Experimental study of the thermo-mechanical behaviour of alumina-silicate refractory materials based on a mixture of Algerian kaolinitic clays

refractory materials based on a mixture of Algerian kaolinitic clays

Article  in  Ceramics International · December 2011

DOI: 10.1016/j.ceramint.2011.05.095 CITATIONS 30 READS 102 5 authors, including:

Some of the authors of this publication are also working on these related projects:

Protection against corrosion of aluminum alloys in aqueous mediaView project

Analysis of structural multi-material performance in the presence of delaminationView project B. Amrane

University of Bouira, Algeria

6PUBLICATIONS   33CITATIONS   

SEE PROFILE

S. Djaknoun

University of Science and Technology Houari Boumediene

8PUBLICATIONS   67CITATIONS   

SEE PROFILE

Nadir Mesrati

National Polytechnic School of Algiers

56PUBLICATIONS   276CITATIONS   

SEE PROFILE

All content following this page was uploaded by Nadir Mesrati on 15 August 2020. The user has requested enhancement of the downloaded file.

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:

Experimental

study

of

the

thermo-mechanical

behaviour

of

alumina-silicate

refractory

materials

based

on

a

mixture

of

Algerian

kaolinitic

clays

B.

Amrane

,

E.

Ouedraogo

*

,

B.

Mamen

,

S.

Djaknoun

,

N.

Mesrati

a_{Grenoble-INP,Laboratoire3SR,France} b

UJF,Laboratoire3SR,France c

CNRSUMR5521Laboratoire3SR,BP53,38041Grenoble,cedex9,France d

LaboratoiredeMe´caniqueAvance´e(LMA),FGM&GP,USTHB,BP32,El-Alia,1611Alger,Algeria e

LaboratoireSciencesetGe´niedesMate´riaux(LSGM),EcoleNationalePolytechniqued’Alger,10avenueHassenBadiElHarrachALGER,Algeria Received2August2010;receivedinrevisedform19February2011;accepted20March2011

Availableonline27May2011

Abstract

Intheframeworkofageneralstudycarriedoutonthethermo-mechanicalbehaviourofrefractoryproductsmadefromAlgerianrefractory clays,silica–aluminabricksbasedonamixtureofhalloysitefromDjebelDebbaghandkaolinfromTamazertwerestudied.Thematerial’sinitial chemicalcompositionandphysicalpropertiesarereported.Uniaxialcompressionandthree-pointbendingtestswereperformedundertemperature conditionsfromroomtemperatureto12008Conrepresentativespecimenscutfromthebricks.Theevolutionofthematerial’sresistanceintension andcompressionasofitsmodulusofelasticitywithtestingtemperatureisreported.Thethermalexpansionofthematerialwasinvestigatedaswere opticalandSEMmicrographsatvarioustemperatures.Theevolutionofthematerialrheologicalbehaviourfromquasi-brittle,betweenroomand 9008C,toviscousathighertemperatureswasshownbybothbendingandcompressiontests.Thegeneralbehaviourofthematerialwithincreasing temperaturewas analysedthrough thevariousmicrostructure investigations and thepresenceof micro-crackinginducedby thedifferential expansionsofthemultiplephasesinpresence.

#_{2011PublishedbyElsevierLtdandTechnaGroupS.r.l.}

Keywords: E.Refractories;Alumina-silicate;Thermo-mechanicalproperties;Kaoliniticclays;Hightemperatures

1. Introduction

Alumina silicate refractories are extensively used in the metallurgical, ceramic and glass industries, representing an importantmaterialintheoverallrefractorymarket[1,2].These materials are made primarily from refractory and kaolinitic clays,whichareusedbothincrudeandgrogforms.Chamotte (fired refractory clay)transforms refractories intostable high temperaturecompounds,andlimitstheshrinkageandcracking during the drying and firing steps[3,4]. In the crude unfired form, the clays act as a binder. Silica–alumina refractory materialsgenerallyhaveaheterogeneousmicrostructurewitha largegrainsize distributionandhighporosity[5–7].

Themineralogicalcompositionofrefractorymaterialsbased on kaolinitic clays consists of mullite (3Al2O3 2SiO2) and

silica,bothincrystallineform(quartzorcristobalite)andinthe amorphous phase. The mullite phase results from the transformation of kaolinite [Al2O32SiO22H2O]during high

temperature thermal cycle [5]. Mullite-based ceramics have important chemical physical properties, including good resistance to corrosion and creep, are high temperature materials and exhibit a low thermal expansion coefficient [5,8–10].Theseadvantagesconfertomullite-basedrefractories the ability tobearsevereserviceconditionsencountered ina varietyofapplications. Thevitreous phaseresultingfrom the reactionsbetweenfreesilicaandalkalinesgovernsthe thermo-mechanicalbehaviourofalumina-silicaterefractorymaterials, limitingtheiruseathightemperatures[2,11].Thecontentofthe vitreousphasevariesaccordingtotherawmaterials,chemical compositionsandfiring temperature.

HallohysitefromDjebbelDebbagh(DD3)andkaolinfrom Tamazert(KT)arethemainrefractoryclaysdepositsminedin Algeriaformanufacturingofsilica–aluminarefractories.These products are commonly used in the ceramics industry to www.elsevier.com/locate/ceramint

Available online at www.sciencedirect.com

CeramicsInternational37(2011)3217–3227

*Correspondingauthorat:Laboratoire3SR,UMR5521,BP53,38041 Gre-noblecedex09,France.Tel.:+33476825297.

E-mailaddress:Evariste.Ouedraogo@grenoble-inp.fr(E.Ouedraogo). 0272-8842/$36.00#_{2011PublishedbyElsevierLtdandTechnaGroupS.r.l.} doi:10.1016/j.ceramint.2011.05.095

manufacture kilns, and as kiln furniture. However, these materials frequentlydeteriorate andwear becauseof thermal andmechanicalconstraints.Thishasanegativeimpactonthe environmentandonthe priceoffinished products.

Thelongtermaimofthisstudyistonumericallysimulateby finite element analysis the behaviour of bricks made from Algeriankaolinandrefractoryclayinserviceconditions.Todo this, it isnecessary todetermine thematerial behaviourwith increasing temperature, and to determine relevant equations and the evolution of the material’s characteristic parameters withincreasingtemperature.Isothermaltestsatvarioustesting temperatureswereconductedtoaccomplishthis.The present study,whichisanextensionofapreviousworkdevotedtothe optimisation of the formulation and the conditions of elaboration [12], presents a thermo-mechanical characterisa-tionofindustrialsilica–aluminarefractorybricksmadefroma mixture of DD3 clay and KT. The thermal and physical propertiesofmaterialsmadewiththeseclaysweremeasured, and the microstructurecharacterised by XRD andby optical andSEMobservations.

2. Materialsandexperimentalprocedure 2.1. Rawmaterials

The samples studied were made from a mixture selected among variouscompositions of DD3 clay and KT[12]. The clayswereusedincrudeandcalcinatedforms(chamotte).DD3 claywasextractedfromthequarryofDjebbelDebbagh,andis greyincolourandrichinalumina(Table1).Itsmineralogical compositionconsistsmainlyofhallohysite,kaolinite,andsmall quantities of quartzandnon-plasticminerals suchas goetite, calciteandplagioclase(Table2).Thegrainsizedistributionof

DD3 clay (Table 3) is predominately fine particles, which

explains its high plasticity index (23). Because of the fine particles,thisclayexhibitsapronouncedlinearshrinkageupon firingthatcauseslargecrackformationinfired products.KT clay isawhite-coloured kaolinextracted fromthe deposit at TamazertintheeastofAlgeria.Itisrichinquartz(Tables1and 2)andhasalowplasticityindex(8)[12,13].Becauseofitshigh

content in kaolin, KT clay forms mullite when fired at high temperature.Inthisstudy,KTclayismixedwithDD3clayto reduceshrinkagethatespeciallyleadstomicrocracking.The firedgrogaddedtothemixturewasproducedbyfiringDD3at 13508C,andhasthemaincharacteristicsshowninTables4and 5,respectively.

2.2. Experimentaltechniques 2.2.1. Preparationofthebricks

To obtain representative material characteristics, fired sampleswerecutfromcommercialprocessedbricksthatwere amixtureof30% wt.DD3,35% wt.KTkaolinand30%wt. chamotte.Theoptimumrawmaterialsformulationwasselected based on research conducted in a previous study [12], and represented the optimisationof physicochemical andthermal propertiesof sinteredsamples inthatstudy.

Fired bricks, were produced commercially following the stepsindicatedintheflowchartofFig.1.Thebestproduction conditionsfrompreviousstudies[12,13]wereused.Mixedraw materialsweremoistenedbyadding8%waterbyweight,then agedfor24h.Testbricks,hereafter indicatedasBSAA,were shapedbyuniaxialcompactionina100tfrictionpressunder 35MPapressure.Afterdryingat608C,thebrickswerefiredat 13508Cinatunnelkilninanoxidisingenvironmentfor33h. 2.2.2. Characterisation ofthe bricks

ThefiredpropertiesoftheBSAAbrickswereevaluatedand aregiveninTable 6.Theapparentdensity andopen porosity were characterisedbyArchimedes’ method,andthe absolute density was determined using a helium picnometer. As the major properties of the refractory materials are intimately connected to theirmineralogical composition, some samples were finelycrunched andanalysedby X-raydiffraction.The differentphasesformedafterfiringat13508Cwereidentified usingtheXPERTDATACOLLECTERsoftware. TheBSAA bricks microstructure was evaluated using scanning electron microscopy.

ThethermalbehaviourofBSAAsampleswascharacterised upto12008CaccordingtotheCHEVENARDmethod[14]by usingaDI24ADAMELLHOMARGYdilatometerataheating rateof 58C/min. The material’s refractorinesswas evaluated Table1

Chemicalcompositionsofhalloysite(DD3)andKaolin(KT)[12,13].

Oxides KTkaolin Halloysite

SiO2 69.86 39.87 Al2O3 19.29 38.36 Fe2O3 0.72 1.14 CaO 0.4 0.78 Na2O 0.13 0.48 MnO – 0.46 SO3 0.03 0.45 MgO 0.4 0.24 K2O 2.67 0.20 P2O5 – 0.02 TiO2 0.4 0.02 Cr2O3 – 0.01 Calc.loss 6.31 17.27 Total 100.21 99.69 Table2

Mineralogicalcompositionofhalloysite(DD3)andkaolin(KT)[12].

Clays Majorminerals Minorminerals

DD3 Hallohysite,kaolinite,quartz Calcite,goetite,plagioclase KTkaolin Kaolinite,muscovite,quartz Goetite,rutile,feldspaths

Table3

Particlesizedistributionofhalloysite(DD3)andkaolin(KT)[12].

Clays >1mm(%) 0.01a` 1mm(%) <0.01mm(%)

DD3 – 43.56 56.68

according totheDIN 51730 standard[15]bymeasuringthe melting pointusingmacroscopicexaminationduringheating. ThermalshocktestswereconductedaccordingtotheDIN 51068standard[16].Sampleswereheatedfor15minat9508C and immediately quenched in room temperature water for

3min.After drying for 2h at 1108C, the heating/quenching cyclewasrepeateduntilsamplesbrokeintoseveralfragments. 2.2.3. Thermo-mechanicaltests

Modulus ofrupture mechanicaltestingwasundertakenon high-temperaturemechanical equipmentspeciallybuiltatthe Table4

Mineralogicalcompositionofthechamotte[12].

Constituents Content(%)

Mullite 55

Quartz 10

Amorphousphase 32

Rutile 3

Fig.1. SchematicrepresentationofthemanufacturingprocessoftheBSAAbricks.

Table5

Grainsizedistributionofthechamotte[12].

Grainsizerange Content(%)

1.25–3mm 26.06

0.315a` 1.25mm 29.91

<0.315mm 44.03

3S-R Laboratory of the National Polytechnic Institute of Grenoble.Thisset-upconsistedof aZWICK Z400E electro-mechanicaltestingmachineequippedwithaPYROX16008C electrical furnace and an exterior differential displacement measuringdevice.Thisset-upcouldbeconfiguredtoperform uniaxial compression or three-point bending tests at various temperatures [17]. Compression and three-point bending specimens were made by sawing fired refractory bricks into desiredshapesusingadiamondsaw.Thecubicspecimensfor compressiontestswerethengroundtoensurethatthebearing faces were parallel within 30mm. The specimens were calibrated to an initial height of 40mm. Fig. 2 presents a viewofaspecimenplaced intheexperimental set-up. Three-pointbendingspecimensconsistedofbarswithasquare cross-section of 25mm2 and a length of 160mm. The distance betweenbottomsupportswas120mm.Thethreelocalareasof thespecimenbearingthetestsupportsforthree-pointbending weretreatedtominimiseinitialgeometricaldefects.However, theinitialpositioningofthethree-pointbendingapparatuswas known to influence initial deflection of some of the load-deflection curves.

The uniaxial compression and three-point bending tests were carried out at ambient temperature, 5008C, 7008C, 9008C, 10008Cand12008C.The loadingprocedureinboth compressionandthree-point bendingconsisted ofapplyinga preliminaryload(2KNincompression,0.05KNinflexion)at a rateof 0.5mm/min toeliminate positioning defects of the specimens. For tests conducted at elevated temperatures, a standardthermalcyclewasapplied:sampleswereheatedfrom

roomtemperaturetothetarget testtemperatureatthe rateof 2008C/h,followedbya2hholdatthetesttemperaturebefore the mechanicalloadingwas appliedinisothermalconditions. Aftertesting, samples were cooledto roomtemperature ata coolingrateof1508C/h.Inbothcompressionandthree-point loading, the mechanical loading consisted of imposing a 0.1mm/mindisplacement ratetothe upper plunger.Thetwo LVDTsensorsofthedifferentialdisplacementdevicesrecorded the specimen’s heightvariation in compression testsand the specimen’sdeflectioninbendingtests.Itthenbecamepossible to display force–displacement or force–deflection curves generated from each test. The stresses and strains in compressiontestsweredeterminedbythefollowingequations: s¼F

a2 (1)

e¼i

a (2)

whereF is the actualmeasured load,a isthe specimen side length and i is the average of the sensor indications. The evolutionofthefracturesurfacevs.temperaturewasexamined usingan opticalandscanning electronmicroscopy.

3. Resultsanddiscussion 3.1. Physicalproperties

ThephysicalcharacteristicsoftheDD3andKTtestbricks averaged andlabelled as BSAAbricks are listedin Table 6. Several studies carried out on refractory clay materials [3,6,11,18–20] showed thatthe sintered brick characteristics areinfluencedbythemineralogicalcompositionof clays,the chamotte content and the firing temperature. In general, porosity increases with the inert grain ratio, whereas firing shrinkage decreases. The firing temperature of brick made using the clays also influences properties, with higher temperatures enhancing densification by the formation of a liquidphase.Theliquidphaseisformedbyreactionsthatoccur

Fig.2. ViewofaprefiredBSAAsampleintheuniaxialcompressionset-upduringthepreparationphase. Table6

PhysicalcharacteristicsofBSAAbricksfor13508CprefiredBSAAsamples.

Characteristics Values Apparentdensity(g/cm3) 2.00 Absolutedensity(g/cm3) 2.60 Waterabsorption(%) 11.14 Openporosity(%) 22.34 Dryingshrinkage(%) 0.65 Firingshrinkage(%) 1.24

in the presence of free silica, alumina and alkaline, in accordance with the ternary SiO2–Al2O3–R2O system. DD3

clay is known for its high plasticity, which allows the introduction of a considerable quantity of inert matter (chamotte) to reduce the heating shrinkage and to ensure dimensional stability during thermal treatments. KT kaolin contains higher quartz content and a lower quantity of fine particlesthanDD3clay,aswellasalowerplasticityindex.The incorporationofKTkaolininthemixtureenablesadjustmentof the physicalpropertiesofthe firedmaterial.

Refractoriness of a material is strongly dependent on microstructural parameters such as grainsize andshape, the relativevolumeofsolidandvitreousphases,theviscosityofthe vitreousphase,andsampleporosity[14,21].Refractorinesscan beevaluatedbyexaminingtheevolutionofthemorphologyof test samples with increasing temperature under their own weight(withoutanyotherload).Refractorinessinthispaperis definedasthetemperaturecorrespondingtothemomentwhena material begins to lose its shape (melting point).As can be clearlyseeninthephotospresentedinFig.3,thesamplestested heremaintainedtheirshapewithoutundergoingany deforma-tion upto16008C.

ThemaincrystallinephasesidentifiedbyXRDaftersamples were fired at 13508C were mullite, quartz, cristobalite and rutile(Fig.4).TheSEMmicrographofFig.5clearlyshowsthe mullite phase identified by its needle like shape. It is well known thatthe reactions between free silicaandalkaline, in

particular potassium oxide that is present in excess in the Tamazertkaolin,leadtotheformationofavitreousphasethat governs the thermo-mechanical behaviour of the materials. Mullite results from the transformation of phyllosilicate mineralssuch as kaolinite,hallohysite andmuscoviteathigh temperaturesaccordingtothefollowingreactions [1,19]:

Kaolinite½Al2O32SiO22H2O !

450 _C

Metakaolinite½2Al2O33SiO2 ! 980 _C

SiO2þ2Al2O33SiO2 !

11001200 _C

SiO2þMullite½3Al2O32SiO2

Muscovite½K2O 3Al2O36SiO22H2O1200!CMullite þSilica-potassiumrich glassy

A previousstudy [3]showed that excesssilica inthe raw materialsmixturepartiallytransformstocristobalitebeginning at13008C,andthenbecomestotallyamorphousafterfiringat 14008C. The various silica phases that compose the fired material do not have the samethermal expansion coefficient Fig.3. Refractoriness(deformation)ofBSAAsamplesat:(a)208Cand(b)16008C.

Fig.4. XRDpatternsofBSAAbrickafterfiringat13508C.

Fig.5. SEMmicrographofaBSAArefractorybrick(halloysiteandKTkaolin brick)firedat13508C(M:Mulliteneedles;Q:Quartz;G:Amorphousphase;P: Pores).

Fig.6. MicrostructureofaBSAAbrickfiredat13508C. -0,4 -0,3 -0,2 -0,1 0 0,1 0,2 0,3 0,4 0,5 0,6 1200 1000 800 600 400 200 0 Temperature (°C) DL/L0 (%) α = 5 10-6 K-1

Fig.8. Thermalexpansioninairmeasuredfromroomtemperatureto12008C for13508CprefiredBSAAsamples.

valuedependingonthefiringtemperature.Furthermore,thea– b quartz transformation at 5708C is accompanied by a significantvariationinvolume[1,14].Thisleadstoamarked crackdevelopmentaroundthegroggrains(chamottegrains)in the matrix due to large tensile hoop stresses during sample cooling.AscanbeclearlyseeninthephotoofFig.6,cracksand elongatedporesareobservedatinterfacesbetweengroggrains andthematrix.Fig.7illustratestheheterogeneouscharacterof the sample microstructure byoptical micrographs,indicating

0 20 40 60 80 100 120 7 6 5 4 3 2 1 0 Axial strain (%)

Axial stress (Mpa)

25 °C 0 20 40 60 80 100 120 7 6 5 4 3 2 1 0 Axial strain (%)

Axial stress (Mpa)

500 °C 0 20 40 60 80 100 120 7 6 5 4 3 2 1 0 Axial Strain (%)

Axial stress (Mpa)

700 °C 0 20 40 60 80 100 120 7 6 5 4 3 2 1 0 Axial strain (%)

Axial stress (Mpa)

900 °C 0 20 40 60 80 100 120 7 6 5 4 3 2 1 0 Axial strain (%)

Axial stress (Mpa)

1000 °C 0 20 40 60 80 100 120 7 6 5 4 3 2 1 0 Axial strain (%)

Axial stres (Mpa)

1200 °C

Fig.9. Stress–strainuniaxialcompressiondiagramsfor13508CprefiredBSAAsamplesatdifferenttestingtemperatures:25,500,700,900,1000,12008C. Table7

MeanthermalexpansioncoefficientofBSAAbricksprefiredat13508Cvs. temperature.

Temperature(8C) Thermalexpansion(106/8C)

200 6.66

400 5.78

600 5.68

800 5.51

1000 5.91

the fractured surface of samples at the different test temperatures.

Dilatometric measurements up to 12008C in air were conductedonsamples cutfromthe prefiredBSAAbricks. As shown in Fig. 8, the dilatometric curve of the material is characteristicofasilicaaluminarefractorywithhighalumina content (greater than 40%) [1]. Mean thermal expansion coefficientsduringheating(a)werecalculatedandarereported

inTable7.Theevolution of(a)was notedtohaveamarked

increase from 208C to 2008C then through 3008C, that indicatedthe presenceof cristobalite,whose a–b transforma-tionoccursbetween1708Cand3008C.Consistentexpansion behaviour was observed from 300 to 12008C, leading to stability in the thermal expansion behaviour of the bricks. Becauseofthepermanentvolumechangeduetothecristobalite phase change, the cooling rate must be adjusted when the material is used as kiln furniture. In general, a values were

similartothoseofcorrespondingcommercialaluminasilicate refractory bricks (ranging from 4.5 to 8.51068C1) [11,22].

Thermalshocktestingwasconductedonwholebricks.After samplesweresubjectedto15thermalshockcyclesofheatingat 9508C followed bywater quenching, test brickdamage was quantifiedbyevaluatingopenporosity.Theincreaseinporosity wasnotgreaterthan2%foranytestsample.Thisconstitutes satisfactorybrickbehaviourwhereabrupttemperaturechanges mightoccur.

3.2. Thermo-mechanicalproperties

Fig. 9 plots the stress/strain curves of compression tests carriedoutatvarioustemperaturesonfiredBSAArefractories. Thecurveshavebell-shapeformscharacteristicofquasi-brittle materials[23,24].Duringtesting,materialsendureamaximal stress(resistance),andthenlosethosestressesbecauseofhigh temperature material softening during heating. Successive lossesofresistanceareobservedafterthepeakat7008C.The strain reached when the stress is amaximum (strain at peak stress) indicates the ability of a material to deform before collapsing,thuscorrespondingtosomethinglikeductility[25]. Fig. 10 presents the maximal stress reached at the various testingtemperatures,whicharevaluesaveragedfromthreetests at each temperature. The maximal stress increases with increasing temperature up to a maximum value at 9008C, then decreases at higher temperatures. Fig. 11 displays the evolutions with temperature of material’s elasticity modulus and strain values recorded at peak stress. The modulus of elasticity (Young’s modulus) has been calculated using the stress–straincurvebymeasuringtheslopeofthelinearpartof thecurvefrom10to30%ofthemaximalstress.Therecorded valuesaretheaveragevaluesofthreetestsateachtemperature. TheYoung’smodulusisconstantatlowtemperatures,reachesa maximumfrom700to9008C,andthendecreasessharplywith increasingtemperature.Thisevolutionevokesthatofalumina refractoryconcrete[24],andisslightlydifferentfromresultson kaolinitic clay mixtures reported by Kolli [11], where the Young’s modulus was constant at low to intermediate temperatures before decreasing at higher temperatures. The Young’smodulusvaluesinFig.11arelowcomparedtothoseof puredenseceramics,whichcanbeexplainedbythepresenceof porosity and the multiphase composition of other phases presentinthesamples.Partofthisdifferencecouldalsobedue toanoverestimationofthespecimenheightvariationduetothe specimen residual parallelism defects, leading to an under-estimation of the Young’s modulus. However, the values measuredareofthesameorderandareconsistentwithnumbers reportedincomparable studies. Forinstance,an 8GPavalue wasobtainedatroomtemperatureinthisstudyinsteadof the 9GPaobtainedbyKolli[11]atthesamefiringtemperaturefor comparable mixtures.Hence,the modulusvaluesobtainedin this study seem to be low, as they are less than 10GPa at temperaturesunder6008C,but similartovaluesobtainedby others at higher temperatures. The sharp decrease in the

0 20 40 60 80 100 120 1500 1250 1000 750 500 250 0 Testing temperature (°C)

Maximal axial stress (MPa)

Fig.10. Evolutionofthematerialcompressionstrengthvs.testingtemperature for13508CprefiredBSAAsamples.

0 2 4 6 8 10 12 14 1400 1200 1000 800 600 400 200 0 Testing temperature (°C)

Elastic modulus (GPa)

0,0 2,0 4,0 6,0 8,0 10,0

Strain at maximal stress (%)

Elastic modulus Strain at maximal stress

Fig.11. Dependenceofthemeanvalueofthematerial’sYoung’smodulusand thestrainatmaximalcompressivestressvs.testingtemperaturefor13508C prefiredBSAAsamples.

modulusofelasticityabove9008Cinthisstudyissimilartothe workbyKolli.

The decrease of the modulus of elasticity at high temperature and the stablevalues at low temperaturecan be explained. At room temperature (after firing), the weak modulus is due to internal damage induced by the thermal expansionmismatchesbetween phasesthatoccurs duringthe

cooling. When the temperature increases to 9008C, the material absorbs the initial damage because the residual stressesduetothermalexpansionmismatchdecreasegradually, leadingtoaprogressiveincreaseinYoung’smodulus[22].The viscous phase begins to appear at 9008C and above, accelerating the material deformation, leading to a decrease of theYoung’smodulus.

0 0,2 0,4 0,6 0,8 1 1,2 0,8 0,6 0,4 0,2 0 Displacement (mm) Recorded force (KN) 25 °C 0 0,2 0,4 0,6 0,8 1 1,2 0,8 0,6 0,4 0,2 0 Displacement (mm) Recorded force (KN) 500 °C 0 0,2 0,4 0,6 0,8 1 1,2 0,8 0,6 0,4 0,2 0 Displacement (mm) Recorded force (KN) 900 °C 0 0,2 0,4 0,6 0,8 1 1,2 1,5 1 0,5 0 Displacement (mm) Recorded force (KN) 1000 °C 0 0,2 0,4 0,6 0,8 1 1,2 1,5 1 0,5 0 Displacement (mm) Recorded force (KN) 1200 °C

Fig.12. Force–displacementcurvesduringthree-pointbendingtestsfor13508CprefiredBSAAsamplesatdifferenttestingtemperatures.

Fig.11also presentsthe strain reachedatmaximalstress. Thisparameterreachedaminimumat7008C,followedbyan increasing trend at higher temperatures. The minimum observed at 7008C is characteristic, and seems to be a consequenceofahealingofthematerialmatrix,leadingtothe decreaseofthe strainatmaximalstress.

Fig. 12illustrates the resultsof the three-point bendtests performed at various temperatures, which were similar to results for compression tests, particularly withregard to the changeofbehaviourbetween700and9008Ccharacterisedby an increase of the strain before failure occurs. As we are interestedinstudyingtheoccurrenceofdamageinthematerial, wefocusedonmaterialschangesinsamplesthatcouldexplain observedphysicalpropertychangesinthedesiredtemperature ranges. Bending test specimens were cut from commercial processed bricks usinga diamond saw, aprocess that might generate surface micro-cracks in the material. Although the specimen contact areas were fixed, defects in the sample preparation may have led to an overestimation of deflection measurements, causing tensile strength measured by three-pointbendingteststobelower.Thesechangesarenotthought tohaveamajorimpactonresultsbecauseourinterestfocused on changes in strength of the material behaviour with increasing temperature. If the formation process had an influence, it would have been the same for all specimens. The occurrence of viscous phases in samples at high temperatures,however,tendstominimisegeometricaldefects. From 20 to 5008C, possibly as high as 7008C, material behaviour is brittle, with an abrupt failure occurring once tensileresistancewasreached(Fig.12).Apost-peakdomainis present,withthepost-peakstress gapsevolvingdifferentlyat eachtesttemperature.At9008C,forinstance,theevolutionof the load before its peak is strongly non-linear, contrary to observations at lower temperatures. However, once the peak wasreached,thefailurewassevere,andthematerialresistance tofailure falls toavery lowvalue, contrary tothe relatively progressivedeclineatlowertemperatures. Itseemsthatonce the macro-crack initiates at 9008C, it propagates directly throughthematerial,whereasatlowertemperaturesitremains constrained,leadingtoaprogressivelossofresistance.Hence,

intermsof microstructureinterpretation,the materialfracture seemstobeinter-granularatlowandintermediatetemperatures, ratherthantrans-granularat9008C(Fig.7).Thisbehaviouris consistentwithexistingmaterialmicrostructure:inter-granular micro-crackscreatedbythepreviouscoolingphasearepresentat low temperatures, tendingtobe reduced andeveneliminated withincreasingtemperatureduetothermalexpansion.Forhigher temperatures(1000–12008C),theload–displacementcurvehas anopenbell-shapeformthatindicatesthepresenceoftheviscous phase[11,23,25].At12008C,thematerialresistancebecomes verylow,asthecurveinFig.12indicates.

Fig.13representsthevariationofthe maximumloadwith the testing temperature superimposed with the evolution of compressivestrengthvs.temperature.Thegeneraltrendofthe two curves is similar, with maximum of both strengths at 9008C.Thisbehaviourindicatesthatmechanismsdrivingthe coalescenceofmicro-crackstoformmacro-cracksaresimilar, andareofthesamenatureintensionandincompression.Itis importanttonotethatnonewphasesappearduringheatingat temperatureslowerthanthefiringtemperature.

4. Conclusion

Silica–alumina refractory materials made from such halloysiteandkaolinclaysarewidelyusedinthemanufacture ofceramicproducts(askilnfurnitureandbricks).Becauseuse ofthesematerialsislimitedathightemperature,knowledgeof their thermo-mechanical and deformation properties in fired refractorymaterialsisnecessary.Thus,mechanicaltestsrelated tomodulusofrupture andcrushingstrength werecarriedout from 258C to 12008Cusing Algerian halloysite and kaolin clays,aswellasastudyofmicrostructurechangesoccurringin thesecompositions.Fromthesestudies,thefollowing conclu-sionsweredrawn:

The material’s thermo-mechanicalbehaviour, evidencedby the compression tests and confirmed by the bending tests carriedoutinthetemperaturerangeof25–12008C,evolved fromquasi-brittledamageableatroomtemperatureto visco-plastic damagebetween 900and10008C.

The globaltrendof thematerial compressionresistancevs. temperature is similar to that of tensile resistance: the maximalstressincreaseswithtemperatureuptoamaximum value at9008C,thendecreases athighertemperatures. OpticalandSEMmicrographsof thefracturesurfacesafter

heating at different temperatures revealedhighly heteroge-neousmicrostructures.Specialattentionshouldbepaidtothe amount of kaolin clay, in a composition which is highly siliceous.Indeed,allotropictransformationsoffreesilicaare wellknowntoinduceinterfacialcracksleadingtoadecrease insomemechanicalproperties.Thisbehaviourmayexplain the weakening of the Young’s modulus values at low temperatures.

Dilatometrictestsshowaslightandregularvariationof the thermalexpansioncoefficientbetweenof 300and11008C, whichhelpstoprovideagoodthermalshockresistancetothe material. 0 20 40 60 80 100 120 1400 1200 1000 800 600 400 200 0 Testing temperature (°C)

Bending test maximal load (daN) ₀

20 40 60 80 100 120 140

Compression strength (MPa)

Maximum bending load Maximum compression stress

Fig.13. Evolutionofthemaximalbendingloadand maximalcompressive stressvs.testingtemperaturefor13508CprefiredBSAAsamples.

References

[1] G.Aliprandi,Mate´riauxre´fractairesetce´ramiquestechniques,Septima, Paris,1979,pp.1–612.

[2] P.Lapoujade,Traite´ pratiquesurl’utilisationdesproduitsre´fractaires,H. Vial,Vial,Paris,1986,p.236.

[3] G.N.Djangang,A.Elimbi,U.C.Melo,G.L.Lecomte,C.Nkoumbou,J. Soro,J.P.Bonnet,P.Blanchart,D.Njopwouo,Sinteringofclay–chamotte ceramiccompositesforrefractorybricks,Ceram.Int.34(2008)1207–1213. [4] F.A.C.Minlheiro,M.N.Freire,A.G.P.Silva,J.N.F.Holanda,Densification behaviourofredfiringBraziliankaoliniticclay,Ceram.Int.31(2005) 757–763.

[5] M. Lalithambika, Fireclay bricks from Keralaclays, Interceramics 6 (1988)17–19.

[6] M.Dondi,M.Marsigli,I.Ventury,Microstructureandmechanical prop-ertiesofclaybricks:comparisonbetweenfastfiringandtraditionalfiring, Br.Ceram.Trans.98(1)(1999)12–18.

[7] C.N.Djangang,A.Elimbi,U.C.Melo,G.L.Lecomte,C.Nkoumbou,J. Soro,J.Yvon,P.Blanchart,D.Njopwouo,Refractoryceramicsfromclays ofMayouomandMvaninCameroon,Appl.ClaySci.39(2008)10–18. [8] K.S. Madziasni,L.M.Brown, Synthesisandmechanicalpropertiesof stoichiometricaluminiumsilicate(mullite),J.Am.Ceram.Soc.55(11) (1972)548–552.

[9] M.A. Sainz,F. Serrano, J.M. Amigo, J.Bastida, A.Caballero,XRD microstructural analysis of mullites obtained from kaolinite–alumina mixtures,J.Eur.Ceram.Soc.20(2000)403–412.

[10] I.Ganesh,J.M.F.Ferreira,Influenceofrawmaterialtypeandoftheoverall chemicalcompositiononphaseformationandsinteredmicrostructureof mulliteaggregates,Ceram.Int.35(2009)2007–2015.

[11] M.Kolli, M.Hamidouche, G.Fantozzi,J.Chevalier,Elaborationand characterizationofarefractorybasedonAlgeriankaolin,Ceram.Int.33 (2007)1435–1443.

[12] B.Amrane,Etuded’unprojetdeproductiondebriquesre´fractaires silico-alumineusesa` basedematie`respremie`reslocales, Me´moired’Inge´nieur, Universite´ deBoumerde`s,Alge´rie,1993.

[13] A. Chakri, Elaboration et caracte´risation des briques re´fractaires de chamottea` partirdesmatie`respremie`reslocales,The`sedeMagiste`re, Universite´ deSe´tif,Alge´rie,1995,pp.125–132.

[14] C.Jouenne,Traite´ dece´ramiquesetdemate´riauxmine´raux,Ed.Septima, Paris,1975,p.268.

[15] NormeDIN51730. [16] NormeDIN51068.

[17] S.Djaknoun,E.Ouedraogo,A.AhmedBenyahia,Fracturethoughnessof high performance concrete on three-point bending notched beam at elevatedtemperature,Adv.Mater.Res.89–91(2010)159–164. [18] M.Hajjaji,S.Kacim,M.Boulmane,Mineralogyandfiringcharacteristics

ofaclayfromthevalleyofOurika(Morocco),Appl.ClaySci.21(3/4) (2002)203–212.

[19] M.Hamidouche,M.A.Madjoubi,K.Loucif,H.Osmani,M.Kolli,M. Gonon,G.Fantozzi,Processingandcharacterizationofrefractorymadeof Algeriankaolin,in:ProceedingsoftheEighthConferenceonCeramics, INSAdeLyon,France,3–5September,2002.

[20] M.Hamidouche, M.A.Madjoubi, K.Loucif, H. Osmani, M.Kolli, M.Gonon,G.Fantozzi,Etudedelatransformationde 3nuances de kaolinenfonctiondelatempe´rature,Silic.Indus.65(11–12)(2001) 119–124.

[21] M.J.Ribeiro,J.A.Labrincha,Propertiesofsinteredmulliteandcordierite pressedbodiesmanufacturedusingAl-richanodisingsludge,Ceram.Int. 34(2008)593–597.

[22] B.Coope,AnIntroductiontoRefractories,RawMaterialsforthe Refrac-toriesIndustry, 2nded., IndustrialMaterials,1986.

[23] J.Lemaitre,J.L.Chaboche,Me´caniquedesmateriauxsolides, 2emeed., Dunod,Paris,1996.

[24] E.Ouedraogo,N.Prompt,Hightemperaturemechanicalcharacterisation ofanaluminarefractoryconcreteforBlastFurnacemainthrough.PartII: materialbehaviour,J.Eur.Ceram.Soc.28(2008)2867–2875. [25] M. Roosefid, E. Ouedraogo, N.Prompt, A.Miras, Caracte´risationet

mode´lisationducomportementthermome´caniquedebe´tonsre´fractaires silico-alumineux,in:17e`meCongre`sFranc¸aisdeMe´canique(CFM2005), Troyes,Septembre,2005.

B.Amraneetal./CeramicsInternational37(2011)3217–3227 3227

View publication stats View publication stats