e for n -thermal microwave e ingle multimode h c /m spower used in a molybdenum disilicide...

International Microwave Power Institute 42-2-47

Jon Binner, Bala Vaidhyanathan, Jianxin Wang, Duncan Price, Mike Reading

IPTME,LoughboroughUniversity,Loughborough,UKDept.ofChemicalSciences,UniversityofEastAnglia,Norwich,UK

[email protected]

Clear evidence for the microwave effect has been observed during experiments in which a variety of materials have been heated using experimental systems that allowed both conventional and conventional-microwave hybrid heating. A hybrid single mode cavity has been used to investigate the microwave effect during phase changes in silver iodide, barium titanate and benzil, whilst a hybrid multimode cavity has been used to investigate the microwave effect during sintering and annealing of a range of ceramic materials with different dielectric properties. Although evidence for the microwave effect was not found in every case, where it was found the results could not be explained purely in terms of temperature gradients within the materials.

Submission Date:29November2007 Acceptance Date: 7May2007

Publication Date: 25July2008

EvidEncE for non-ThErmal microwavE EffEcTs Using singlE and mUlTimodE hybrid

convEnTional/microwavE sysTEms

Keywords: Microwave effect, hybrid microwave systems

INTRODUCTION

Many investigators have reported unexpectedeffects resulting from the use of microwaveradiationasanalternativeenergysourceduringtheprocessingofmaterials.Thishas includedapparent evidence for acceleratedkinetics forarangeofprocessesinceramic,polymericandorganic systems [Binneret al., 1995;Bochet al.,1992;Booskeet al.,1991;Fathiet al.,1991;Giguere,1992;Janney&Kimrey,1991;Janney,Kimreyet al., 1991;Katzet al., 1991;Lewis,1992;Willert-Porada et al., 1992]; enhancedsinteringofceramicpowdercompacts,includinglowersinteringtemperatures[Janney,Calhounet al.,1991;Janney&Kimrey,1988];andreducedactivation energies [Janney & Kimrey, 1988;Janney & Kimrey, 1991; Janney and Kimrey

et al.,1991;Lewis,1992]. It isnowgenerally,thoughbynomeansunanimously,acceptedthata‘microwaveeffect’exists.Oneofthereasonsfor theremaininguncertainty is the inability tovary theenergy sourcewithout simultaneouslyaffecting awide rangeofothervariables.Forexample,whilstmicrowaveheatingexperimentsare performed in a microwave applicator thecorresponding conventional experiments aretypicallycarriedoutinaseparate,radiantfurnaceof very different specification. Research has now been performed ona number of materials systems using twosystemsthatwerecapableofheatingmaterialsusing either conventional, microwave or thesimultaneouscombinationofthetwo(knownashybrid)heating.Inbothcasestheenergysourcecouldbevariedwithoutaffectingawiderangeofothervariables. In the first series of experiments, a hybrid

42-2-48 Journal of Microwave Power & Electromagnetic Energy ONLINE Vol.42,No.2,2008

singlemodecavityhasbeenusedtoinvestigatethemicrowaveeffectduringphasechangesinsilver iodide, barium titanate and the organiccompoundbenzilviaexaminationofthethermalproperties, including the specific heat capacity. ThisfollowsworkbyRobbet al.[2002]usingtemperature resolved in situ powder X-raydiffractionthatindicatedanomalousbehaviourinthephasetransitionforsilveriodide.Underthe influence of conventional heating thestructuraltransitionwasdetectedattheexpectedtemperature,howeverwhenheatedby2.45GHzmicrowave radiation the transition occurredsome50°Clowerthanexpected. In the second series, a hybrid multimodecavityhasbeenusedtoinvestigatethemicrowaveeffectduringsinteringandannealinginarangeof ceramic materials with different dielectricproperties,viz.alumina,zirconiaandzincoxide.Inbothcases,itwaspossibletoheatthesamples

using exactly the same heating profile but with varyingamountsofmicrowaveenergy.

EXPERIMENTAL

Single mode cavity

Aschematicdiagramofthehybridsinglemodecavity is shown in Figure 1; it has been de-scribedindetailelsewhere[Binneret al.,2005].Arectangularwaveguide(notshown)wasusedtolaunchmicrowaveradiationfromacontinu-ouslyvariable500Wmagnetronoperatingat2.45GHzintoacylindricalcavitycontainingthespecimenholderatitsaxis.Motorisedchokesatthetopandbottomofthecavitywereadjustedsothat the E-field within the cavity (measured by loopantennasorthogonaltothespecimen)wasmaximised.Ancillarytuningbyamanual3-stubtunerinthelaunchsectionwasemployedsoas

Figure 1. Schematic diagram of single mode test cavity showing arrangement for heating sample and positions of temperature sensors.


to minimise reflected power. By these means the cavitycouldbeoperatedinaTE111modewiththe maximum field intensity at the locus of the sampleposition. Conventional heating of the sample wasachievedbypassingcompressedair,heatedbya750Wprocessgasheater,aroundthespecimenholder.The temperature of the specimen wasmonitored either by a fluoroptic thermometer (Luxtronmodel790)orbyathermalimagingcamera (FLIRSystemsThermovision®A40).The fluoroptic thermometer was calibratedaccordingtomanufacturer’sinstructionsusingan ice-water bath as a single reference point.The thermal imaging camera was calibratedfor temperature by measuring the emissivityof the sample held at known temperatures inaconventionalfurnace.Thepowersuppliedtotheheaterandmagnetron,inputairtemperatureandsampletemperaturewererecordedbyA/Dconverters (PicoLogADC-16 and TC-08).A temperature controller (Eurotherm 2408)was used to control the power to the heatingsystem which could either be operated withpure conventional heating, pure microwaveheating or a hybrid fashion with fixed amounts ofmicrowaveenergybeingsuppliedinadditiontoautomaticcontrolofsampletemperatureviathesurroundingairtemperature. The system was used to examine phasechanges in three different materials, viz.silver iodide (AgI), barium titanate (BaTiO3)and benzil (diphenylethanedione, C6H5CO)2),via examination of the thermal properties ofthesematerials.TheAgIandBaTiO3powderswere made into pellets, the former of ~83%of theoreticaldensityand the latter~60%,byuniaxial die pressing, whilst the benzyl waspre-meltedbyheatingthetubeinairsothatineach case the fluoroptic thermometer made good contactwiththespecimens.Thesampleswere

heatedusingbothpureconventionalandhybrid(microwaveplusconventional)heating.

Multimode cavity

Several series of pellets measuring 13 mmdiameter by 5 mm thickness were made bydie pressing a range of >99% purity ceramicpowdersthatvariedintermsofbothchemicalcompositionandparticle size (Table1).Eachpellet was sintered at one of two differenttemperatures for one hour using exactly thesame temperature/time profile whilst varying the fraction of microwave and conventionalpowerusedinamolybdenumdisilicideelement-basedhybridmicrowavefurnace[Wanget al.,2006].The latterwas capableofoperating inpureconventionalormicrowave/conventionalhybrid mode1; the microwave frequency was2.45GHzandupto2kWofmicrowavepowerwasavailableifrequired. Foreachsinteringrunwithinagivenseriesthe microwave power level was fixed at agivenvaluebetween0and1000W,in200Wincrements,andtheconventionalpowervariedtoprovidethesamesinteringschedule.Allthesampleswereheldat500oC for1hourat thebeginningofeachsinteringruntoensurethatthermal equilibrium was reached before thetemperature was increased to the final value chosen.Thisminimisedthemagnitudeofanytemperaturegradientsdeveloped.Thesinteringtemperaturesweredeliberatelyselectedtoavoidachieving full densification so that differences couldbeobservedasafunctionofthemicrowavepowerandpowderparticlesizeused.Throughoutthework,aminimumofatleast3sampleswassinteredforeachdatapoint. Inafurtherseriesofexperiments,samplesweresinteredforonehouratseveraldifferenttemperatures, from where little densification was

1 Pure microwave heating was not possible with this system since the molybdenum disilicideelements absorbed a small but finite amount of microwave energy that was then radiated back onto thesamples.


achieved through to almost full densification, using just pure conventional and hybridheating involving1kWofmicrowavepower.Thisallowedadirectcomparisonoftheeffectof sintering temperature on any enhancementcausedbytheuseofmicrowaves.Onceagain,in each case the temperature/time profile was identicalforeachpairofsamplessintered. For the annealing experiments submicronzinc oxide disks with >98.4% of theoreticaldensity and an average grain size of ~0.6 μm wereused.Thediskswereannealedinthehybridfurnacefor1,3or5hrsat900,1000,1100,1150or1200oCusingeitherpureconventionalheatingorhybridheatinginvolving1kWofmicrowavepower and a balance of conventional heatsufficient to achieve an identical temperature/time profile, as described before. Throughoutthework,thetemperaturewas

controlled using a Luxtron optical fibre ther-mometer (M10, Luxton Corporation, USA)insertedintoanarrowholedrilledintothecentreof the flat face of each sample such that the tem-peraturewasmeasuredatthecentreofthebody,asindicatedinFigure2.Thetemperatureattheedgeofthesampleswasalsomonitoredusingasecondopticalthermometer,whichallowedthetemperaturegradientacrosseachsampletobedetermined.Theradiationfromtheheatingele-mentswasshieldedviatheuseofanon-metallic,very low dielectric loss ceramic tube aroundthe optical thermometers to achieve accuratetemperaturemeasurement.Theaccuracyofthemeasurementswasindependentlyassessedusingaseparateseriesofmeasurementsmadeofthemeltingpointofvanadiumpentoxideandfoundtobe±<3oC[Binneret al.,2003].Inaddition,athermalimagingcamerawasusedtoobserve

Table 1. Average particle size, source of the powders, typical microwave absorbtion (at room temperature and 2.45 GHz), the green densities achieved after uniaxial pressing and

sintering temperatures used.

Powder Average particle size Source Tan δ1× 102 Green density/%

Sintering temperature / oC

ZnO

‘Submicron’~0.15 μm SigmaAldrich 4.8 ~58% 680&780

‘Micron’~0.92 μm SigmaAldrich 8.4 ~58% 780&900

3-YSZ

‘Nano’~20nm MEL 0.6 ~46% 900&1000

‘SubmicronD’0.24 μm

HSY-3U,Daiichi - ~45% 1100&1200

‘SubmicronM’0.12 μm MEL 1.9 ~44% 1220&1260

10-YSZ ‘SubmicronM’0.25 μm MEL 0.4 ~45% 1220&1260

Al2O3

‘NanoS’~26nm SigmaAldrich 0.5 ~50% 1400

‘NanoIH’<15nm

Produced‘in-house’ - ~35% 1180&1260

1.Loosepowdersmeasuredat2.22GHzand300oC.


the temperaturegradientsacross thefaceofanumberofsamplesduringthesinteringruns. Green sample densities were determinedbysimplemeasurementof thesampledimen-sions and mass, whilst the sintered densitieswere measured by theArchimedes methodusing mercury.Themicrostructures of all thespecimenswereexaminedonbothfractureandsliced,polishedandchemicallyetchedsurfacesusing field emission gun scanning electron mi-croscopy (FEG-SEM), the latter allowing theaveragegrainsize tobedeterminedusing thelinearinterceptmethod.

RESULTS AND DISCUSSION

Single mode cavity

Noeffectonthephasetransformationwasob-servedwiththeuseofmicrowaveenergyforei-therbenzilorBaTiO3.However,Figure3showsresultsobtainedonheatingsilveriodideat1°Cmin-1exposedtoincreasinglevelsofaconstantbackgroundofmicrowavepower.Thedifferencebetweentheairtemperatureintheproximityofthe sample and the sample temperature (ΔT) isplottedagainstthesampletemperatureinananalogous manner to a DifferentialThermalAnalysis (DTA) measurement [Chen & Dol-limore(1995)].Withnomicrowavepower,thecurveshowsthecharacteristicendothermicpeakaccompanying the normal phase transition at147°C.Inthepresenceofmicrowaveenergythe

shapeofthecurvechangesduetotheincreasedcoupling of the α-phase with microwaves com-pared to the β-phase, resulting in a sharp drop in conventionalheatingpowerrequiredtomaintaintheprogrammetemperature.Itappearsthatthephasetransitionisshiftedtolowertemperaturesunder the influence of increasing levels of mi-crowaveenergy. Ratherthanusealinearrisingtemperatureprofile, the sample temperature could be pro-grammedtooscillatebetweentwotemperatures.Theratiooftheamplitudesofthespecimenandairtemperaturesisthenproportionaltotheheatcapacityofthetestspecimen.ThisisthebasisofACcalorimetry[Kraftmaker,2002]anddataforsilveriodideobtainedintheabsenceofanymicrowave field using the present apparatus is showninFigure4(a).Inthisparticularmeasure-ment,thespecimentemperaturewasoscillatedby±2°Caboutameanvalueoveracycletimeof 2 minutes for a period of 10 minutes andthenthemeanvalueincrementedby2°Csoastoperformastep-wisetemperaturesweep.Thephasetransitionofsilveriodideisaccompaniedbyapeakinheatcapacityatthenormaltran-sition temperature and there is a reduction inheat capacity from the β-phase to the α-phase. Figure4(b)showssimilardataforsilveriodidemeasuredwithabackgroundof50Wmicrowavepower.Thepeakinheatcapacityatthephasetransition was not detected due to difficulties intemperaturecontrolduringtheactualphasetransitionitselfbuttheoccurrenceofthephasechangecanbedetectedbythecharacteristicdrop

Figure 2. Schematic of temperature monitoring system for sintering and annealing experiments; the sample temperature is controlled by the optical fibre thermometer at the centre of the sample body.


inbaselineheatcapacitybetween128and130°Cwhich agrees with the transition temperatureobtainedunderthesamemicrowavepowerusingthe‘DTA’approach,seeFigure3. Ratherthanchangetheaveragetemperatureof the sample, anexperimentwascarriedoutwhereby thespecimen temperaturewasoscil-lated between 132 and 128°C over a periodof two minutes as the microwave power wascycledinastepwisefashionbetween0Wand70W.Figure5showsaplotofthespecimen’sapparentheatcapacityasafunctionofmicro-wave power and it appears that silver iodidecan be transformed reversibly between β-AgI and α-AgI under quasi-isothermal conditions by the influence of microwave radiation. Again, these dataareconsistentwiththoseshowninFigures3and4(b). Whilst theDTA-andACcalorimetry-likedata provide evidence for a non-thermal mi-crowave effect in silver iodide, the presenceof temperature gradients within the specimencannotbediscounted.Athermalimagingcamerawas therefore used in place of the fluoroptic ther-mometertomonitorthespecimen’stemperatureduringpuremicrowaveheating.ExampledataareshowninFigure6for150and75Wmicro-

wavepower.Changesinheatingrateoccurredaround90°Cduetoincreasedcouplingofthespecimen with the microwave field suggesting the formation of some of the α-phase. Accompa-nyingthisphenomenonwasadramaticincreaseinthermalgradient(determinedbythedifferencebetweentheminimumandmaximumtempera-turesofthespecimen)acrossthespecimen.Thushotspotsappearedinthesampleconcurrentwiththe formation of α-AgI. Below 90°C, the sample wasfairlyuniformintemperature.Abovethistemperature,inhomogeneitiesinthesampleand/or microwave field could have caused hot spots to appear and thus form α-AgI locally. Alterna-tively,agenuinenon-thermalmicrowaveeffectwould cause nuclei of α-AgI to be generated ata lower temperature thannormaland thesethenleadtolocalisedsuperheating.Inordertodecidebetweenthesepossibilities,itwouldbenecessarytoimprovethespatialresolutionofthethermalimagingcameraand/ordeviseameansof 3-dimensionalmappingof the temperaturedistributionwithinthesample,ordesignanewcavitythatcouldhandlelargersamples.

Figure 3. DTA data for silver iodide heated at 1°C min-1 under increasing levels of microwave power (inset shows expansion of 0 W curve in transition region).


(a)

(b)

Figure 4. (a) Heat capacity of silver iodide measured by AC calorimetry. (b) AC calorimetry data for silver iodide with a 50 W microwave power background.

Figure 5. AC calorimetry data for AgI at 130°C under different levels of microwave power.


Multimode cavity

Sintering ExperimentsThe final density of the ZnO ceramics after sintering at 680, 780 and 900oC are shownin Figure 7. The relatively low final densities achievedweredeliberatesince itensured thatany differences in densification were visible. It can be clearly seen that densification was enhanced by the presence of a microwave field, with the enhancement becoming more signifi-cantthegreaterthelevelofmicrowavepoweruseddespitethefactthatthethermalhistoriesofeachpairofsampleswereidentical.Figure8 shows FEG-SEM micrographs of fracturesurfaces of the submicron ZnO ceramics sin-teredat680oCusingconventional andhybridheating, itcanbeseenthat thesampleheatedusing1kWofmicrowaveenergyhassinteredto a significantly greater extent confirming the densityresultsinFigure7.Similarresultswereobtained for the 3-YSZ powders, Figure 9, and Al2O3powders,Table2,althoughthemagnitudeoftheeffectwassmallerintheselessmicrowaveabsorbingmaterials.Whilsttheeffectintheverylowdielectricloss(at2.45GHz)aluminawas

particularlysmall,itwasrepeatable. These results are in agreement with thework reported by Xieet al. [1999], in whichthreekindsofceramicswithdifferentdielectriclosses,Al2O3, Ce-Y-ZrO2andthelead-basedre-laxor ferroelectric PMZNT, were sintered using 2.45GHzmicrowaveandconventionalheating.Larger increases in densification were observed duringmicrowavesintering in thehigher lossCe-Y-ZrO2 and PMZNT compared to the lower lossAl2O3.Nevertheless,therehavebeenreportsin the literature in which a significant apparent enhancementwasobservedwhensinteringalu-mina[Janney&Kimrey,1988;Brosnanet al.2003].Janney&Kimrey[1988]used28GHzmicrowavestosinterdopedaluminaat~250oClower than with conventional heating. Sinceceramics suchasaluminaabsorbmicrowavesmorereadilyathighermicrowavefrequenciesthisresultisunderstandable,howevertheresultsobtained by Brosnan et al. [2003] using 2.45GHzmicrowavesmustbeinterpretedwithcaresincetheywereobtainedusingseparatemicro-waveandconventionalsinteringfurnaceswithdifferenttemperaturemeasurementtechniquesandheatingrates.

Figure 6. Temperature vs time profiles of silver iodide heated by 150 and 75 W microwave power.


When 3-YSZ and 10-YSZ submicron pow-der samples were sintered using the hybridfurnace,thelatterdisplayedconsistentlyhigherdensification [Wang et al.,2006],aswouldbeexpectedduetothehigherfractionofvacanciesin the crystal structure, and the densification en-hancementarisingfromtheuseofmicrowaveswasalsomorepronouncedinthishigherionicconductivity and more microwave absorbinggradeofpartiallystabilisedzirconia.Thesere-sultsareingoodagreementwiththosereportedbyJanneyetal(1993)whofoundthat12mol%CeO2doped zirconia displayed a much lowerdensification enhancement during microwave sinteringthan8mol%Y2O3 doped ZrO2,thelatterhavingahigherionicconductivitybyafactorof~100[Nightingaleet al.,1997],thoughitshouldbenotedthatagainseparatemicrowaveandconventionalfurnaceswereused.Theseresultsareperhapsnotsurprisingsinceionicconductivityinzirconiasisrelatedtothepresenceofoxygenvacanciesinthelattice,afactorthatalsopositivelyaffectssintering. Thegreatesteffectassociatedwiththeuseofhybridheatingoccurredintheintermediatestagesofsintering(i.e.theregionbetweenabout

65–90%oftheoreticaldensity),withtheeffectdecreasing with increasing densification, Figure 10(a). The basic shape of the densification curve was the same forbothpure conventional andhybrid heating,suggesting that thefundamentalmechanismcouldbesimilarwhenmicrowavesareused. The final average grain size of the samples followedasimilar trend; Figure 10(b) showsthe results for the submicron ZnO powder. The largest enhancement in densification between that obtained with pure conventional heatingandhybridheatingusing1kWofmicrowavepoweroccurredat680oC for the submicron ZnO powder,witha23%densityenhancement,Table3.The equivalent figure for the submicron M 10-YSZ at1220oCwas17.5%whilstthatforthesubmicronM 3-YSZ at the same temperature was 15%. The effectsofparticlesizecanalsobeobservedintheseresults, with finer particle sizes leading to greater enhancement at an equivalent temperature asmight be expected. For example, the submi-cron ZnO powder sintered to roughly the same degree of densification at 680oCasthemicronZnO powder did at 780oC.The comparativeenhancements in densification were 23% and 18%respectively.

Figure 7. Final density curves of a) ‘submicron’ ZnO hybrid sintered at 680oC and 780oC and b) ‘micron’ ZnO hybrid sintered at 780oC and 900oC.

(a) (b)


Figure 9. Final density curves vs MW power for 3-YSZ for a) ‘nano’ powder sintered at 900 and 1000oC and b) ‘submicron D’ powders sintered at 1100 and 1200oC.

Table 2. Final densities of Al2O3 pellets sintered at different temperatures using con-ventional and hybrid heating, the latter with 1000 W of microwave power, but the same

temperature/time profiles for each series.

Final density‘Nano IH’ ‘Nano S’

Sintering temp. 1180oC 1260oC 1400oCPure conv. 57±0.5% 61±0.5% 84±0.5%Hybrid (1000 W mw) 59±0.5% 62±0.5% 85±0.5%

Figure 8. FEG-SEM micrographs of fracture surfaces of submicron ZnO sintered at 680oC using a) conventional heating and b) hybrid heating using 800 W of microwave power.

(a) (b)

(a) (b)


Figure 10. Conventional and hybrid (using 1 kW of microwave power) sintering of the submi-cron ZnO powder. a) The greatest effect occurred at relatively low sintering temperatures, i.e.

in the intermediate stages of sintering. b) The final average grain size of the samples followed a similar trend.

Table 3. Enhancement in densification observed between that obtained using conventional and hybrid heating, the latter with 1000 W of microwave power, at the sintering

temperatures indicated.

Powder Particle size Sintering temperature / oC

Δ% Theor density1

ZnO‘Submicron’ 680 23.3

780 7.2

‘Micron’ 780 17.7900 8.6

3-YSZ

‘Nano’ 900 12.81000 6.7

‘SubmicronD’ 1100 12.41200 3.1

‘SubmicronM’ 1220 15.11260 10.5

10-YSZ ‘SubmicronM’ 1220 17.51260 12.5

Al2O3

‘NanoS’ 1400 1.2

‘NanoIH’ 1180 2.41260 1.8

1. Difference between the final density obtained after sintering at the designated temperature for one hour usingconventionalandhybridheating,thelatterwith1000Wofmicrowavepower.

(a) (b)


Asfarastheauthorsareaware,theseresultscanonlybeexplainedbytwopossibilities.Thefirst is that the so-called ‘microwave effect’ is genuine.However, the second is that thedif-ferent levels of microwave and conventionalpower used have resulted in significantly differ-entthermalgradientswithinthesamplessinceitisknownthatmicrowaveheatingleadstoan‘inverse temperature gradient’, i.e. the centreof thebody ishotter than thesurface[Binner&Cross,1992].BasedontheresultsinFigures7(a)and(b),thesegradientswouldhavetobeatleast100oC in the ZnO pellets during sintering since inboth cases thedensity achievedwithconventionalplus1kWofmicrowavepowerisactuallyhigherthanthatachievedwithpurecon-ventionalheatingusinga100oChighersinteringtemperature.Asindicatedearlier,themagnitudeof the effect in the 3-YSZ is smaller, however thetemperaturegradientswouldstillhavehadtobesubstantial.Table4showsthedifferencesintemperaturebetweentheedgeandcentreofthepelletsasmeasuredbythetwoopticalthermom-etersatthestartandendofthe1hourisothermalholdperiodforthesubmicronzincoxidepelletssinteredat680oC.Itcanbeseenthatatthestart,immediately after the heating period finishes, there is a small but finite temperature difference

withthesurfaceofthesamplebeinghotter.Asexpected, over the1hr hold this temperaturegradientreducesinallcasesasthetemperatureequilibrates;withthehighermicrowavepowerlevels a very slight inverse temperature profile actually forms with the temperature gradientbeing negative indicating that the sample ishotterinthecentre.However,ineverycasethemeasuredtemperaturegradientsare<10oC.Thisis significantly smaller than would be required to explain the densification effects observed above. Thisresultissupportedbythelackofanyvisiblegradientinthegrainsizeorothermicrostructuralfeaturesacrossthediameterofthesamples. In a further attempt to eliminate thetemperaturegradientpossibility,ahighresolutionthermalimagingcamerawasusedtomonitorthetemperature profile across the face of the samples duringthesinteringschedule.Itwasfoundthatthesamplessinteredbyhybridheatingalwaysdisplayed a slightly lower average surfacetemperatureandasmallergradient thanthosesinteredbypureconventionalheating;eventhelargest temperature gradients were ≤10oC. EvenEventhough these gradients were measured on thesurfaceofthesamplesandtheywillbelargeratthecentreofthecompacts,theycannotexplainthevariation insintereddensitiesobservedas

Table 4. Temperature gradients across the sample radius, as measured by the optical thermometers, for the submicron zinc oxide pellets sintered at 680oC as a function of the

amount of microwave (mw) power used.

Power Temperature gradient at sintering temperatureΔT = Tedge – Tcentre / oC

Startofisothermalhold EndofisothermalholdConventional 10 5

Hybrid(200Wmw) 8 5Hybrid(400Wmw) 7 5Hybrid(600Wmw) 8 4Hybrid(800Wmw) 6 -1Hybrid(1000Wmw) 8 -3


a function of microwave power.Very similarresults were obtained for the other materials[Wanget al.,2006]; in every case the maximum;ineverycasethemaximumtemperature difference was far too small toaccount for the greater densification observed whenhybridheatingwasused.Iftemperaturegradientscannotbeusedtoexplaintheresultsthen this therefore appears to be firm proof of theexistenceoftheso-calledmicrowaveeffectduringceramicssintering.

Annealing ExperimentsFigure 11 shows the polished and etchedmicrostructures at the centre of the ZnO disks beforeannealingandafterannealingat1000oCand1100oCfor5hoursusingpureconventionaland hybrid heating. It can be seen that graingrowth was enhanced by the use of hybridheating, a phenomenon confirmed by the grain sizedata shownafter5hoursofannealing inFigure 12. The figure reveals that the onset of

significant grain growth during conventional heating was at around 1150oC, whilst for thehybrid heated samples it occurred at about1050oC.By1200oC,thedifferenceingrainsizebetweentheconventionallyandhybridannealedsamples was significantly reduced, supporting the idea that microwaves have their greatesteffectduringtheearlytointermediatestagesoftheprocess [Wanget al.,2006].Thisappearsto be confirmed by Figure 13, which shows thattheslopesofthelinesat1100and1150oCweredifferentfortheconventionalandhybridcasesalthoughtheywereessentiallythesameat1200oC. The difference in temperature, ΔT, thatexistedacrosstheradiusofthedisksthatwereannealedfor5hoursasmeasuredbythetwoopticalthermometerswasalwaysfoundtobelessthan20oC,fartoolittletoaccountforthesignificantly larger average grain sizes at 1100 and1150oC[Binneret al.,2007]. The kinetic grain growth equation can be

Figure 11. FEG-SEM micrographs of the centre of the submicron ZnO pellets: a) before anneal-ing; and after annealing for 5 hours, b) at 1000oC using conventional heating, c) at 1000oC using hybrid heating, d) at 1100oC using conventional heating and e) at 1100oC using hybrid heating.

The hybrid heating anneals involved 1000 W of microwave power.


Figure 12. Average grain sizes after 5 hours for annealed ZnO samples as a function of anneal-ing temperature and location on the cross section of the sample (see inset) with conventional and

hybrid heating.

Figure 13. Average grain sizes as a function of time for annealed ZnO samples after convention-al (C) and hybrid (H) heating at 1100oC, 1150oC and 1200oC.


expressedbyequation1[Coble,1961]:

G G Ktn n− =0 (1)

WhereGistheaveragegrainsize,nisanintegerthatdependsonthediffusionmechanism, t istimeandKcanbeexpressedbytheArrheniusequation:

K K QRT

= −0 exp( ) (2)

whereK0isthepre-exponentialconstantofthematerial,Qtheactivationenergy,Rthegascon-stantandTtheabsolutetemperature.Hence,thekineticgraingrowthequationcanbewrittenas:

G G K QRT

tn n− = −0 0 exp( ) (3)

InthisexpressiontheinitialgrainsizeG0canbeneglected because it is significantly smaller than thegrainsizeGattimet.Thus,equation3canbe simplified to:

G K QRT

tn= −0 exp( ) (4)

Thishasbeenwidelyappliedbymanyresearch-erstocalculatethegraingrowthexponentvaluen,e.g.[Bennisonet al.,1983],byplottingintheform:

log log logGn

Kn

t= +1 1

(5)

Where theslopeof the logGversus log t line

is 1/n, the grain growth exponent. Hence thesmaller thevalueofn, thegreater the rate ofgraingrowth. Figure 14 illustrates the results for ZnO an-nealedat1100oCwhilstTable5showsthevaluesfornasafunctionoftemperature.Itcanbeseenthatat1100and1150oCtheywere3.3duringpureconventionalheatingand1.4duringhybridheatingrespectively,whilstat1200oCtheybothdecreasedtoaround1.5.Manyresearchershavereportedgraingrowthresultsduringsinteringofundoped ZnO using pure conventional heating andthevalueofnobtainedis~3inthe1100-

Figure 14. Log grain size versus log time for ZnO pellets annealed at 1100oC using conven-tional and hybrid heating. The grain growth

exponents, n, are shown on the plot.

Table 5. Grain growth exponent values for un-doped, submicron ZnO as a function of tem-perature for conventional heating and hybrid heating. Errors are those arising from the fit

to the data.

Annealing temperature/ oC

Grain growth exponent, n

Conventional heating Hybrid heating

1100 3.3±0.1 1.4±0.11150 3.3±0.1 1.4±0.11200 1.6±0.2 1.5±0.2


1150oCtemperaturerange[Duttaet al.,1970;Sendaet al.,1990], showinggoodagreementwiththepresentresults.Thefactthatthevalueofnat1200oCduringconventionalheatingandat1100–1150oCduringmicrowave-basedheatingisbothlowerandsimilarsuggeststhattheeffectofusingmicrowavesissimplytoacceleratetheconventionalprocesssothatithappensroughly100oCsooner,inlinewiththeresultsfromthesinteringstudy[Wanget al.,2006].

CONCLUSIONS

Two completely different investigations havebeenpursuedintotheso-called‘microwaveef-fect’,onebasedonthemeasurementofthermalpropertiesassociatedwithphasechangesusingasinglemodecavityandtheotherthedetermi-nationofsinteringandannealingcharacteristicsofceramicmaterialsusingamultimodecavity.However,inbothcasestheequipmentallowedheatingtobeaccomplished,inthesamecavityandusingexactlythesameexperimentalcon-figuration, with either pure microwave, pure conventional or any combination of hybridheatingandtheresultsshowclearlythatami-crowaveeffectcanbeobserved.Thesinteringinvestigationalsoshowedarelationshipbetweenthemagnitudeof the effect and thedielectriccharacteristics of the material; the higher thedielectriclossthegreatertheeffect.Apossibleexplanationof theeffect in termsof tempera-ture gradients was discounted in both cases.Although temperaturegradientswerepresent,asexpected,theyweresubstantiallytoosmalltoaccountfortheresultsobserved.

ACKNOWLEDGMENTS

The authors gratefully acknowledge fundingfromEPSRCintheUK,researchgrantnumbersGR/R94220andGR/R52435.

REFERENCES

BennisonS.J.andHarmerM.P.(1983).“EffectofMgOsoluteonthekineticsofgraingrowthinAl2O3.”J. Am. Ceram. Soc.,66(5)C90-C92.

BinnerJ.G.P.andCrossT.E.(1992).“Useoftheinversetemperature profile in microwave processing ofadvancedceramics.”Mat. Res. Soc. Proc.,269,pp.357-362.

BinnerJ.G.P.,HassineN.A.andCrossT.E.(1995).“Thepossibleroleofthepre-exponentialfactorinexplain-ingtheincreasedreactionratesobservedduringthemicrowavesynthesesof titaniumcarbide.”J. Mat. Sci,30,pp.5389-5393.

BinnerJ.G.P,PriceD.M.,ReadingM.andVaidhyanathanB. (2005). “Hybrid microwave / conventionallyheatedcalorimeter.”Review of Scientific Instruments,76064902,pp.1-4.

BinnerJ.G.P.,VaidhyanathanB.andWangJ.(2003).“Acomparative study of temperature measurementsduringmicrowaveprocessing.”Proc. 9th Int. Conf. on Microwave and High Frequency Heating,pp.477-480,Loughborough,UK.

BinnerJ.,WangJ.,VaidhyanathanB.,JoomunN.,KilnerJ.,DimitrakisG.andCrossT.E.(2007).“Evidenceforthemicrowaveeffectduringtheannealingofzincoxide.”J. Amer. Ceram. Soc.,90(9),pp.2693-2697.

BochP.,LequeuxN.andPilusoP.(1992).“Reactionsinter-ingofceramicmaterialsbymicrowaveheating.”Mat. Res. Soc. Proc.,269,pp.211-216.

BooskeJ.H.,CooperR.F.,DobsonI.andMcCaughanL.(1991),“Modelofnonthermaleffectsonionicmobil-ityduringmicrowaveprocessingofcrystallinesolids.”Ceram. Trans.,21,pp.185-192.

BrosnanK.H.,MessingG.L.andAgrawalD.K.(2003).“Microwavesinteringofaluminaat2.45GHz.”J. Am. Ceram. Soc.,86(8),pp.1307-1312.

ChenD. andDollimoreD. (1995). “Thepossibilityofmanufacturing a single-pan differential thermalanalyzer unit.” Thermochim. Acta, 249, pp.259-267.

Coble R.L. (1961). “Sintering crystalline solids: IIExperimental test of diffusion models in powdercompacts.”J. Appl. Phys.,32(5),pp.793-799.

DuttaS.K.andSprigsR.M.,(1970).“Graingrowthinfullydense ZnO.” J. Am. Ceram. Soc.,53(1),pp.61-2.

Fathi Z., Ahmed I., Simmons J.H., Clark D.E. and Lodding A.R. (1991). “Surface modification of sodiumaluminosilicate glasses using microwave energy.”Ceram. Trans,21,pp.623-629.

Giguere R.J. (1992). “Microwave assisted organic-synthesis.”Mat. Res. Soc. Proc.,269,pp.387-396.

Janney M.A., Calhoun C.L. and Kimrey H.D. (1991).“Microwave sintering of zirconia-8 mol% yttria.”


Ceram. Trans.,21,pp.311-318.Janney M.A., Jackson M.L. and Kimrey H.D. (1993).

“Microwave sintering of ceria-doped zirconia.”Microwave sintering of ceria-doped zirconia.””Ceram. Trans.,36,pp.101-108.

JanneyM.A. andKimreyH.D. (1988). “Microstructureevolution inmicrowavesinteredalumina.”Ceramic Powder Science II, Am. Ceram. Soc.,pp.919-924.

JanneyM.A.andKimreyH.D.(1991).“Diffusion-controlledprocesses in microwave-fired oxide ceramics.” Mat. Res. Soc. Proc.,189,pp.215-227.

JanneyM.A.,KimreyH.D.,SchmidtandM.A.andKiggansJ.O.(1991).“Graingrowthinmicrowave-annealedalumina.” J. Am. Ceram. Soc.,74(7),pp.1675-1681.

KatzJ.D.,BlakeR.D.andKenkreV.M.(1991).“Microwaveenhanceddiffusion?”Ceram. Trans.,21,pp.95-105.

KraftmakerY. (2002). “Modulation calorimetry andrelatedtechniques.”Physics Reports,pp.3561-117.

LewisD.A.(1992).“Microwaveprocessingofpolymers-anoverview.”Mat. Res. Soc. Proc.,269,pp.21-31.

NightingaleS.A.,WornerH.K.andDunneD.P.(1997).“Microstructuraldevelopmentduringthemicrowavesinteringofyttria—zirconiaceramics.”J. Am. Ceram.Soc.,80(2),pp.394-400.

Robb G.R., HarrisonA. and WhittakerA.G. (2002).“Temperature-resolved, in-situ powder X-raydiffraction of silver iodide under microwaveirradiation.”Phys. Chem. Comm.,5,pp.135-137.

SendaT.andBradtR.(1990).“GraingrowthinsinteredZnO and ZnO-Bi2O3ceramics.”J. Am. Ceram. Soc.,73(1),pp.106-14.

WangJ.,BinnerJ.,VaidhyanathanB.,JoomunN.,KilnerJ.,DimitrakisG.andCrossT.E.(2006).“Evidenceforthemicrowaveeffectduringhybridsintering.” J. Amer. Ceram. Soc.,89(6),pp.1977-1984.

Willert-PoradaM.,KrummelT.,RohdeB.andMoormannD. (1992). “Application of microwave heating tocombustion synthesis and sintering ofAl2O3-TiCceramics.”Mat. Res. Soc. Proc.,269,pp.199-204.

Xie Z., Yang J., Huang X. and Huang Y. (1999). “Microwaveprocessingandpropertiesofceramicswithdifferentdielectricloss.”J. Eur. Ceram. Soc., 19,pp.381-387.

e for n -thermal microwave e ingle multimode h c /m spower used in a molybdenum disilicide...

Documents