Materials Science and Engineering A266 (1999) 211–220

Use of partially oxidized SiC particle bed for microwave sinteringof low loss ceramics

Peelamedu D. Ramesh a,*,1, David Brandon a, Levi Schachter b

a Department of Materials Engineering, Technion-Israel Institute of Technology, Haifa 32000, Israelb Department of Electrical Engineering, Technion-Israel Institute of Technology, Haifa 32000, Israel

Received 17 March 1998; received in revised form 17 November 1998

Abstract

A 1 kW hybrid microwave system was fabricated using a partially oxidized SiC powder bed and used to sinter smallcomponents of low-loss insulating ceramics. Samples of 3% yttria stabilized zirconia/20% alumina (3YZA) and 99% alumina weresintered to final densities of 99%(3YZA) and 95%(99% alumina). Partially oxidized silicon carbide (b-SiC) powder was used assusceptor (preheater). By comparison, unoxidised b-SiC powder, which couples well with microwaves at room temperature,exhibited thermal runaway above 400°C. It could be possible that at high oxidation levels the connectivity between SiC particlesin b-SiC powder bed might become depercolated, and a model similar to one proposed for silicon nitridation under microwavescould be appropriate in understanding the microwave absorption phenomenon. Stable temperature measurements at variouspositions in the system confirmed that the maximum temperature was at the center of the 3YZA sample. The plot of the relativetemperature difference between the sample and its surroundings as a function of sample temperature resulted in a bell-shapedcurve with a clear maximum at around 800°C, associated with the rapid increase in radiation heat transfer above this temperature.Experiments also confirmed that sintering occurred at lower temperatures in a microwave field when compared to conventionalsintering. © 1999 Elsevier Science S.A. All rights reserved.

Keywords: Hybrid microwave heating; Oxidized silicon carbide; Depercolation; Sintering

1. Introduction

Sintering of ceramics is an attractive area of mi-crowave processing [1]. The advantages include rapidheating rates, short processing times, low power re-quirements and reduced reaction with the atmosphere.Materials with a high loss tangent can couple withmicrowaves at room temperature while less ‘lossy’ ma-terials need higher initial temperatures in order tocouple. Some fine powders, such as carbon [2], siliconcarbide [3] and vanadium oxide [4] couple efficientlywith microwaves at room temperature, whereas otherssuch as zirconia [5] and alumina [6], only couple at hightemperatures. The dependence of dielectric constant (o %r)and loss tangent (tand) on temperature are needed to

predict the power absorption characteristics of materi-als in a microwave field.

Kitchen microwave ovens provide inexpensive multi-mode cavities for experimental materials processing.The modes generated within such a cavity result in acomplex electric field distribution, so that the specimenposition within the cavity is an important variable inthe heating process. More uniform electric fields may beachieved either by operating at higher frequencies or byusing mode ‘stirrers’. Larger specimens (a high fillingfactor) absorb microwaves more efficiently and result inhigher heating rates, but have the disadvantage of fieldnon-uniformity and resonances within the specimen(which may result in cracking of the sample). Heatgeneration is also affected by the depth of penetration(Dp), in the sample. As the temperature of the materialrises, microwave penetration decreases, leading to local-ized surface heating due to non-uniform power absorp-tion [7,8]. Smaller workpieces experience a uniformelectric field, but the heating efficiency is then very low

* Corresponding author. Tel.: +49-721-6087922; fax: +49-721-174263.

E-mail address: hr04@rz.uni-karlsruhe.de (P.D. Ramesh)1 Now at Institut fur Keramik im Maschinenbau, Universitat Karl-

sruhe, Haid-und-neu-Str. 7, 76131 Karlsruhe, Germany.

0921-5093/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved.

PII: S0921 -5093 (99 )00017 -9

P.D. Ramesh et al. / Materials Science and Engineering A266 (1999) 211–220212

Fig. 1. The microwave hybrid heating configuration. All dimensions are in mm. The temperature was measured at positions a, b, c, and d,corresponding to the outer insulation, the inner insulation, the alumina fiber board and the sample, respectively. Measurement of temperature onthe alumina fiberboard surrounding the sample ‘c’ is a measure of susceptor temperature, since the sample is well-separated from the fiberboard.

due to the low mass [9] and can only be improved byusing a single mode cavity [10].

Since the dielectric loss properties increase with tem-perature, low loss materials can be made absorptive byraising their temperature. Most commonly, a roomtemperature susceptor surrounds the specimen and sup-plies thermal energy by heat transfer, simultaneouslyimmersing the specimen in both thermal and electricfields2. Computer modeling of microwave heating indi-cates that, despite a decrease in total electric fieldintensity, the field at the specimen is not significantlyscreened by using a preheat susceptor [7].

Several preheating configurations have been sug-gested. De’et al. [11] used a SiC-lined susceptor aroundthe specimen to sinter Al2O3 compacts. An array of SiCrods [12] has been used to sinter 8Y–ZrO2 (a ‘picketfence’ arrangement). Using this configuration, densitiesas high as 99% have been reported for specimen sinter-ing temperatures of 1100°C, ca. 200°C lower than the

conventional sintering temperature of the same powder.Carbon [13] has also been tried as a preheating source,while LaCrO3 and mixtures of Al2O3+LaCrO3 havealso proved to be suitable susceptor materials (Spotz etal. [14]).

In this study, we report the use of a partially oxidizedb-SiC powder compact as a susceptor (pre-heater) tosinter small components of low-loss ceramics (Al2O3

and ZrO2). The susceptor powder surrounds the speci-men cavity and heats the low-loss specimen to tempera-tures of the order of 1250°C.

2. Experimental

2.1. The hybrid microwa6e heating system

Various configurations [11,12] of hybrid heating haveproved successful for ZrO2 and Al2O3 compacts. Poormicrowave absorbers such as porous magnesia (MgO)and quartz (SiO2), form the container walls of therefractory cavity, and alumina fiber board and fibrousalumina are used as thermal barriers. A schematic of

2 In this study radiation heat transfer within the microwave ovenassembly is taken to refer to thermal radiation and not to theexternally imposed microwave radiation field.

P.D. Ramesh et al. / Materials Science and Engineering A266 (1999) 211–220 213

the present hybrid heating arrangement is shown in Fig.1. The loss tangent values of MgO and SiO2 are verylow and these ceramics remain transparent to mi-crowaves even above 1000°C. b-SiC powder (600 grit)was used as the susceptor material. About 30–40 g. ofthe SiC powder was placed in the gap between thealumina fiberboard inner cavity and the quartz wall.The quantity of susceptor powder determines the elec-tric field at the sample and was selected to ensure thatthe specimen was adequately exposed to the electricfield at high temperatures, even though the microwavepenetration depth into the SiC powder bed decreases asthe temperature rises. Temperature measurements weremade at selected sites, labelled a, b, c and d in Fig. 1,using a type R (Pt–Pt13%Rh) thermocouple within analumina support tube. The specimen temperature Td

was measured by inserting the thermocouple into asmall hole at the center. The temperature at the Al2O3

fiber board Tc was taken as an estimate of the tempera-ture of the silicon carbide susceptor. Temperatures werealso determined across the thickness of the porousMgO crucible Ta and Tb. The present system wasdesigned to generate a moderate thermal gradient in thecentral region: a critical consideration for uniform sin-tering and crack prevention.

2.2. Temperature measurement

Simultaneous temperature measurement in a continu-ous electromagnetic field using a thermocouple is usu-ally impossible as the thermocouple–field interactionleads to gross errors. A very thin metal sheath sur-rounding the thermocouple could be used to overcomethis problem, but contact between the thermocouple tipand the metal sheath must be avoided, while, shieldingreduces the response of the thermocouple. In this work,the microwave energy was pulsed and an alumina cov-ered thermocouple, with the tip uncovered, was usedfor calibrating temperature measurements. Tempera-tures were measured using a Pt–Pt 13% Rh thermocou-ple (type R, 0.5 mm wire). The stable temperaturereadings agreed with readings obtained with a platinumsheathed thermocouple. While the temperature readingsare judged to be reliable, they are still average readingswhich do not necessarily reflect the energy conversionprocesses associated with the microwave power absorp-tion. In particular, the relation between the measuredtemperature and the effective temperature in the sinter-ing process may not be linear.

2.3. Absorption characteristics of SiC, ZrO2 and Al2O3

Polar substances absorb microwaves to convert mi-crowave energy to thermal energy. The generalizedequation for the absorbed power is written as

Pmw=Pinc(1−p2)(1−e−2ad) W m−3 (1)

where Pmw is the total power absorbed by the material.The reflection coefficient, p, and attenuation constant,a, mainly determine the total power absorbed in thematerial. Low values of p and a correspond to insulat-ing materials, in which a limited microwave absorptionresults. Metals reflect microwaves and have high valuesof p and a.

For a homogeneous material, reflection constant [15]can be written as

p=1−o r*

1+o r*(2)

and attenuation constant a,

a=2pfc�o0o r

’

2�1/2!�

1+� s

2pfo0o r%

�2n1/2

−1"1/2

(3)

where o r* is the RMS value of the complex dielectricconstant, s is the electrical conductivity of the material,f=2.45 GHz, c=3×108 m s−1, and o0 and o r% are thefree space permittivity and the relative dielectric con-stant of the material, respectively.

At a fixed microwave frequency, only the materialproperties or=o0o r% and tand determine the total powerdissipated within the sample. Both these parametersdepend strongly on temperature [16].

Experimental time versus temperature curves forfresh powders of SiC, Al2O3 and ZrO2 are given in Fig.2 for increasing power loadings. Experiments were con-ducted with 50 g. of each powder sample in an aluminacrucible without any preheating powders. The figureclearly indicates the strong absorption of the SiC pow-der in comparison with the other powders. A fewminutes exposure, even at low power levels (450 W),raises the temperature of fresh SiC powder above1600°C, with thermal runaway occurring after 90 minof irradiation. More unexpectedly, after three successiveruns of the now partially oxidized SiC powder themicrowave heating curve exhibits much better couplingat the lower temperatures with no runaway at hightemperatures.

All subsequent hybrid sintering experiments wereperformed with a partially oxidized SiC powder suscep-tor for two reasons: (1). Repeated runs generated nearlyidentical power/temperature/time conditions. (2) Thecontrolled coupling of the electric field together withthe moderate temperature maximum ($1250°C at 1kW) fulfilled the basic prerequisites for a hybrid pre-heating system.

2.4. Penetration depth

Microwave absorption for planar thick samples canbe characterized by a parameter termed the penetrationdepth, DP. The primitive equation to calculate the

P.D. Ramesh et al. / Materials Science and Engineering A266 (1999) 211–220214

depth of penetration for microwaves in a homogeneousmaterial is given by

DP=lo

2p tand o %r(4)

At a frequency of 2.45 GHz, the free space wavelengthlo=122.4 mm. The temperature dependence of Dp wascalculated from the data of o r% and o r¦ for Al2O3 and SiCavailable in the literature [16]. At 1350°C, 2.45 GHzmicrowaves penetrate 1 mm in SiC and 93 mm in solidalumina. Alumina only absorbs microwaves signifi-cantly above 800°C, SiC absorbs at room temperaturebut at high temperatures heat generation is limited tothe surface layer and the susceptor will screen thesample. It follows that both the thickness and thetemperature of the susceptor layer must be limited.

In general, if the penetration depth exceeds samplethickness, the material behaves transparent. Volumetricheating results, if the penetration depth is comparablewith the sample thickness. Otherwise, if the sample istoo thick, absorption limits to the surface.

In the case of an oxide layer surrounding the suscep-tor powder particle, such as SiC powder particle sur-rounded by silica, the above Eq. (4) can not beapplicable. For a compositionally changing ceramicsystem, a more complex equation involving terms forconnectivity and percolation, is needed to describe themicrowave penetration depth.

2.5. Materials and methods

Specimens with two different compositions and ofvarious shapes (see Fig. 8) were used for the sinteringstudies. The compositions used were 3Y2O3–ZrO2/20Al2O3 (YZA) and 99% Al2O3. The nominal particle sizeof the YZA powder was 20 nm. All YZA samples were

prepared by gelled tape-casting [17] as shown in theflowchart (Fig. 3). Cylindrical pellets of YZA 15 mm indiameter and 4–5 mm in thickness were obtained fromthe gelled tape by uniaxial pressing. The measuredgreen density of these samples was 50–55% of theoreti-cal. Al2O3 green samples were supplied in the form ofsmall cylinders (16 mm height and 3 mm diameter). Amaximum of three YZA pellets (approx. 15 gm.) and 15Al2O3 specimens (ca. 5 g) were placed separately in thefiber-board specimen cavity (Fig. 1).

A kitchen microwave oven was used (multimodecavity, Sanyo Super Showerwave, 1000 W maximumpower). The microwave power was pulsed at a constantrate, but with the pulse duration adjusted for differentpower levels (at maximum power the power supply wascontinuous). Stable temperatures were recorded in-situbetween pulses.

No ‘sparking’ associated with microwave–thermo-couple interaction was observed. The hybrid pre-heat-ing system (Fig. 1) was placed at the center of themicrowave oven cavity and insulated by alumina fiber-board. Even at high specimen temperatures the heat-losses due to convection from the insulated system werelow and no overheating of the oven was observed.

3. Results and discussion

3.1. Oxidation of SiC powder

When the SiC powder bed is microwave heated totemperatures above 1000°C in ambient atmosphere,oxidation phenomena takes place. The oxidation phe-nomena [18] that takes place in SiC is briefly discussedbelow. The possible oxidation reactions for SiC couldbe

Fig. 2. Effect of power-dependent microwave absorption on the heating of SiC, Al2O3 and ZrO2 powders without any preheater (susceptor).

P.D. Ramesh et al. / Materials Science and Engineering A266 (1999) 211–220 215

Fig. 3. Flowchart for YZA pellet formation by tape-casting.

where b is the mass gain per unit volume of SiO2.Values of Kpa were calculated for oxidation of silicon,based on oxygen diffusion through a silica layer, andcompared with the experimental SiC oxidation data.

From a thermodynamical point of view, various pos-sible reaction rate limiting steps in the SiC oxidationreaction are discussed in detail elsewhere [18]. Thesesteps include (i) inward O2 diffusion through SiO2 layer(ii) reaction at the SiC � SiO2 interface and (iii) outwarddiffusion of CO molecules. On observation of a higheroxidation rate of silicon compared to SiC, it was sug-gested that more than one rate-limiting step is involvedin SiC oxidation, i.e. a mixed interface-reaction/diffu-sion controlled oxidation takes place in the case of SiC.

3.2. Absorption by different materials

Some attempt has been made to understand theabsorption characteristics of powders as opposed tosolid samples. An amount of 15 g of sample powderwas placed in an alumina crucible which was sur-rounded by 50 g of fresh SiC powder (as susceptor).The heating curves for different sample powders at aconstant power level of 650 W are shown in Fig. 4. Themeasured temperatures reflect both the microwave ab-sorption by the powder sample (the sample temperaturewas measured using a Pt–Pt13%Rh thermocouplewhich was covered by a thin alumina tube in order toavoid the thermocouple tip touching the sample pow-der) and mixed mode heat transfer from the SiC suscep-tor powder: firstly, if microwave absorption by thepowder sample were negligible, at all temperatures,then all the heating curves should be similar, with thesample temperature only dependent on heat transferthrough the sample powder. Secondly, the lower ther-mal conductivity of ZrO2 as compared to Al2O3 shouldresult in a higher Al2O3 temperature at a given time.However, in Fig. 4, the ZrO2 samples reach highertemperatures confirming that ZrO2 powder absorbs mi-crowaves much more efficiently than Al2O3.

The vast difference in the heating curves betweenunoxidized SiC powder bed and oxidized SiC powderbed clearly suggests that in these two materials, themicrowave absorption nature is not identical. When aSiC powder compact is microwave heated to tempera-tures above 1000°C in an ambient atmosphere, thesurface of SiC particle starts oxidizing leading to for-mation of a SiO2 layer. This system can now be viewedas a compositionally changing ceramic system, which issimilar to nitridation of silicon under microwaves,where a conducting silicon phase continuously getsconverted to a non-conducting silicon nitride phase. Afinite difference model developed by Skamser et al. [19]gives a deeper insight into the microwave absorptionnature in a compositionally changing ceramic system.Indeed, the model was successful in predicting the

SiC (s)+O2 (g)USiO2 (s)+C (s) (a)

and

SiC (s)+3/2O2 (g)USiO2 (s)+CO (g) (b)

The reaction rate constant, Kp, can be written as

Kp=(Dx)2

t=

2DO2

0,*(pO2

g −pO2

i )a

(5)

where D0,*O2

is the tracer diffusion coefficient in silica, a

is the number of moles of oxygen required to form onemole of silica and pg

O2and pi

O2are the oxygen partial

pressures at the oxygen � silica and SiC � silica interfaces,respectively. Dx is the SiO2 layer thickness throughwhich oxygen diffusion takes place. Depending on the a

value, one of the above described reactions ((a) or (b))is chosen.

The above Eq. (5) shows a parabolic behavior, whichcan also be expressed in terms of SiO2 mass gain, Km,

Km=Kpb2 (6)

P.D. Ramesh et al. / Materials Science and Engineering A266 (1999) 211–220216

temperature and composition distributions inside a re-action bonded silicon nitride specimen [20]. During thenitridation reaction, silicon was gradually converted tosilicon nitride, and at the intermediate stage, the mate-rial was two phasic nature. Despite an insulating layerformation, their experiments revealed the existence ofsilicon connectivity at low conversion levels. At around25% conversion, the electrical conductivity was maxi-mum, thereafter started decreasing with more and moreconversion, and reached below the detection limit be-yond 70% conversion. At this region, the silicon parti-cles lost their connectivity and became completelydepercolated, clearly indicating that there was a neckformation due to sintering of silicon particles, occurredat the initial stages of reaction. As long as the connec-tivity among silicon particles persisted, silicon behaved

like a conductor, reflecting most of the microwaveenergy. But, when the silicon particles lost their connec-tivity and became depercolated, the system behaved likean insulator, absorbing more microwave energy. Thisled to non-uniform local heatings in the specimen,where the more nitrided regions heated to higher tem-peratures compared to the less nitrided ones. Thomas etal. [20] have obtained a reasonably accurate empiricalmodel of the microwave heating characteristics of theRBSN system by assuming it as a homogenous conduc-tor at low conversion levels and like a conductor loadeddielectric (CLD) at high conversion levels.

Let us consider the microwave heating process offresh SiC powder bed. At the initial stage, the unoxi-dized silicon carbide powder is homogeneous, having acharacteristic dielectric behavior. At the initial stages of

Fig. 4. Heating curves for various ceramic powders surrounded by fresh SiC susceptor powder (constant power loading).

Fig. 5. Experimental time-temperature profiles at various sites within the cavity (see Fig. 1) on a YZA sample.

P.D. Ramesh et al. / Materials Science and Engineering A266 (1999) 211–220 217

Fig. 6. Plot of [1−Tc/Td] versus sample temperature [Td] (YZA specimen).

the heating process (i.e. before oxidising), neverthelessthe Eqs. (1)–(3) can be used to calculate the reflectioncoefficient, the attenuation constant and the powerabsorption. The temperature dependence of conductiv-ity for yellow green coloured SiC shows a semiconduc-tor behavior with an increased conductivity on heatingfrom room to high temperatures. This dielectric lossfactor increases sharply at temperatures beyond 400°C,which makes the material to show a thermal runawaybehavior beyond this temperature. As the SiC powderbed gradually oxidize to SiO2 (see section 3.1), thesystem gradually slips away from homogeneity, consist-ing of two different solid phases coexisting with poros-ity. With the continuous increase in SiO2 volumefraction during microwave heating, in a normal sce-nario, it can be assumed that the semiconducting SiCparticles are gradually isolated by an electrically insu-lating layer of SiO2. This would cause a gradual de-crease in the tangent loss of the SiC, as their volumefraction decreases during the oxidation phenomenon.

Since the above case is somewhat analogous to sili-con nitridation under microwaves, we believe that asimilar modelling procedure could be possible. On thesame basis, we speculate that at the intermediate stage(partially oxided), the conductivity of SiC could be ofthe same order of magnitude as that of unoxidised SiC,which might suggest that the silicon carbide particlessintered partially and still remain well connected. But,the silicon carbide phase would become completelydepercolated at higher oxidation levels resulting in avery low conductivity. This could mean that at lowconversion levels, SiC might behave like a semiconduc-tor and at high conversion levels, it might transform toa semiconductor loaded dielectric (SLD) material.

The microwave absorption behavior of such a mate-rial is more complex and a mathematical treatmentsimilar to nitridation of silicon reaction could be help-ful in understanding the microwave absorption natureof this material. The above given Eqs. (2)–(4) forreflection coefficient, attenuation constant and the pen-etration depth respectively, becomes invalid at highconversion levels and a new set of equations similar tothe one proposed by Neelakanta and Park [21] might bemore appropriate. More importantly, in the energybalance equation, the SiC depercolation, and chemicalphenomena of oxidation should properly be accountedin order to get accurate temperature distributions in thepowder bed. This discussion is purely speculative andno conductivity measurements were made in order toconfirm the percolation phenomenon.

3.3. Temperature profiles

Temperature versus time profiles measured as a func-tion of power input for the different sites in the hybridsystem of Fig. 1 are shown in Fig. 5. The highesttemperatures were recorded at the center of the YZAsample Td, (1340°C), whereas the highest temperaturemeasured in the alumina fiber board Tc, was 1250°C.Curve Td diverges increasingly from curve Tc above400°C, consistent with increasing microwave couplingof the YZA specimen. This divergence reaches a maxi-mum at 800°C and then decreases, consistent withincreasing radiation heat transfer. The relative tempera-ture divergence was plotted as [1−Tc/Td] against thesample temperature [Td], Fig. 6 and can be qualitativelyexplained as follows. At low temperatures (B400°C)there is no YZA specimen-microwave coupling, and themeasured temperature at the center of the sample is

P.D. Ramesh et al. / Materials Science and Engineering A266 (1999) 211–220218

Fig. 7. Radiant heat transfer from a YZA specimen calculated from the experimental temperature values.

predominantly the result of heat transfer from thesusceptor powder. Above 400°C [1−Tc/Td] increases asthe sample starts to couple and absorb microwaveenergy. As heat transfer due to radiation, proportionalto Td

4, increases, above 800°C the temperature differ-ence again decreases, leading to the observed bell-shaped curve in the [1−Tc/Td] versus Td graph.

Heat transfer due to radiation can be calculated fromthe appropriate values for the parameters in the radia-tion equation. The radiant heat loss from the samplecalculated from experiments as a function of the sampletemperature [Td] is shown in Fig. 7. As predicted, theradiation loss increases rapidly above 800°C, but thensaturates as the temperature difference (Td−Tc) de-creases at the highest temperatures.

3.4. Sintering

In most reports on microwave sintering of Al2O3 orZrO2 ceramics, researchers have observed a ‘microwavefield effect’. Janney et al. [22] have sintered Al2O3+0.1% MgO compacts to densities of ca. 98% at 1200°Cusing a 28 GHz microwave furnace, a temperaturenearly 300°C lower than that required for conventionalsintering. Using the ‘picket fence’ SiC susceptor ar-rangement, the same authors have reported [23] sinter-ing of ZrO2–8% Y2O3 compacts to densities above 99%at 1195°C, ca. 150°C lower than the conventional firingtemperature. Fang et al. [24] have reported much re-duced sintering times for hydroxyapatite ceramicsheated in a multimode cavity compared to conventionalsintering [25]. They showed that the same soaking timeand sintering temperature in the microwave field gavehigher densities and finer microstructures, and that this

effect was the same at all temperatures. Enhanceddensification in a microwave field was also observed forZrO2–3% Y2O3 compacts by Nightingale et al.[26].Even though there is still much to explain in mi-crowave-field assisted densification, the effect appearsreal and significant.

In the present work, specimen densities of 99% of thetheoretical value were achieved for YZA and about95% for Al2O3 samples. Example of the various greenand sintered samples are given in Fig. 8. Uniform sizereduction (density) was observed when a batch of 15Al2O3 samples were sintered together. The microstruc-ture of the sintered YZA samples showed a uniformdistribution of the two phases (the Al2O3 grains aredark, Fig. 9a). The grain size of the Al2O3 samples in

Fig. 8. Sintered and unsintered samples of different composition usedin the present work.

P.D. Ramesh et al. / Materials Science and Engineering A266 (1999) 211–220 219

Fig. 9. Scanning electron micrographs of (a) YZA, r=99%, (Al2O3

grains appear dark) and (b) Al2O3, r=95%.

samples with different soaking times are shown in Fig.10. The sintering temperature for YZA was 1340°C(Fig. 5) with soaking times from 5 to 60 min. Thetemperature of the Al2O3 sample was not recorded.The faster sintering of the YZA samples is attributedto greater microwave absorption in ZrO2 as comparedto Al2O3.

4. Conclusions

Room temperature microwave absorption by siliconcarbide powder was utilised to construct a microwavehybrid heating configuration. Partially oxidized SiCpowder proved a very satisfactory room temperaturepreheater, generating reproducible power/time/temper-ature conditions. Since the SiC powder bed was conti-nously oxidizing at high temperatures generated bymicrowave energy, it could be considered as a compo-sitionally changing ceramic system. It is suggested thata model similar to the one proposed for silicon nitri-dation could be appropriate in explaining the observeddifference in the microwave absorption nature betweenpure and partially oxidized SiC powder compacts.Temperature measurements indicated that a zirconiasample reached the highest temperature in the hybridsystem. The temperature increase of zirconia at lowtemperatures (B800°C) is primarily by thermal radia-tion from the susceptor. At higher temperatures radia-tion cooling of zirconia occurs, while the continuingtemperature increase is due to microwave power ab-sorption by the zirconia. Using this simple hybridheating unit, it was possible to sinter small samples ofyttria stabilized zirconia and 99% alumina to densitiesabove 95%.

Fig. 9b was about 1 mm, with no exaggerated graingrowth but some porosity at the triple junctions.

Microwave sintering curves for Al2O3 and YZA

Fig. 10. Density change with sintering time for YZA (at 1340°C) and Al2O3 specimens.

P.D. Ramesh et al. / Materials Science and Engineering A266 (1999) 211–220220

Acknowledgements

The authors acknowledge the Wolfson Centre forInterface Science for providing a research fellowship(PDR) and the Faculty of Materials Engineering forproviding the experimental facilities. The research wassupported by AFIRST. L. Schachter’s research on mi-crowave ceramic sintering was supported by the IsraeliMinistry of Science.

References

[1] W.H. Sutton, Ceramic Bull. 68 (2) (1989) 376.[2] P.D. Ramesh, K.J. Rao, Adv. Mater. 7 (2) (1995) 177.[3] P.D. Ramesh, Ph. D thesis, Indian Institute of Science, Banga-

lore, 1995.[4] K.J. Rao, P.D. Ramesh, Bull. Mater. Sci. 18 (4) (1995) 447.[5] Y.L. Tian, B.S. Li, J.L. Shi, Y.P. Xu, J.K. Guo, D.S. Yen, in:

D.E. Clark, F.D. Gac, W.H. Sutton (Eds.), Ceram. Trans. 21,The American Ceramic Society, Westerville, OH, 1991, p. 577.

[6] J.D. Katz, R.D. Blake, J.J. Petrovic, Ceram. Eng. Sci. Proc. 9(1988) 725.

[7] M.F. Iskander, MRS Bull. 18 (1993) 30.[8] K.G. Ayappa, Rev. Chem. Eng. (in press).[9] W.H. Sutton, in: W.H. Sutton, M.H. Brooks, I.J. Chabinsky

(Eds.), Mater. Res. Soc. Proc. 124, Materials Research. Society,Pittsburgh, PA, 1988, p. 287.

[10] M.P. Sweeny, D.L. Johnson, in: D.E. Clark, F.D. Gac, W.H.Sutton (Eds.), Ceram. Trans. 21, The American Ceramic Society,Westerville, OH, 1991, p. 365.

[11] A. De’I. Ahmad, E.D. Whitney, D.E. Clark, in: D.E. Clark,F.D. Gac, W.H. Sutton (Eds.), Ceram. Trans. 21, The AmericanCeramic Society, Westerville, OH, 1991, p. 319.

[12] M.A. Janney, C.L. Calhoun, H.D. Kimrey, J. Am. Ceram. Soc.75 (2) (1992) 341.

[13] R.D. Blake, J.D. Katz, in: D.E. Clark, J.R. Laia, W.R. Tinga(Eds.), Ceram. Trans. 36, The American Ceramic Society, West-erville, Westerville, OH, 1993, p. 459.

[14] M.S. Spotz, D.J. Skamser, D.L. Johnson, J. Am. Ceram. Soc. 78(2) (1995) 1041.

[15] A.C. Metaxas, R.J. Meredith, Industrial microwave heating,Peter Peregrinus, London, 1983.

[16] J. Batt, J.G.P. Binner, T.E. Cross, N.R. Greenacre, M.G.Hamlyn, R.M. Hutcheon, W.H. Sutton, C.M. Weil, in: D.E.Clark, D.C. Folz, S.J. Oda, R. Silbeglitt (Eds.), Ceram. Trans.59, The American Ceramic Society, Westerville, OH, 1995, p.243.

[17] T. Carisey, A. Laugier-Werth, D.G. Brandon, J. Eur. Ceram.Soc. 15 (1995) 1.

[18] K.L. Luthra, J. Am. Ceram. Soc. 74 (1991) 1095.[19] D.J. Skamser, J.J. Thomas, H.M. Jennings, D.L. Johnson, J.

Mater. Res. 10 (12) (1995) 3160.[20] J.J. Thomas, D.J. Skamser, H.M. Jennings, D.L. Johnson, J.

Am. Ceram. Soc. 79 (1996) 2458.[21] P.S. Neelakanta, J.C. Park, IEEE Trans. Microwave Theory

Techn. 43 (6) (1995) 1381.[22] M.A. Janney, H.D. Kimrey Jr., in: G.L. Messing, E.R. Fuller

Jr., H. Hausner (Eds.), Ceram. Trans. Ceramic Powder ScienceIIB, The American Ceramic Society, Westerville, OH, 1988, p.919.

[23] M.A. Janney, C.L. Calhoun, H.D. Kimrey, in: D.E. Clark, F.D.Gac, W.H. Sutton (Eds.), Ceram. Trans. 21, The AmericanCeramic Society, Westerville, OH, 1991, p. 311.

[24] Y. Fang, D.K. Agrawal, D.M. Roy, R. Roy, J. Mater. Res. 9(1994) 180.

[25] A. De’I. Ahmad, E.D. Whitney, D.E. Clark, Ceram. Eng. Sci.Proc. 11 (1990) 1743.

[26] S.A. Nightingale, D.P. Dunne, H.K. Worner, J. Mater. Sci. 31(1996) 5039.

.