandthepowderaswellasbetweenthedie andthepowderforeasyremoval.Thisensembleisknownasthe stack. The powders were sintered in vacuum (residual cell pressure about 5Pa). Apulse configuration of twelve pulses (onepulseduration3.3ms)followedbytwoperiods(6.6ms)of zerocurrentwasused.Anopticalpyrometer,focusedonalittle hole at the surface of the die, was used to measure the temperature.Aheatingrateof3008C/minwasusedfromroom temperatureto7008C,wherea1mindwellwasapplied,and from700to13008C.Adwellof5minwasappliedat13008C. Coolingratewas1008C/min.Notethatforthefirstpartofthe study, the applied uniaxial pressure was kept at aminimum (5MPa),i.e. the contact pressure,during the fullcycle. The sinteredspecimensaredesignatedCOMS1,CAS1andUS1in the following. The sintered specimens are pellets 8mm in diameter and about 2mm thick. The sintering experimental conditions aresummarized inTable 1.
Forthesecondpartofthestudy,adwelltimeofeither3or 9min was applied at 13008C and the maximum uniaxial pressure was increased to 100MPa. It was applied by four different ways,increasinglyearlyinthecycle:duringthelast minuteofthedwell(samplesUS2(3)andUS2(9)),duringthe
firstminuteofthedwell(US3),duringthe firstminuteof the 700–13008C ramp(US4), during the first minuteof the RT-7008Cramp(US5).Thesinteredspecimensarepellets8mmin diameter and about 2mm thick. The sintering experimental conditionsare summarizedinTable1.
Forthelastpartofthestudy,threepellets20mmindiameter werepreparedbySPS,twousingtheUpowderandoneusingthe COMpowder. SpecimenUS6 was consolidatedin conditions similartothatusedforUS2(9).SpecimenUS7andCOMS2were consolidatedinconditionssimilartothatusedforUS4,except that the dwell time was doubled (6min). The sintering experimentalconditionsaresummarizedinTable2.
2.3. Characterization
Thespecificsurfaceareaofthepowders wasmeasuredby theBET method(MicrometricsFlow SorbII2300)using N2
adsorptionatliquidN2temperature.Detection and
identifica-tion of the crystallized phases was performed by X-ray diffraction(XRD,CuKaradiation,BrukerD4Endeavor).The
powders(metalizedwithPt)wereobservedby field-emission-gun scanning electron microscopy (FESEM, JEOL JSM 6700F).
Thedensityofthesinteredspecimenswascalculatedfrom the weight and dimensions after removal of the graphitic surfacesheetbyalightpolishing.Thepellets8mmindiameter werecutintheirmiddlealongthepressingaxisusingadiamond blade.OnehalfwasusedforXRDinvestigationsperformedon thesemi-circularsurfaces,firstontheunpolishedone,thenon samplesgroundeverdeeper,inordertorevealthecrystallized phasespresent at variousdepths into the material.The other halfwasusedasacross-section,whichwaspolishedtoa1mm diamond suspension and was observed by FESEM. For the pellets 20mm in diameter, the top side was only slightly polished in order to reveal the surface composite layer. By contrast,thebottomsidewasgroundinordertorevealthecore ofthe specimen. BothsurfaceswereobservedbyFESEM.A small portion of thesamplewas cut andobservedas across section.
Table1
Specificsurfaceareaofthestartingpowder(Sw);SPSexperimentalconditions:maximumtemperature(T),uniaxialpressure(P),dwelltime(t),density(r); characterizationofthespecimens:compositionofthecore,compositionandthickness(ds)ofthesurfacelayer,sizerange(dCo)oftheCoparticles.sp,spinel,a, a-Al2O3;nm,notmeasured. Specimen Sw (m2/g) T (8C) P (MPa) t (min) r (g/cm3)
Corecomposition Surfacecomposition dS (mm)
dCo (mm)
COMS1 55 1300 5 0 3.2 sp(+Co+a) Co+a Ill-defined 0.03/0.50–1.0
CAS1 78 1300 5 0 2.6 Co+a Co+a no 0.45–0.60 US1 18 1300 5 0 3.4 sp(+Co) Co+a 240 0.80–1.0/2.0–3.5 US2(3) 18 1300 100a 3 3.8 sp(+Co) Co+ a 202 nm US2(9) 18 1300 100a 9 3.5 sp(+Co) Co+ a 436 nm US3 18 1300 100b 3 3.9 sp(+Co) Co+ a 161 nm US4 18 1300 100c 3 4.1 sp(+Co) Co+a 98 nm US5 18 1300 100d 3 4.1 sp(+Co) Co+a 57 nm
aPressureappliedduringthelastminuteofthedwell. b Pressureappliedduringthefirstminuteofthedwell.
cPressureappliedduringthefirstminuteofthe700–13008Cramp.
2.4. Indentationandfriction tests
Indentationtests(10Nfor10sinairatroomtemperature) wereperformedonboththetopandbottompolishedsurfacesof the20mmpellets(i.e.thesurfacecompositelayerandthecore) byloadingwithaVickersindenter(ShimadzuHMV2000).The calculated microhardness (HV) values are the average of 10 measurements.
Dryfriction experiments were performed using a ball-on-reciprocatingflatgeometry.Analuminaball(TCP-C-AA-0063, CSM, Switzerland) 6mm in diameter was used against flat samplessurfaces.Thenormalloadwasfixedat5and10Nand theslidingspeedwasfixedat5cm/s.Thereciprocatingstroke was 20mm and the test was performed for 500 cycles. The frictional force transferred to a load cell was recorded throughoutthetests.
3. Results anddiscussion 3.1. Powders
Onlythepeakstypicalof CoAl2O4arepresentattheXRD
patternsof thethreepowders(Fig. 1).Thecrystallite sizewas evaluatedbyapplyingScherrer’sequationonthe(220),(311) and(440)peaks.Theobtainedvalues,aftersubtractionofthe instrumentalbroadeningobtained byroutinecalibrationof an aluminasample,aresimilarandwereaveraged.Thecrystallite size is equal to 347nm, 213nm and 324nm for powdersCOM,CAandU,respectively.Thespecificsurfacearea isequalto55,78and18m2/gforpowdersCOM, CAandU, respectively(Table1).FESEMobservationsrevealthatpowder COM(Fig.2aandb)ismadeupoflooseaggregatesabout10– 70mminsize,consistingof25nmprimarygrains.PowderCA (Fig.2candd)ismadeupofslightlyporousgrainsbelow40mm insize,consistingoffineprimarygrains(<25nm).Bycontrast, forpowderU(Fig. 2eandf),thegrains aredense, formedof sinteredprimarygrains(ca.50nm)withsomelargepores.These observations are in reasonable agreement with the XRD and specificsurfaceareadataandwithearlierworks[10–12].
3.2. Sinteredspecimens
Thedensityisequalto3.2,2.6and3.4g/cm3forCOMS1, CAS1andUS1,respectively(r–Table1).TheXRDpatterns
(Fig. 3a) of the surface of all three specimens show the Co (111)peakandotherpeaksaccountingfora-Al2O3.Nospinel
peaksaredetected.Thisrevealstheformationofacomposite layer at the surface. The three specimens were ground to remove this layer and expose the core of the samples. The correspondingXRDpatternsareshowninFig.3b.ForCOMS1, the spinelpeaks are detected alongwith the Co (111)peak (weak)anda-Al2O3peaks(veryweak).ForCAS1,peaksofCo
and a-Al2O3 onlyare detected, thus similarly to the surface
XRDpatterns.Bycontrast,onlyspinelpeaksandaveryweak (111)CopeakaredetectedforUS1core.Thecross-sectionsof the specimens were observed by FESEM (Fig. 4). The presented resultsare for the top sideof the specimens, close totheupperpunch,butitwasverifiedthatthesameresultsare obtained close to the bottom punch. In the FESEM images (back-scatteredelectronimagesinchemicalcontrastmode),the Coparticlesappearaswhitedots,thespinelphaseaslight-gray grains and a-Al2O3 as dark-gray grains (although the latter
compounds are difficult todistinguishfrom each other).The average diameter of the Co particles (dCo – Table 1) was
evaluated by measuring the diameter of about one hundred
Fig.1. XRDpatternsofthedifferentCoAl2O4powders:COM,CAandU. Table2
Uniaxialpressure(P),dwelltime(t),density(r),thickness(ds)ofthesurfacecompositelayer,size(dCo)oftheCoparticlespopulation(s),fractionofsurfacearea occupiedbyCoparticles(SCo),Vickersmicrohardnessofthecore(HVcore)andsurface(HVsurf),averagefrictioncoefficientofthecore(mcore)andsurface(msurf)for appliedloadsof5and10N,forthespecimenspreparedbySPS(maximumtemperature13008C).
Specimen P (MPa) t (min) r (g/cm3) dS (mm) dCo (mm) SCo (%) HVcore (GPa) HVsurf (GPa) mcore msurf 5N 10N 5N 10N US6 100a 9 3.5 420 0.3/3.0 23 15.5 9.6 0.28 0.31 0.27 0.51 US7 100b 6 3.9 145 0.3/1.5 24 14.9 8.3 0.30 0.30 0.61 0.58 COMS2 100b 6 4.1 65 0.3–1.0 17 15.9 13.6 0.31 0.34 0.23 0.31
aPressureappliedduringthelastminuteofthedwell. b Pressureappliedduringthefirstminuteofthe700–1300
particlesonsuchimages.ForCOMS1(Fig.4a),thetransition between the composite layer and the core is ill-defined, possiblybecausethecorealsocontainsCoanda-Al2O3.The
compositelayer(Fig.4b)ismadeupofareas containingCo particles ofmarkedlydifferentsizes(0.03and0.5–1.0mm). ForCAS1(Fig.4candd),thespecimenisveryporousandis homogeneous (0.45–0.60mm Co particles are observed everywhere), inagreementwithXRD data.TheCA powder is the one with the higher specific surface area, therefore reduction was easier and the Co-Al2O3 composite was
formed throughout thesample and thusno boundaryexists. ForUS1,thetransitionbetweenthecompositelayer(240mm thick) and the core is sharp (Fig. 4e). The microstructure resemblesthatofCOMS1withtwopopulationsofCoparticles butwithsignificantlyhighersizes(0.80–1.0and2.0–3.5mm) (Fig.4f).
SeveralspecimensweresinteredusingpowderUbecause the transition between the core and the composite layer is sharp.Theuniaxialpressureisappliedincreasinglyearlyfor theUS2,US3,US4andUS5specimens,respectively(Section
2.2 and Table 1). For all samples, the XRD patterns (not shown) of the surface and core are similarto that for US1 (Fig. 3), i.e. Co and a-Al2O3 are detected at the surface
whereasCo(veryweak)andCoAl2O4aredetectedatthecore
(Table1).Thethicknessofthecompositelayer(dS–Table1),
measured on FESEM images (not shown) similar to that shownonFig.4e,isshowninFig.5versusthedensityofthe specimens; sample US1 with no applied pressure was also includedforcomparison.Applyingthepressureearlyin the cyclefavorsdensification(4.09g/cm3forUS5)buthampers the transformation of CoAl2O4 into Co and Al2O3, the
thicknessofthecompositelayerbeingthelowest(57mm)for Fig.2. LowandhighmagnificationFESEMimagesoftheCOM(aandb),CA(candd)andU(eandf)CoAl2O4powders.
US5.NotethatforUS4thethicknessisalmostdoublethanfor US5foradensityonlyslightlylower(40.7vs4.09g/cm3).For US2,increasingthedwelltimefrom3to9min(US2(3)and US2(9))at 13008Cfavors the formation ofa muchthicker layer(436vs202mm)attheexpenseofdensification(3.5vs 3.8g/cm3).
It is proposed thatthe cobaltparticles are precipitated as described by reactions (1) and (2) during the SPS process because of a combination of high temperature and low O2
partialpressure:
CoAl2O4 !Co1ÿxAl2O4þxCo (1) Co1ÿxAl2O4 !ð1ÿxÞCo þ Al2O3þ½O2 (2) In reaction (1), the precipitation is not total and an aluminum-rich spinel [10,11] is formed. For the surface of thespecimens, oreventhebulkofthe samplefromthe more reactive powder (CA), the aluminum-rich spinel becomes unstableandtheprecipitationproceedswiththeformationof a-Al2O3 along with more Co particles. Note that reduction
processesasdescribedbyreactions(3)and(4)cannotberuled out,although reaction (4) isconsidered,from theanalysis of earlierresults[9],tobequiteunlikelyexceptfor theextreme topmostsurface ofthesample:
CoAl2O4þCO!Co þAl2O3þCO2 (3)
CoAl2O4þ½C!Co þAl2O3þ½CO2 (4)
Thus,ahigherspecificsurfaceareaoftheCoAl2O4powder
willfavortheescapeofgases(O2,CO2)fromthesampleand
the transformation into a Co–Al2O3 compositewill progress
deeper into the sample. However, these processes are super-imposedwiththoseassociatedwithdensificationandnotably differential sintering phenomena: firstly, the more compact partsofthepowder,i.e.thosewiththesmallergrainsand/orthe more agglomerated ones, become still denser; then the precipitation/reduction processes occur, accounting for the formationofareascontainingverysmallCoparticles,inthese dense parts,andareaswithlargerCo particles inthenot yet dense parts where surface coalescence is still possible. Eventually,thelatterareasbecomedenser.
The powder with the higher specific surface area (CA, 78m2/g)isthemorereactivebuttheCAS1sampleistheless denseone(2.6g/cm3).BycontrastpowderU(18m2/g)isthe least reactive powder and produces the denser sintered specimen (3.4g/cm3). Thus, it seems that transformation hampersdensificationduetotheevolvedgas.Theresultsonthe differentUSspecimens couldalso reflectthepossible roleof openporosityintheprocess.Applyingthepressureearlyinthe cycleatlowtemperature(US4andUS5)mayfavortheclosing ofporosity,whichwoulddecreasethepossibilityofO2and/or
CO2leavingthesample,beforetheformationofCobyreaction
(2)and/or(3)isthermallyactivated.Thereactionzoneandthus thecompositelayerarethinner,asobservedinapreviousstudy ontheformationofFe–Al2O3layers[9].Thisalsorevealsthat
therelationshipsbetweentransformationanddensificationare quitecomplex,because each oneisabletohamperthe other. Applying the pressure during the first minute of the 700– 13008Cramp(asforUS4)isagoodcompromiseifonewants toobtainaCo–Al2O3compositelayerabout100mmthickwith
aspecimendensityover4g/cm3.Applyingthepressureduring thefirstminuteoftheroomtemperature–7008Cramp(asfor US5) is a good compromise if one wants to obtain dense CoAl2O4,afterremovalbygrindingofathincompositelayer.
Fig.3. XRDpatternsoftheCOMS1,CAS1andUS1specimenspreparedby SPS.(a)Surface;(b)core.
3.3. Microhardness andfriction behavior
Thesinteringexperimentalconditions,densityandthickness ofthe compositelayerfor specimens US6,US7andCOMS2 aresummarizedinTable2.AsmentionedinSection2.3,thetop side of the specimens was only slightly polishedin orderto reveal the surface composite layer and the bottom side was groundinordertorevealthecore.ForUS6,thecompositelayer (Fig. 6a) shows areas containing Co particles of markedly
differentsizes(ca.0.3and3.0mm)(dCo–Table2),asobserved
aboveforCOMS1andUS1(Fig.4).Thelateapplicationofthe pressure favored the growth of the Co particles. For US7 (Fig.6b),theresidualporosityissignificantlyhigher,andthere arestilltwopopulationsofCoparticles(ca.0.3and1.5mm), althoughthegrowthhasbeenlimited.Bycontrast,thereisonly onepopulation(ca.0.3mminsize)forCOMS2(Fig.6c),with onlysomecoalescenceatthegrainjunctions(particlesca.1mm insize). The proportion of surface area occupied by the Co Fig.4. FESEMimagesofacross-sectionofthecompositespreparedbySPS:COMS1(aandb),CAS1(candd)andUS1(eandf).
particles,determinedbyanalysisofsimilarFESEMimages,is equal to about 23,24 and 17% for US6, US7 and COMS2, respectively(SCo–Table2).TheHVvaluesmeasuredforthe
compositelayers(HVsurf–Table2)arefairlylowforUS6and
US7(9.6and8.3GPa,respectively),whichcouldresultfroma lack of densification and an exaggerate growth of the Co particles[15].ThevaluefoundforthesurfacelayerofCOMS2 (13.6GPa)corresponds tothat reported [15] for a bulkCo– Al2O3compositewithaCocontentequaltoabout45wt%and
thesizeoftheCoparticlesequaltoabout0.75mm.Thereare marked differences between the specimens regarding the friction behavior (msurf – Table 2 and Fig. 7). ForUS7, the
frictioncoefficientisalwayshigherthanforanAl2O3specimen
preparedbySPS[5].Thecontactstabilization,toavalueabout doublethatforAl2O3,isslow,allthemoresowhentheloadis
increasedfrom5to10N.Thiscouldreflecttoomuchcontact between the sample and the alumina ball because of a continuouspull-outofCoparticlesandAl2O3grains,duetothe
low microhardness and relatively high porosity of the compositelayer.By contrast, thefriction coefficient for US6 andCOMS2islowerthanfortheAl2O3specimenfora5Nload
(Fig.7a)andtheaveragevaluesarelow(msurf=0.27and0.23,
respectively).Thiscouldreflectthelubricatingroleof theCo particles.Fora10Nload(Fig.7b),thefrictioncoefficientfor US6andCOMS2isinitiallymuchlowerthanforAl2O3,butthe
curvegetsprogressivelynoisierstartingatabout100cyclesand thereisastrongincreaseat180cyclesforUS6,andamilder oneforCOMSatabout160cycles,revealingmajorchangesin thecontact, probablybecause of thepulling-outandpossible oxidationof the Co particles. The friction results are on the wholebetter forCOMS2thanforUS6, whichcould reflecta combination of higher hardness and a more homogeneous microstructure,withonlyonepopulationof Coparticles.
FESEM observations showed that the core of all three samplesismadeupofCoAl2O4andaverysmallproportionof
Coparticles.Themicrohardnessvaluesaresimilar,intherange 14.9–15.9GPa (HVcore – Table 2). The average friction
coefficients (mcore – Table 2) are similar too, in the range
0.28–0.31fora5Nloadandintherange0.30–0.34fora10N load.TheyaresignificantlylowerthanthatofanAl2O3sample
(about0.40).Courbiereetal.[16]havereportedthatCoAl2O4
layersgrownonAl2O3showsasimilarorslightlylowerfriction
coefficientthanAl2O3,butshowhigherwear,inatesthowever
totallydifferent(steelball,waterlubrication,muchhigherload) Fig.5. Thicknessof thesurfacecomposite layerversus thedensityofthe
specimenstheUS2,US3,US4andUS5specimenspreparedbySPS.
Fig. 6. FESEM image of the polishedsurface of US6 (a), US7 (b) and COMS2(c).
thanthepresentone.Tothebestofourknowledge,itisthefirst timethatsuchfrictiondataarereportedforbulkCoAl2O4.The
tribological properties of CoAl2O4 warrant more studies, in
particularwearwill bereportedelsewhere. 4. Conclusions
Materials with a Co–Al2O3 composite coating over a
CoAl2O4corewerepreparedduringSPSofCoAl2O4powders.
TheCoparticlesareprecipitatedbecauseofacombinationof hightemperatureandlowO2partialpressure.Ahigherspecific
surfaceareaoftheCoAl2O4powderfavorstheescapeofgases
(O2, CO2) from the sample and thus the progress of the
precipitationdeeperinto thecore,butthishampers densifica-tion.Applyingthepressureearlyinthecycleatlowtemperature toincreasedensificationalsofavorstheclosingofporosity,thus decreasingthepossibilityofgasestoleavethesample,resulting inathinnercompositelayer.Applyingthepressureduringthe first minuteof the 700–13008Crampallowsonetoobtain a Co–Al2O3 composite layer about 100mm thick, with a
specimendensity over4g/cm3.Applyingthe pressureduring
thefirstminuteoftheroomtemperature–7008Cramppermits toobtaindenseCoAl2O4,after removalbygrindingofathin
composite layer. The friction behavior of the Co-Al2O3
composites against an Al2O3 ball depends strongly on the
sample microstructure, residual surface porosity and applied load. For a 5N load, specimens show a friction coefficient lowerthanthatfoundforareferenceAl2O3specimenandthe
averagevaluesarelow,whichcouldrevealthelubricatingrole ofthesubmicrometerCoparticles.However,fora10Nload, the initially very low friction coefficient shows a strong increaseat160–180cycles,revealingsomemajorchangeinthe contact, probably because of the pulling-out and possible oxidationoftheCoparticles.Interestingly,CoAl2O4wasfound
tohave aVickersmicrohardness inthe range 14.9–15.9GPa andaveragefrictioncoefficients(0.28–0.31fora5Nloadand 0.30–0.34fora10Nload)lowerthanthatofanAl2O3sample
(about0.40).Tothebestofourknowledge,it isthefirsttime thatsuch frictiondataarereported forbulkCoAl2O4.
Acknowledgments
SPSwasperformed atthePlateformeNationale CNRSde Frittage-Flash (PNF2, Toulouse). Electron microscopy was performed at TEMSCAN, the ‘‘Service Commun de Micro-scopieElectronique’’, Universite´ Paul-Sabatier.
References
[1] A.K.Dutta,A.B.Chattopadhyaya,K.K.Ray,Progressiveflankwearand machiningperformanceofsilvertoughenedaluminacuttingtoolinserts, Wear261(2006)885–895.
[2] S.T.Oh,S.J.Yoon,Y.H.Choa,Y.K.Jeong,K.Niihara,Wearbehaviorof nano-sizedmetalparticlesdispersedAl2O3nanocomposites,KeyEng. Mater.317–318(2006)369–372.
[3] G.dePortu,S.Guicciardi,C.Melandri,F.Monteverde,Wearbehaviourof Al2O3–MoandAl2O3–Nbcomposites,Wear262(2007)1346–1352. [4] P.Stempfle´,F.Pollet,L.Carpentier,Influenceofintergranularmetallic
nanoparticlesonthefrettingwear mechanismsofFe–Cr–Al2O3 nano-compositesrubbingonTi–6Al–V,Tribol.Int.41(2008)1009–1019. [5] J.GurtSantanach,C.Estourne`s,A.Weibel,A.Peigney,G.Chevallier,Ch.
Laurent,MechanicalandtribologicalpropertiesofFe/Cr–FeAl2O4–Al2O3 nano/microhybridcompositespreparedbySparkPlasmaSintering,Scr. Mater.64(2011)777–780.
[6] T.Rodriguez-Suarez,J.F.Bartolome´,A.Smirnov,S.Lopez-Esteban,R. Torrecillas,J.S.Moya,Slidingwear behaviorofalumina/nickel nano-compositesprocessedbyaconventionalsinteringroute,J.Eur.Ceram. Soc.31(2011)1389–1395.
[7] A.Mukhopadhyay,R.I.Todd,Microstructureandmechanicalproperties ofAl2O3matrixnanocompositesproducedbysolidstateprecipitation,J. Eur.Ceram.Soc.30(2010)1359–1372.
[8] J.GurtSantanach,C.Estourne`s,A.Weibel,A.Peigney,G.Chevallier,Ch Laurent, Sintering as a reactive sinteringtool for the preparation of surface-tailored Fe–FeAl2O4–Al2O3 nanocomposites, Scr. Mater. 60 (2009)195–198.
[9] J.GurtSantanach,C.Estourne`s,A.Weibel,G.Chevallier,V.Bley,Ch. Laurent, A.Peigney,Influenceof pulse currentduring Spark Plasma Sinteringevidencedonreactivealumina–hematitepowders,J.Eur.Ceram. Soc.31(2011)2247–2254.
[10] A.Navrotsky,O.J.Kleppa,Thethermodynamicsofcationdistributionin simplespinels,J.Inorg.Nucl.Chem.29(1967)2701–2714.
[11] B.Hallstedt,ThermodynamicassessmentofthesystemMgO-Al2O3,J. Am.Ceram.Soc.75(1992)1497–1507.
Fig.7. FrictioncoefficientoftheCo–Al2O3layersversusnumberofcyclesfor appliedloadsof5N(a)and10N(b)againstanAl2O3ball.
[12] K.C.Patil,Advancedceramics:combustionsynthesisandproperties,Bull. Mater.Sci.16(1996)533–541.
[13] W.Li,J.Li,J.Guo,Synthesisandcharacterizationofnanocrystalline CoAl2O4spinelpowderbylowtemperaturecombustion,J.Eur.Ceram. Soc.23(2003)2289–2295.
[14] E.Flahaut,Ch.Laurent,A.Peigney,CatalyticCVDsynthesisofdouble andtriple-walledcarbonnanotubesbythecontrolofthecatalyst prepara-tion,Carbon43(2005)375–383.
[15] W.P.Tai,T.Watanabe,PreparationandmechanicalpropertiesofAl2O3 reinforcedbysubmicrometerCoparticles,J.Mater.Sci.33(1998)5795– 5801.
[16] M.Courbiere,R.Trabelsi,D.Treheux,C.Beraud,C.Esnouf,G.Fantozzi, Growthandwearofsuperficialternarymetaloxidesuponpurealumina, in:P. Vincenzini (Ed.),High Tech Ceramics, ElsevierSc. Pub.B.V., Amsterdam,1987,pp.2599–2608.
