PrQsiding Pertemuan Ilmiah Sains Materi III$erpong, 20 -21 Oktoher 1998 ISSN 1410-2Bi7c

SYNTHESIS OF FUNCTIONALLY-GRADED CERAMIC MATERIALS!

VIA LIQUID INFILTRATION TECHNIQUE Sd6

OJ, s. P~apa1, LM. Lowz, B.H. O'Connor, L. Atmaja3I Physics Department, Faculty of Mathematics and Science, Institute of Technologi Sepuluh Nopember (ITS), Surabaya

2 Applied Physics Department, Curtin University of Technolofy, Perth, Western Australia3 Chemistry Depa~~nt,Faculty of Mathematics and Science, Institute of Technologi Sepuluh Nopember (ITS), Surabaya

ABSTRACT

SYNTHESIS OF FUNCTIONALLY-GRADED CERAMIC MATERIALS VIA LIQUm INFILTRATION TECH-NIQUE. Liquid infiltration technique has been used to synthesise several functionally-graded ceramic materials. In this process,

powder as the matrix was uniaxially pressed and then presintered at approximately 1000 °c to obtain green bodies with porosity

of 40-45% to increase their strength for liquid infiltration. The green bodies were then fully immersed with suitable liquids. Theinfiltrated green bodies were sintered at high temperature to allow the formation of new phases and densification of the ceramics.The technique has been applied to produce functionally-graded mullite-aluminiun titanate/alumina, aluminium titanate/zirconia-alumina, and zirconium titanate/zirconia composites. The graded profile of each material was characterised using Rietveld x-raydiffraction technique, microanalysis, and electron microscopy. Results showed that the functionally-graded ceramic compositescan be produced by infiltration method.

ABSTRAK

SINTESIS BAHAN KERAMIK UBAHAN GRADUAL DENGAN METODE INFILTRASI CAIR. MetOOe infiltrasicair telah dimanfaatkan untuk mensintesis beberapa bahan ubahan gradual. Pada proses ini. serbuk matriks dipres secara uniaksial

dan diprasinter pada suhu 1000 °c untuk mendapatkan prakeramik dengan porositas antara 40-45% dengan kekuatan yang

memadai untuk infiltrasi. Kemudian prakeramik dicelupkan ke dalam cairan prekursor yang sesuai. Prakeramik yang telahdiinfiltrasi kemudian disinter pada suhu tinggi untuk mendapatkan rasa barn dan membentuk keramik parlato Teknik ini telahdikembangkan untuk mensintesis bahan ubahan gradual mullite-aluminium titanatJalumina. aluminium titanatJzirkonia-alumina.dan zirkonium titanatJzirkonia. Karakter gradual pada masing-masing bahan disintesis dengan menggunakan metode Rietvelddifraksi sinar-x. mikroanalisis. dan mikroskopi elektron. Hasil-hasil karakterisasi menunjukkan bahwa teknik infiltrasi memilikipotensi sebagai metode sintesis bahan ubahan gradual.

KEYWORDS

Infiltration, Functionally-graded, Mullite, Aluminium titanate, Zirconium titanate, Alumina, Zirconia, Rietveld analysis.

INTRODUCTIONwas originated by Japanese scientists led by Niino in1984 [2]. Beginning with the development for mechanicaland thermal properties improvement, FGM has foundapplication in broader fields which include electronicsand optics. Methods for producing FGM have alsoadvanced, one of which is liquid infiltration technique.Type of FGM can be combination of two or more basictype of material, ie metals, ceramics, and polymers.

The idea of producing functionally-gradedceramic materials via liquid infiltration technique was in-spired by the success of synthesising such materialsby sol-gel process. In this process, a powder is mixedwith a solution containing a phase precursor. When themixture is processed, usually by firing, formation of anew phase is allowed or a new phase is introduced intothe former matrix resulting in a functionally-graded com-posite. An alumina! zirconia functionally-graded

In the past, nxtuirements for industry were mainlyfocused on materials which exhibit uniform properties.Among them are dispersed, laminated, and fiber-reinforced composites. There is no doubt that thesematerials are still required. However, a novel type ofcomposite has been developed, termed as functionalIy-graded materials or functional gradient materials (FGM).This type of material displays character which is spatiallyinhomogeneous but changes between spatial positionsare continuous. From phase distribution point of view,FGM lies between homogeneous or dispersed andlaminate composites. It has been shown that thecontinuous change of characters in FGM providessuperior macroscopic properties over other type ofmaterials, especially when residual thermal stresses istaken into account [1]. The idea of synthesising FGM

207Suminar Pratapa dkk.

Prosiding Pertemuan Ilmiah Sains Materi 111Serpong, 20- 21 Oktoher 1998 ISSN 1410-2897

I

material can be produced [3] by infiltrating a solutionof Al(NO')3. 9H2O into yttria-tetragonal zirconia poly-crystal (Y -TZP). Other systems which have been pro-duced include mullite/alumina with TEOS as infiltrant[4], mullite/ZTA with TEOS as infiltrant [5], and alu-minium titanate/alumina with TEOT as infiltrant [6].

In this paper, synthesis of three other types offunctionally-graded ceramic materials which have beenproduced via infiltration technique is reviewed. Theyare mullite-aluminium titanate (A1)/ZTA with TEOS andTEOT as infiltrants [7], AT/nroonia-alumina [8], and zir-conium titanate (ZT)/zirconia [9]. Infiltrant for the latertwo materials was a solution containing titanium cWo-ride. This paJx:r also reports the characterisation of thematerial to reveal their graded character.

Australian Standard AS 1774.5 (1989): TheDetenninationof Density, Porosity, and Water Absorption [10].

Infiltration process was conducted bY fully im-mersing the presintered ceramic with the certain solu-tion for 1 -24 hours. In some cases, control samples,ie. uninfiltrated samples. were also made. The sampleswere then sintered at high temperature to allow the for-mation of new phases and densification of the ceramic.The synthesis of the FGM is then finalised. The gen-eral procedure of synthesising a FGM can be illustratedusing a flow-chart as shown by Figure I.

Characterisation of the graded profile in theFGMs was made using x-ray diffractometry, microanaly-sis and electron microscopy. The x-ray diffraction tech-nique was not only used to monitor the formation ofnew phases but also to determine the phase composi-tion within the samples. This quantitative phase analy-sis was performed using the whole-pattern Rietveldphase analysis LHPM (Hill and Howard 1986). Meth-ods of calculating weight fractions using Rietveld'ZMV' approach following Pratapa et al. [11] and Hilland Howard [12] were used.

EXPERIMENTAL METHOD

Raw materials used are tabulated as in Table 1.The table also shows the shape. dimensions. sinteringtemperature and porosity of both green body and as-fired ceramics.

The powder for each system was wet-milled withpure water as the wetting media. Certain amount ofNH.OH was added into each mixture before milling. Thepurpose of adding NH.OH was to take up the pH up to10-12 in order to facilitate the break up of the spray-dried grains. The as-milled mixture was dried in an ovenat 100 °C for 24-48 hours. The dried mixture was thensieved until free-flowing with a consecutive grid-sizesieves.

RESULTS AND DISCUSSION

There are two indications which can be used topredict that the synthesis is successful, ie. mass gain-ing and x-ray diffraction pattern from the smface of eachsample. Results showed that the green bodies hadgained a certain amount of precursor indicated by thepresence of positive mass difference between as-firedand uninfiltrated samples. Mass increase of 4.2% wasfound both in the AT/zirconia-alumina (24 hours immer-sion) and ZT/zirconia (1 hour immersion) systems. Thex-ray diffraction pattern showed that new phase(s)formed in the material. The new phases were believed

Uniaxial press in a metal die was applied into thepowder mixture to produce the shape. A partial sintering

at 1000 °c was then used to increase its strength butmaintain the porosity prior to infiltration. The apparentporosity of the samples was measured following the

Table Raw materials, shape, dimension, porosity and given heat treatment in the processing oftheFGMs.

No

fSystem Powder shapelDlfnensionIPorositYISintering Tern(%) perat~r~ (~CL

ilnfiltrant

mullite-AT/ZTA

alumma-Zlfcorna(85:15% by_~ei~ht)

~

TEDSand TEOT

13.5mm diam2mmheil1.ht

35-40 1550for 3hours

aisc

I

A T/Zi~conial-alumina

alumina-zirconia(90: 10% by

wei~ht)

solution

containingtitaniumchloride

45 1550for 3hours

oar :immx 12 mmx 60 mm

Zf7Zlfconiai solution

containingtitaniumchloride

13.5mm diam3mmh~~

44 1350for 2hours

OISCZlfcoma

i<unstabilised)

Note: TEOS: tetraethyl orthosilicate as precusor of SiO2TEOT: tetraethyl orthotitanate as precursor of TiO2Solution containing titanium c\oride as precursor of TiO2Porosity: for the green bodies.

Prosiding Pertemuan Ilmiah Sains Materi IIISerpong, 20 -21 Oktober 1998 ISSN 1410-2897

to be formed via sintering reaction between the matrixand the introduced precursor. The reaction occured ineach system can be summarised as follows.1). inthemullite-AT/ZTAsystem:

3 Al2O) + 2SiO2 ~ 3Al2O).2SiO2 (1)

(mullite)Al2O) + TiO2 ~ Al2 TiOs (2)

2). in the AT/zirconia-alumina system:AI2O) + TiOz ~ Al2 TiO,

3) .in the ZT/zirconia system:

zrO2 +TiO2 ~ZrTiO4 (3)While the successful synthesis can only be pre-

.Infiltration of green bodies with suitableinfiltrant(s)

Sintering at high temperature for morethen 2 hours allowing new phase

formation and densification

Functionally grade Materials (FGMs)are produce

Figure I. Liquid infiltration steps to produce func-

tionally-graded materials (FGMs).

were three meas~res from the refinement which wereused to determine the completion of each refinement.Firstly, figures-of-merit from the refinement show thatthe refinement are acceptable [8,9,11], ie the Bragg R-factor are generally less than 10% and the "goodnessof fit" (GoF) are less than 3%. The second measure isthe difference plot which also indicates that the rmne-ments are acceptable since the difference profiles arenegligible. Representative difference plot from the AT/zirconia-alwnina FGM is depicted in Figure 4. The finalmeasure is the standard deviation of each refined pa-rameters which is generally less than 10%. The successin the refinement led to the ability to make use of thecell parameters and scale factor for calculating thtweight fraction of each phase.

In the AT/zirconia-aIwnina system, the absoluteweight fraction of AT is 44.5 wt% on the surface andreduces with depth to 9.5 wt% at 0.3 rom, and then to5.3 wfO/o at 1.5 rom. On the other hand, the aIwnina con-tent increases with depth from 44.4 wfO/o at the surfaceto 80.2 wfO/o at 0.3 mm, and then to 85.7 wfO/o at 1.5 rom.

The quantitative Rietveld phase analysis for theZT/zirconia system showed that the weight fraction ofZT reduces with depth ie from 12.2% at the surface to8.4% at 0.2 rom, then to 4.4% at 0.5 rom, and finally to0% at the center of the FGM. On the contrary, theweight fraction of zirconia increases with depth, ie from87.8% at the surface to 91.6%atO.2 mm. then to 95.6%at 0.5 mm and finally to 100% at the center.

The continuous change of the phase content inthe FGMs can be expressed as follows (following ex-pression derivation given by Hirai [1] and Markworthet aI. [13J.1). AT/zirconia-aIwnina system [8J:

a.ATVAT =-194.0d+49.3 O.O~d~O.2

=-5.75d+ll.6 O.2£d£1(4)

dicted by those two approaches, the presence of thegraded profile in the materials can be observed bydiffmctometry, microanalysis or electron microscopy. X-my diffraction patterns were collected from variousdepths in order to, both qualitatively and quantitatively,see the graded composition within the materials. X-mydiffraction patterns of the AT/zirconia-alumina and Zf/zirconia systems are shown by Figure 2 dan 3. As canbe seen from the figures, the AT and ZT peaks reducein height with depth. By contrast, the alumina and zir-conia peak intensities increase with depth. This resultindicates the qualitative gradual composition in thesamples. Detail information about the refinement andcalculation can be found in other reports [8,11].

Applying Rietveld refinement into the collectedpatterns resulted in the quantitative composition. There

Suminar Pratapa dkk. 209

Prosiding Pertemuan Ilmiah Sains Materi IIISerpong, 20 -2I Oktober 1998 ISSN1410-2897

0

';::

(6)

2). ZTlzirconia system [9]:a. zr

Wzr= 7,1 cP -19.2 d+ 12.1b. ztOz

W = -7,1 cP+ 19.2 d+77.9

where W is weight fraction.(7)

Figure 5. X-ray characteristic ennnisions of TiKa,AIKa, and ZrL a from the AT/zirconia-alumina system at several depths.

around surface is much sharper than that in the latersample. This difference is believed to be caused by thedifferent infiltration time.

From the x-ray diffraction measurements, it canbe claimed that the infiltration technique is one powerfulways for producing FGMs, especially ceramic-cerarnicsystem. The diffraction method, on the other side, is apotential technique for revealing the compositiondistribution within FGMs.

There are other techniques can be used to revealthe qualitative graded profile of a FGM, ie. microanalysisand electron microscopy. They have been utilised toassess such profile in the mullite-AT/ZTA and AT/zirconia-alumina systems. Making use the microanalysis,Ti and Si dot map and their characteristic x-ray intensitiescan be acquired from the cross-sectioned surface of theFGMs (pratapa and Low 1997, Pratapa 1997). Figure 5

shows the x-ray characteristic emissions ofTiKa., AtKa.,

and ZrLa with depth of the AT/zirconia-alumina samplemeasured using energy-dispersive x-ray microanalysis.

The figure displays the graded character in the TiKa

Figure 3. X-ray diffraction patterns from the ZT/zirconia system at several depths, (a) 0 mm,(b) 0.3 mm, (c) 0.7 mm. Z: zirconia, ZT:

zirconium titanate

These expressions show that a graded com-position has occurred both in the AT/zirconia-aluminaand ZT/zirconia systems. They also show a slightdifference in the trend of phase content culVes. Whilethe culVe for the later system is simply parabolic, that forthe former is divided into two linear cUIVes. In the formersample, the gradual change in composition within area

Suminar Pratapa dkk.210

Prosiding Perlemuan Ilmiah Sains Materi IIISerpong, 20 -21 Oktober 1998 ISSN 1410-2897

of 600 I.lIn and finally to 3.7 I.1m at depth of 1000 I.1m. This

implies that the presence of AT tends to control the grain

growth of alumina.

CONCLUDING REMARK

It can be concluded from the studies that infil-tration method has been successfully applied to pro-duce functionally-graded mullite-AT/ZTA, AT/zirconia-alumina, ZT/zirconia ceramic materials. Rietveld x-raydiffraction method has been shown as a novel and po-tential way in determining the graded composition inFGMs. Other techniques, such as microanalysis andelectron microscopy, can also be used to characterisethe graded profile of FGMs.

Figure 6a. Back-scattered SEM micrographs of the AT!

zirconia-alumina sample at 20 Jlm depth.

ACKNOWLEDGEMENT

S.P. thanks AusAID for providing MSc scholar-ship and Higher Education Project, ADB-Loan INO-1253for funding the research on the zirconium titanatlzirco-nia system.

REFERENCES

[1]. T.lllRAI (1996). Functional Gradient Materials. InPrecessing of Ceramics (part 2) (ed. Brook, R. J.).VCH VerlagsgeseUscham mbH, Weinheim.

[2]. M. KOIZUMI (1992). The Concept ofFGM. InSecond International Symposium on FunctionallyGradient Materials (ed. Holt, J. B., Koizumi, M.,Hirai, T. andMunir,Z.A.),pp. 3-10. The AmericanCeramic Society, WesterviIle, Ohio, San Fransisco,CA.S.J. GLASS AND D.J. GREEN (1987). SurfaceModification of Ceramics by Partial Infiltration.Adv. Ceram. Mat. 2(2),29-31.B.RMARPLEANDD.J. GREEN. (1990).MuIIite/Alumina Particulate Composites by InfiltrationProcessing: Ill, Infiltration and Characterization.J.Am. Ceram. Soc. 73(12),3611-3616.I.M. LOW, R SKALA, R RICHARDS AND D.S.PERERA (1993). Synthesis and Properties of NovelMullite-Zirconia-toughened Alumina CompositesJ.Matet: Sci. Left.n, 1985-1987.

--.' I.M. LOW, R.D. SKALA AND D. ZHOU (1996).Synthesis of FunctionaUy gradient Aluminiumtitanate! Alumina Composites. J. Matet: Sci. Lett.15,345-347.S. PRATAPA AND I.M. LOW (1996). Synthesisand Properties of Functionally-gradient Aluminiumtitanate-MulIite-ZTA Composites. J. Matet: Sci.Lett. 15, 800-802.

I. S. PRATAPA (1997). "Synthesis and Character ofa Functionally-graded Aluminium Titanate!Zirconia-Alumina Composite". M.Sc. Thesis,Curtin University of Technology, Perth, Western

Figure 6b. Back-scattered SEM micrographs of the AT!zirconia-alumina sample at 300 ~m depth.

[3].

[4].

[5]

[6)Figure 6c. Back-scattered SEM micrographs of the AT/

zirconia-alumina sample at 600 11m depth.

[7).intensity which relates the graded AT content within thematerial.

The use of electron microscope with back-scattering mode has shown that the amount of AT grainsreduced with depth while that of alumina increased withdepth (Pratapa 1997). Representative results arepresented in Figure 6. More interestingly, the grain size

of alumina increased with depth, ie 2.1 ~m at depth of20

~m to 3. 0 ~m at depth of 300 ~m then to 3.2 J.Un at depth

[8]

Suminar Pratapa dkk. 211

Prosiding Pertemuan Ilmiah Sains Materi IIISerpong, 20- 21 Oktoher 1998 ISSN 1410-2897

Australia.191. S. PRATAPAandL. ATMAJA(1998)"Synthesis

and Characterisation of Functionally-gradedZrTiO 4/ZrO2 Ceramic Materials via InfiltrationMethod", "Starter Grant" Research Project Report,ADB Loan 1253-INO, Contract No. 237/HEP/X!

C.SP.97/SG (20 Octckr 1997).[10]. AUSTRALIAN STANDARDS (1989). Refractories

and Refractory Materials -Physical Test Methods,Method 5: The Detennination ofDensity, Porosityand Water Absorption: 1774.5, Standard Australia.

[11]. S. PRATAPA. B.H. O'CONNOR, AND I.M.WW(1998). Use of Compton Scatter Measurements forAttenuation Corrections in Rietveld PhaseAnalysis with an External Phase CompositionStandard. Powder Diffraction, 13(1), 1-5

[12]. RJ. H1l1.. ANDC.J. HOWARD (1987). QuantitativePhase Analysis from Neutron Powder DiffractionData Using the Rietveld Method. J. App/. Cryst.

20,467-474.[13]. A.J. MARKWORrn. K.S. RAMESH AND W.P.

PARKS, JR. (1995). Review Modelling StudiesApplied to Functionally Graded Materials. J.Matet: Sci. 30, 2183-2193.