Colloidal stability-slip casting behavior relationship in slurry of mullite synthesized by the USP method

Other uses, including reproduction and distribution, or selling or

licensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of the

article (e.g. in Word or Tex form) to their personal website or

institutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies are

encouraged to visit:

Colloidal

stability–slip

casting

behavior

relationship

in

slurry

of

mullite

synthesized

by

the

USP

method

Remzi

Go¨ren

,

Bahri

Ersoy

,

Cem

O

¨ zgu¨r

*

,

Talip

Alp

a_{DumlupinarUniversity,DepartmentofCeramicEngineering,Ku¨tahya43100,Turkey} b

AfyonkocatepeUniversity,DepartmentofMaterialsScienceandEngineering,Turkey c

UniversityofYalova,FacultyofEngineering,Turkey

Received4May2011;receivedinrevisedform21July2011;accepted25July2011 Availableonline30thJuly2011

Abstract

Thisstudypresentstheoutcomeofaresearchconcerningtherelationshipbetweenthecolloidalstabilityofmullitepowderssynthesizedbythe

USP(ultrasonicspraypyrolysis)methodanditsslipcastingbehavior.Thecolloidalstabilityofmulliteslurryhasbeeninvestigatedunderthree

differentpHconditions(4.5,8.9and10.9)derivedfrompH-dependentzetapotential(ZP)curves.EmployingthesepHvalues,mulliteslurrieswith

50wt.%solidcontentwerepreparedandslipcast.ThemicrostructuresofdriedandsinteredspecimenswereexaminedusingSEM.Itisconcluded

thatthepHsignificantlyinfluencesthestabilityandinturntheslipcastingbehaviorofthemulliteslurry.Inordertopreparehomogeneousand

stablemulliteslurryforefficientslipcastingitispreferabletoutilizeabasicratherthananacidicmedium.HighpH(i.e.10.9)tendstoleadstomore

closelypackedmulliteparticlesresultinginahomogeneousmicrostructureandgreaterstructuralintegrity.

#_{2011PublishedbyElsevierLtdandTechnaGroupS.r.l.}

Keywords: A.Slipcasting;D.Mullite;Zetapotential

1. Introduction

Mullite(3Al2O32SiO2)isbecomingincreasinglyimportantin

electronic,optical,andhightemperaturestructuralapplications duetoitsattractivepropertiessuchasexcellenthightemperature strength,resistancetocreepandthermalshock,lowdielectricand thermalexpansionconstants[1].Someimportantapplicationsof mulliteand/ormullitebasedceramicsincludeamatrixmaterial for continuous fibre reinforced ceramic matrix composites (CMCs) used as thermal protection systems for combustion chambersinaircraftturbineenginesandstationarygasturbines, heatexchangeparts,heatinsulatingparts,millingmedia,furnace centertube,refractoriesinthemetallurgicalindustriesforelectric furnace roofs,protectivecoatings,turbineenginecomponents, hotmetalmixersandlowfrequencyinductionfurnaces[1–5].

Slip casting is an attractive and well established forming methodforpowderbasedshapingofceramiccomponentsthat

hasbeenusedforalongtime.Itiswidelyusedintheproduction ofceramics,duetoitsversatilityas aconsolidationprocess to obtain materials with high green density and microstructural homogeneity [6]. In this process, slurry containing a high proportionofsolidwithahighdegreeofstabilityisprerequisite forproductswithhighgreendensity.Unstable(orundispersed) slurrycausesdefectsinthefinalsinteredbodywhichnegatively affectsthefinalmechanicalandphysicalproperties[7].Itiswell knownthattherearetwomainoperatingrepulsiveforceswhich contributetothestabilityofacolloidalsuspensionlikeceramic slurry;(i)electricaldoublelayer(EDL)and(ii)stericforces[7– 10].Stericrepulsiveforcesbetweentwoparticlesgenerallyoccur whentheirsurfacesarecoatedwithapolymerbaseddispersantso thattheypreventtheparticlesfromphysicalcontact.However, theEDLforcesbetweencolloidalparticlesinasuspensionoccur whenoverlappingoftheEDLsoftwoadjacentparticles takes place and can be easily changed through changing of the suspensionpH,orelectrolyte/dispersantaddition.Eachmineral has an IEP (a certain pH value) controlled by the mineral powder’s molecularconstitution. In general, mineralpowders typically flocculate (become unstable) at the IEP and they deflocculate(stabilize) as thepH increasesor decreasesaway

www.elsevier.com/locate/ceramint

Available online at www.sciencedirect.com

CeramicsInternational38(2012)679–685

*Correspondingauthor.Tel.:+902742652031/4312; fax:+902742652066.

E-mailaddress:cozgur@dumlupinar.edu.tr(C.O¨ zgu¨r).

0272-8842/$36.00#_{2011PublishedbyElsevierLtdandTechnaGroupS.r.l.} doi:10.1016/j.ceramint.2011.07.056

fromIEP.Hence,ifchargestabilizationisthedominantstability mechanismforasuspension,thenthezetapotentialcanbeused topredictsuspensionstability[11].

The relevant literature cites several researches on the production andmechanicalandphysicalpropertiesof various mullitecontainingceramiccompositessuchasalumina/mullite, alumina/mullite/zirconia,mullite/zirconia, mullite/zirconia/zir-con andcordierite/mullite formed bythe slip castingmethod

[12–20].However,withtheexceptionoftheresearchcarriedout byCameruccietal.(1998)nootherstudyhasbeenfocusedon suspension parameters such as pH, type and quantity of dispersantandtheireffectsonthecharacteristicsof greenand sinteredproducts.Cameruccietal.discussedtheeffectofsolid contentanddispersant(polyethyleneglycol)ontheviscosityof cordierite–mulliteslurry,andconsequentlyonthepropertiesof bothgreenandsinteredproducts[20].Theyhaveconcludedthat bothparameterssubstantiallyinfluencetheslipcastingbehavior of the slurry. Nevertheless, in order to fully determine the stabilityandflowcharacteristics,andexplainthe stabilization mechanismsofslurriescontainingdifferenttypes ofminerals, thestabilizingconditionsforeachconstituentinthemixturemust be identified.Publishedresearchesonthe microstructuraland physicalpropertiesofproductsmanufacturedfrommullitealone bytheslipcastingarescanty[21–23].Tkalcecetal.investigated theeffectofsinteringtemperatureonthemechanicalproperties andmicrostructureofproductsobtainedbyslipcastingofrefined mullitepowderspreparedbythesol–geltechnique[22]. Burgos-MontesandMoreno,ontheotherhand,examinedtheeffectofthe dispersantonthestabilityofmulliteslurryinwhichthemullite powders were produced by the ‘‘combustion synthesis’’ technique [23]. They obtained strong sintered products from mullite slurry with polyacrylic acid dispersant. The work of Hashiand Senna appears tobe theonly attempttostudy the influenceofpHonthestabilityofmulliteslurryandinturnonthe greenandsinteredproducts[21].Theyexaminedtherelationship betweenthepHandtherheologyofmulliteslurryandporesize distributionofthegreencompacts.Theypointedoutthatproduct slipcastfromshearthinningslurriespreparedatpH8possesa largertotal pore volumethanthose prepared with pH 4. The larger pore volume is attributed to the occurrence of agglomeration at pH 8 which is close to the pH 8.7 values correspondingtotheIEPofmullite.Thepresentresearchaimsto contribute to the relevant literature by providing a detailed discussiononthe effectof pHonthe slipcasting behaviorof mullite slurry without dispersant, and consequently on the microstructure of green compacts and sintered products manufactured from mullite produced by the USP technique. Another objective of this investigation is to resolve the discrepancy between the IEP values of mullite reported by differentresearchers[21,23].

2. Materialsandmethods

2.1. Synthesizingand characterizationofmullitepowder The USP system which was used for the preparation of mullitepowdersandthepowderpreparationstepsinvolvedare

describedindetailelsewhere[24].TheUSPsystemconsistsof anultrasonicnebulizer,aquartztubeandapowdercollecting unit. The starting chemicals were tetraethylorthosilicate (TEOS, Fluka, of 98% purity) and aluminum nitrate nanohydrate(Al(NO3)39H2O,withextrapureMerck),

contain-ing silicon and aluminum species, respectively. For the preparation of the solution, TEOS with a molarity of 1.5 wasaddedtodistilledwaterandmixedwithmagneticstirrerat 500rpmtoformaclearsolutionduringwhich0.2MHNO3was

addedtothewater.Finally,aluminumnitratenanohydratewas dissolved with continued stirring to obtain the required solution. Clarity of solutions as measured by turbidimeter wasregardedasacceptablewhentheturbidityvaluewasbelow 0.1NTU. The solution wasplaced inanultrasonic nebulizer (frequency, 1.63MHz) to produce aerosol in small droplets whichwere thenconveyedthroughaquartztubeplaced ina vertical furnace, where the constituents combine to form mullite.Theairflowrateusedwas1.5l/min.Thetemperature of the tubular furnace was fixed at 10008C. The pyrolyzed powders were collected by filtration using membrane filters (Schleicher &Schuell-NL 16)havingaporesize of 0.2mm. The powders were then characterized by XRD (RigakuTM miniflex),SEM (Leo-1430 VP)and FTIR (BrukerVertex 70 seriesFourierTransformInfraredSpectrometer)techniques.In addition,thermaldecompositionofthesolutionwasstudiedby TG/DTA(PerkinElmer-Diamond)inordinaryairataconstant heatingrateof108Cmin1.

2.2. Zetapotential(ZP) andrheologicalmeasurements In diluteaqueoussolutions, thestabilityofthe synthesized mullite powderswas studied as afunction of pH through ZP measurements that were performed using the electrophoresis method(ZetaSizer NanoZSmodel,Malvern, UK).The dilute suspensionswerepreparedbyadding0.4gofthemullitepowder intodistilledwaterinaglassbeaker.TheZPexperimentswere conductedfortwodifferentsolutionswhichcontained103M and 102M NaCl respectively to maintain a constant ionic strength.40mlsamplesofthesuspensionsthuspreparedwere furtherstirredbymagneticmixerat700rpmfor2minbeforeand after pH adjustments. The solutions with varying pH values rangingfrom2to12wereadjustedtotheappropriatepHusing stock solutions of KOH and H2SO4. pH measurements of

solutionsthewereperformedwithasensitivityof0.01.Priorto ZPmeasurements,ultrasonicvibrationswereappliedtoprevent agglomerationand/or breakupsoft agglomeratesat a20kHz frequencyand35%amplitudefor30s.

Theflowbehaviorofconcentratedsuspensions(slurries) of themullitewasdeterminedusingarotationalrheometer(Bohlin Instruments-CVO) with shear rate control. All rheological measurements were performed at the ambient temperature. Concentrated suspensions of powders were prepared to solid contentof50wt.%withoutusingadispersingagent.Themullite slurrieswere preparedbyadding 10.2g of mullite powderto 10.2gof aqueoussolution,of 103MNaCl.Homogenization wascarriedoutby6hballmillingusingaluminajarandballs. Prior to rheological measurements, pH adjustments of the

R.Go¨renetal./CeramicsInternational38(2012)679–685 680

suspensionswereperformedat4.65,8.9and 10.9,afterwards, theywerestirredfor10min.Flowcurvesplottedasshearstress versusshearrate,wereobtainedundercontrolledratecondition withathree-stagemeasuringprogramwithalinearincreaseof the shear rate from 0.2 to 1000s1 for 2min, a plateau at 1000s1for1min,andafurtherdecreaseto0.2s1for2min. 2.3. Slip castingand microstructureanalysis

ThemulliteslurriespreparedatthreedifferentpHvalues(4.5, 8.9and10.9)wereslipcastonplasterofParismouldstoobtain solid discs (2cm in diameter). After initially dryingat room temperaturefor 48h and subsequentlyat1058Cfor24h,the greencastproductsweresinteredat16008Cusingaheatingrate of58Cmin1andasoakingtimeof5h.SEMobservationswere performed onthe gold-coated specimensafter the dryingand firingprocessesusingthesecondaryelectronsgeneratedbySEM (Leo-1430 VP). In addition, percentage porosity of sintered specimens was measured by Hg-porosimetry (Quantachrome, Poremaster60000).

3. Resultsanddiscussion 3.1. Characterization

TheDTA/TGcurvesoftheclearsolutionwiththechemical constituents mentioned in Section 2.1 up to 12008C is presented inFig. 1.In the DTA curve,there are three peaks at1258C,2508Cand9908Ctwoofwhichareendothermicand one is exothermic respectively. The endothermic peaks correspond to loss of H2O, nitric acid, decomposition of

TEOSandaluminumnitratenanohydrate.Theexothermicpeak at around 9908C indicates that the atomic scale solution preparedbymixingAlandSispecieshasconverteddirectlyto mullite [25–28]. Accordingly, it was thought that it may be possible toobtainmullitepowdersaround9908C.

The mullite powders obtained by the USP method were characterized via XRD and FTIR techniques. X-ray analysis revealed the samples to be amorphous (Fig. 2). However,

spectral analysis by FTIR (Fig. 3), a technique based on vibrations of the atoms of a molecule and successfully employed in the characterization of inorganic compounds (i.e.minerals)aswellasorganiccompounds[29,30]revealed thesampletohavemullite-likemoleculargroups.Asinthecase ofsilicateminerals,thetwopeaksappearedaroundat3400and 1600cm1 maybeattributedtothe OH stretchingvibrations andthedeformationvibrationofadsorbedwater,respectively

[31].Theotherpeaklocatedat3750cm1maybeattributedto the OH stretching vibration arising from the Si–OH groups

[32]. The shoulders around 1152cm1 and 456cm1 are

Fig.1. DTA/TGcurveofthesolution.

Fig.2. XRDpatternofthemullitepowdersynthesizedbytheUSPmethodat 10008C.

assignedtovibrationalmodeof theasymmetricstretchofSi– O–SiandbendingofSi–O–Si,respectively[33,34].Thebandat 567cm1 is ascribed to Al–O stretching vibration modes resulting from AlO6 groups of mullite [35,36]. The band at

751cm1isinagreementwiththeproposed(Si,Al)–O–(Si,Al) bending mode for the band at 737cm1 in mullite [25]. According to the literature [37,38], the fundamental bands assignedtothe SiO4(482, 988, 1107,1131and1168cm1),

AlO4(620,828and909cm1),AlO6(578and482cm1)and

T–O–T (TO4whereTmeans Si or Al) (737cm1)show the

presenceand/orformationofthemullite.Boththeliteratureand thebandsat567,751,882and1137cm1appearedintheFTIR analysisindicatethatthemullitepowderproducedbytheUSP technique includes amorphous mullite phase containing the shortrangeorderofAlO6,SiO4,AlO4andSi–O–Allinkages.

The SEM micrograph of the mullite powder obtained by USP at10008CisshowninFig.4.The powderparticles are shown to be of spherical shape with smooth surfaces and unagglomerated.Themeanparticlesize isabout1.1mm. 3.2. DeterminingtheIEP (isoelectric point)ofmullite

The IEP ofmullite wasdirectlyobtained byelectrokinetic measurementsofZPagainstpH,inthepresenceofindifferent electrolytesofvariousmolaritiesasseeninFig.5.TheIEPalso indicatesthatatthispoint(orpH)thereisnochargeatthesurface, that is,the totalpositive chargeisequal tothe totalnegative charge.ElectrokineticstudiesrevealedthattheIEPformullite wasatpH8.9.Thisisinagreementwiththeresultofastudyby HashiandSenna[21].Burgos-MontesandMoreno,ontheother hand,have reported apH valueof 5.9for the IEP of mullite producedbythecombustionsynthesismethodinthepresenceof background electrolyte, 102M of NH4NO3 [23]. However,

accordingtotheliterature[9],inordertoaccuratelydetermine theIEPofanymineral,itszetapotentialmeasurementsagainst pH must be performed in the presence of indifferent (or background)electrolyteswhichareusuallymonovalent symme-tricalelectrolytessuchaschlorideornitratesaltsofNa,K,NH4

usingtwoorthreedifferentmolarities(i.e.103,102,101M).

Thisisbecausetheindifferentionssuchasmonovalentcationsor anions are generally not able to adsorb in the IHP (Inner HelmholtzPlane)oftheSternlayerintheelectricaldoublelayer (EDL)ofamineralparticle.Hence,theycannotshifttheIEPeven whentheirconcentrationincreases[39].Consequently,asseen clearlyinFig.5,theIEPofthemullitepowdersusedinthisstudy correspondstoapHvalueof8.9whichisveryclosetotheIEP (pH8.7)ofcrystallinemullitestudiedbyHashiandSenna[21]. However,thereisasignificantdifferencebetweentheIEPvalues of mullite obtained from spray pyrolysis method used in the present study andthe combustion synthesis method(pH 5.9) utilized by Burgos-Montes and Moreno [23]. As mentioned earlier, the FTIR spectrum of mullite obtained by the USP method(Fig.3)showsthatthesamplehashydroxylgroupsand/ or watermolecules bound tometallicatoms(Alor Si) inthe structure.Unfortunately,noFTIRspectrumhasbeengivenfor the mullite powder in the study due to Burgos-Montes and Moreno [23]. As is known from the literature[40] hydration increasestheIEP,whereasdehydrationdecreasesit.Thatis,as thethicknessoftheadsorbedwaterlayerontheparticlesurface increases,theshearplaneinthe EDLcanbeexpectedtoshift away from the surfacethereby lowering the ZP, and inturn, shiftingtheIEPoftheparticletolowerpHs.Anotherplausible explanationforthisdivergencecouldbeanycontaminationofthe surfaceofmullitepowdersduringproductionorpretreatments duringZP measurements.According totherelevant literature such a surface contamination may lead to vary the IEPs of mineral powders [41]. On the other hand, the zeta potential increasessharplywithdecreaseinpHfromzeromV atpH8.9to the about 50–60mV at about pH 6 depending on the ionic strengthofthemediumduetodissociationof–MOH(whereM symbolizesSiorAl)groupsorproton(H+)adsorptiononthese groups,e.g.

½MOH þHþ !MOH2þ $MþþH2O

Further decrease in pH does not affect ZP significantly. Beyondthe IEP(pH8.9), increaseinpH reversesthe surface charge of mullite from positive to the negative attaining the

Fig.4. SEMimageofthemullitepowdersynthesizedbytheUSPmethodat 10008C.

Fig.5. VariationofzetapotentialofmulliteasafunctionofpHinthepresence ofindifferentelectrolyte(NaCl).

R.Go¨renetal./CeramicsInternational38(2012)679–685 682

levelof(30–33)mVataboutpH11and60mVatpH12 duetotheadsorptionofOH–ordeprotonation reactions,e.g. ½MOH þOH!MðOHÞ2 $MOþH2O

3.3. Shearstress versusshearratecurves

Stability of slurry with appropriate rheological properties coupled withhighsolid content andpseudoplastic (or shear-thinning)behaviorisfavoredtoyieldproductswithhighgreen density[21].Theflowcurvesobtainedbyplottingshearstress versusshearrateformulliteslipatpH4.5,8.9(IEP)and10.9 arepresentedinFig.6.Allthreeslipsdisplayedapseudoplastic behavior. As showninFig.6,the slurryexhibits amaximum shear stress at pH=8.9 (IEP), a minimum at pH=4.65 and intermediateatpH=10.9.ThisisduetothefactthatatIEPthe EDLrepulsiveforcesarenon-existentsothatparticlescluster together under the influence of van der Walls forces of attraction.Thisinturndecreasestheflowrateandincreasesthe viscosityoftheslurry.SEMdataclearlysupportthesefindings (Fig.7).Similarly,theZPdataisinagoodagreementwiththe viscosity. The shear stress values of the mullite slurries obtainedattheshearratesof1000s1and103MNaClwere 10.6,14.8and12.3PaforpH4.65,8.9and10.9,respectively. UnderthesamepHconditions,theZPvalueswere+58,0and 33mV,respectively.Similarresultswereobtainedinastudy undertakenbyHashiandSennarelatingtotheeffectofpHon the zetapotentialandviscosityofthe mullite[21].

3.4. Microstructureanalysisofgreenandsinteredbodies Greenbodies wereobtainedby castingof the suspensions intoaplastermould.Fig.7depictsthefracturesurfacesofdried green bodies as observed through the scanning electron microscope(SEM).Thefracturesurfaceofthespecimensslip cast at pH 10.9, the SEM micrograph reveals 1–2mmsized mulliteparticleswitharelativelyuniformsizedistribution.As clearlyseenfromFig.7b,mulliteparticlesintheslurrycastat pH8.9(IEP)weretotallyagglomerated.Thatis,thepackingof sphericalmulliteparticlespackedisirregular,inhomogeneous andloose.Thereseemstobenosignificantdifferencebetween thepackingofthemulliteparticlesinspecimenscastfromthe

slurries prepared under acidic (pH 4.5)and basic (pH 10.9) conditions respectively.

Fig. 8 shows fracture surfaces of specimens slipcast and sinteredat16508Cfor5h.TheSEMfractographsindicatethat thesphericalshapedmullitepowdersobtainedbytheUSPtends toassumethewellknown rod-likestructure aftersinteringat 16508C, while the mullite slip cast at pH 10.9 reveals an evidentlymoreuniformsizedistributionandnon-agglomerated surface. Indeed, when looked at Fig. 5 the absolute ZP of mullitepowdersatpH10.9(33to35mV)islowerthanthatof pH4.5(48to53mV)andthusitisexpectedthatthepacking of mullite powdersat pH 4.5isbetter than thoseatpH 10.9

Fig.6. Shearstressversusshearrateformulliteslurriespreparedatdifferent pHs.

Fig.7. SEMviewsofdriedslipcastspecimenspreparedat(a)pH=4.5,(b) pH=8.9(IEP)and(c)pH=10.9.

which will lead to more homogenous microstructure and a compactfinalproductatpH10.9.Butforthiscontrariness,no explanation wasfoundatpresent. Itis importanttonotethat comparison of Figs. 7b and 8b supports the notion that inhomogeneity and loose packing of particles in a green

compactwouldleadtoanon-homogeneousandlooselypacked microstructureaftersintering.Mercury-porosimetry measure-mentsonsinteredsamplesreinforcethisview(Table1).Thatis, thehighestporosity(17.1vol%)wasobtainedforthesintered sample slipcast atpH8.9.

4. Conclusions

Using the ultrasonic spray pyrolysis (USP) method, it has beenshownthattheproductionofsphericalmullitepowders withaverageparticlesizeof1.1mmandinnarrowsizerange isfeasible.

Theiso-electricpoint(IEP)forthemullitepowdersproduced by the USP method has been determined as pH 8.9. As expected, at this pH the viscosity of the slurry is seen its maximum, while packing of mullite particles is the least closelypacking.

Mulliteslurriespreparedwith50%solidcontentanddifferent pHshavebeenshowntoexhibitpseudoplasticflowbehavior. Ithasbeenpossibletoobtainmulliteslurriesofhighstability by simply adjusting the pH level alone. However, SEM observations on cast and sintered products and also Hg-porosimetry measurements on sintered products have evidentlyshownthatinordertoincreasethe ZPof mullite as absolute value by pH adjustment, addition of basic electrolyte such as KOH rather than an acidic electrolyte yieldsbetter results.

NotwithstandingthefactthattheabsoluteZP(33to35mV) inthevicinityofpH11islowerthantheabsoluteZP(48to 53mV)obtainedatpH4.5,themicrostructureoftheproduct prepared from basic mullite slurry has been shown to be unequivocally more homogeneous. This conditionleads to the developmentlowporosity,inturnhighdensity.

References

[1] I.A.Aksay,D.M.Dabbs,M.Sarikaya,Mulliteforstructural,electronic, andopticalapplications,J.Am.Ceram.Soc.74(10)(1991)2343–2358. [2] X.H.Jin,L.Gao,J.K.Guo,Thestructuralchangeofdiphasicmullitegel studiedbyXRDandIRspectrumanalysis,J.Eur.Ceram.Soc.22(2002) 1307–1311.

[3] P. Mechnich, H. Schneider, M. Schmu¨cker, B. Saruhan, Accelerated reactionbondingofmullite,J.Am.Ceram.Soc.81(7)(1998)1931–1937. [4] J.C.Chen,C.S.Chen,H.Schneider,C.C.Chou,W.C.J.Wei,Atomistic calculationsoflatticeconstantsofmullitewithitscompositions,J.Eur. Ceram.Soc.28(2008)345–351.

[5] K.Yoshida,H.Hyuga,N.Kondo,H.Kita,Synthesisofprecursorfor fibrousmullitepowderbyalkoxidehydrolysismethod,Mater.Sci.Eng.B 173(2010)66–71.

[6] Y.Takao,T.Hotta,K.Nakahira,M.Naito,N.Shinohara,M.Okumiya,K. Uematsu,Processingdefectsandtheirrelevancetostrengthinalumina ceramicsmadebyslipcasting,J.Eur.Ceram.Soc.20(2000)389–395. [7] R.Greenwood,Reviewofthemeasurementofzetapotentialsin concen-tratedaqueoussuspensionsusingelectroacoustics,Adv.ColloidsInterface Sci.106(2003)55–81.

[8] S.B.Johnson,G.V.Franks,P.J.Scales,D.V.Boger,T.W.Healy,Surface chemistry–rheologyrelationships inconcentrated mineralsuspensions, Int.J.Miner.Process.58(2000)267–304.

[9] R.J.Hunter,ZetaPotentialinColloidScience, thirdprinting, Academic PressInc.,SanDiego,1988.

Fig.8. SEMmicrostructuresofslipcastspecimenssinteredat16508Cfor5h: (a)pH=4.5,(b)pH=8.9(IEP)and(c)pH=10.9.

Table1

Porosity(asvol%)valuesforthesinteredsamplesslipcastatdifferentpHs.

pH %Porosity

4.5 8.2

8.9(IEP) 17.1

10.9 7.5

R.Go¨renetal./CeramicsInternational38(2012)679–685 684

[10] J.N. Israelachvili, Intermolecular and Surface Forces, second ed., AcademicPressInc.,London,1992.

[11] T.M.Riddick,ControlofColloidStabilityThroughZetaPotential, Zeta-MeterInc.,NewYork,1968,pp.1–372.

[12] C. Aksel, The influence of zircon on themechanical properties and thermalshockbehaviourofslip-castalumina–mulliterefractories,Mater. Lett.57(2002)992–997.

[13] C.Aksel,Mechanicalpropertiesandthermalshockbehaviourofalumina– mullite–zirconiaandalumina–mulliterefractorymaterialsbyslipcasting, Ceram.Int.29(2003)311–316.

[14] C.Aksel,Theeffectofmulliteonthemechanicalpropertiesandthermal shockbehaviourofalumina–mulliterefractorymaterials,Ceram.Int.29 (2003)183–188.

[15] F.Temoche,L.B.Garrido,E.F.Aglietti,Processingofmullite–zirconia grainsforslipcastceramics,Ceram.Int.31(2005)917–922.

[16] L.B.Garrido,E.F.Aglietti,L.Martorello,M.A.Camerucci,A.L. Cava-lieri, Hardness and fracture toughness ofmullite–zirconia composites obtainedbyslipcasting,Mater.Sci.Eng.A419(2006)290–296. [17] N.M. Rendtorff,L.B. Garrido, E.F.Aglietti,Effectof theadditionof

mullite–zirconiatothethermalshockbehaviorofzirconmaterials,Mater. Sci.Eng.A498(2008)208–215.

[18] N.Rendtorff,L.Garrido,E.Aglietti,Mullite–zirconia–zirconcomposites: propertiesandthermalshockresistance,Ceram.Int.35(2009)779–786. [19] L.B.Garrido,E.F.Aglietti,Reaction-sinteredmullite–zirconiacomposites bycolloidalprocessingofalumina–zircon–CeO2mixtures,Mater.Sci. Eng.A369(2004)250–257.

[20] M.A.Camerucci,A.L.Cavalieri,R.Moreno,Slipcastingofcordierite andcordierite–mullitematerials,J.Eur.Ceram.Soc.18(1998)2149– 2157.

[21] Y. Hashi, M. Senna, Relationship between rheological properties of slurriesandporesizedistributionofcastandcompressedcompactsof mullite,PowderTechnol.83(June(3))(1995)187–191.

[22] E.Tkalcec,R.Nass,T.Krajewski,R.Rein,H.Schmidt,Microstructure andmechanicalpropertiesofslipcastsol–gelderivedmulliteceramics,J. Eur.Ceram.Soc.18(August(8))(1998)1089–1099.

[23] O.Burgos-Montes,R.Moreno,Colloidalbehaviourofmullitepowders producedbycombustionsynthesis,J.Eur.Ceram.Soc.27(2007)4751– 4757.

[24] C.O¨ zgu¨r,O.S¸an,PreparationofsphericalanddenseNa2O–B2O3–SiO2 glasspowderbyultrasonicspraypyrolysistechnique,J.Non-Cryst.Solids 356(2010)2794–2798.

[25] H.Tan,Y.Ding,J.Yang,Mullitefibrespreparationbyaqueoussol–gel process andactivation energyof mullitization, J.AlloysCompd.492 (2010)396–401.

[26] D.R.Treadwell,D.M.Dabbs,I.A.Aksay,Mullite(3Al2O3–2SiO2) syn-thesiswithaluminosiloxanes,Chem.Mater.8(1996)2056–2060. [27] B.Bagchi,S.Das,A.Bhattacharya,R.Basu,P.Nandy,Nanocrystalline

mullite synthesisata lowtemperature:effect ofcopperions,J.Am. Ceram.Soc.92(3)(2009)748–751.

[28] F.He,W.T.Petuskey,Lowtemperaturemullitecrystallizationin Al-and Si-alkoxide derived homogeneous gels, Mater. Lett. 63 (2009) 2631–2634.

[29] J.W.Salisbury,L.S.Walter,N.Vergo,D.M.D’Aria,Infrared(2.1–25mm) SpectraofMinerals, TheJohnsHopkinsUniversityPress,London,1991. [30] B.Stuart,W.O.George,P.McIntyre,in:D.J.Ando(Ed.),ModernInfrared

Spectroscopy,JohnWiley&SonsInc.,NewYork,1998,p.175. [31] A.L.Andrade,R.M.de,M.Turchetti-Maia,M.T.P.Lopes,C.E.Salas,R.Z.

Domingues,Invitrobioactivityand cytotoxicityofchemicallytreated glassfibers,Mater.Res.7(2004)635–638.

[32] R.Baranwal,M.P.Villar,R.Garcia,R.M.Laine,Flamespraypyrolysisof precursorasaroutetonano-mullitepowder:powdercharacterizationand sinteringbehavior,J.Am.Ceram.Soc.84(2001)951–961.

[33] A. Hamoudi,L. Khouchaf, C. Depecker,B. Revel, L. Montagne, P. Cordier,Microstructuralevolutionofamorphoussilicafollowing alka-li–silicareaction,J.Non-Cryst.Solids354(2008)5074–5078. [34] Dj.Janackovic, V.Jokanovic, Lj.Kostic-Gvozdenovic, D.Uskokovic,

Synthesisofmullitenanostructuredsphericalpowderbyultrasonicspray pyrolysis,Nanostruct.Mater.10(1998)341–348.

[35] B.Bagchi,S.Das,A.Bhattacharya,R.Basu,P.Nandy,Mullitephase enhancementinIndiankaolinsbyadditionofvanadiumpentoxide,Appl. ClaySci.47(2010)409–413.

[36] J.Temuujin,K.Okada,K.J.D.MacKenzie,Ts.Jadambaa,TheEffectof water vapour atmosphere on the thermal transformation of kaolinite investigatedbyXRD,FTIRandsolidstateMASNMR,J.Eur.Ceram. Soc.19(1998)105–112.

[37] L.S. Cividanes,T.M.B. Campos,L.A. Rodrigues, D.D.Brunelli, G.P. Thim,Reviewofmullitesynthesisroutesbysol–gelmethod,J.Sol-Gel Sci.Technol.55(2010)111–125.

[38] A.Beran,D.Voll,H.Schneider,Dehydrationandstructuraldevelopment ofmulliteprecursors:anFTIRspectroscopicstudy,J.Eur.Ceram.Soc.21 (2001)2479–2485.

[39] M.S. C¸elik, B. Ersoy, Mineral nanoparticles: electrokinetics, in: J.A. Schwarz,C.I.Contescu,K.Putyera(Eds.),EncyclopediaofNanoscience andNanotechnology,Marcel-DekkerInc.,NewYork,2004,pp.1991–2005. [40] G.A.Parks,Theisoelectricpointsofsolidoxides,solidhydroxides,and

aqueoushydroxocomplexsystems,Chem.Rev.65(1965)177–198. [41] R.D.Kulkarni,P.Somasundaran,Effectofpretreatmentonthe