Microwave-assisted preparation and sintering of mullite and mullite-zirconia composites from metal organics

Article abstract of DOI:10.1016/S0025-5408(98)00135-4

Decomposition characteristics of a mixture of aluminum-tri-sec butoxide (ASB) and tetraethyl orthosilicate (TEOS) in ethyl acetate solution by microwave pyrolysis was examined. The decomposed product was used as a second-stage precursor in the preparation of pure mullite (3Al₂O₃·2SiO₂). Mullite-zirconia composites were obtained using the above mixture along with zirconium-iso-propoxide as starting materials. Pellets of mullite and mullite-zirconia composites were sintered in a microwave oven using β-SiC as a secondary heater.

Full text of DOI:10.1016/S0025-5408(98)00135-4

Materials Research Bulletin, Vol. 33, No. 10, pp. 1527–1536, 1998
Copyright © 1998 Elsevier Science Ltd
Printed in the USA. All rights reserved
0025-5408/98 $19.00 ϩ .00
PII S0025-5408(98)00135-4
MICROWAVE-ASSISTED PREPARATION AND SINTERING OF MULLITE
AND MULLITE–ZIRCONIA COMPOSITES FROM METAL ORGANICS
B.G. Ravi, V. Praveen†, M. Panneer Selvam, and K.J. Rao*
Solid State and Structural Chemistry Unit, Indian Institute of Science,
Bangalore 560 012, India
(Refereed)
(Received February 9, 1998; Accepted February 19, 1998)
ABSTRACT
Decomposition characteristics of a mixture of aluminum-tri-sec butoxide
(ASB) and tetraethyl orthosilicate (TEOS) in ethyl acetate solution by micro-
wave pyrolysis was examined. The decomposed product was used as a
second-stage precursor in the preparation of pure mullite (3Al2O3⅐2SiO2).
Mullite–zirconia composites were obtained, using the above mixture along
with zirconium-iso-propoxide as starting materials. Pellets of mullite and
mullite–zirconia composites were sintered in a microwave oven using ␤-SiC
as a secondary heater. © 1998 Elsevier Science Ltd
KEYWORDS: A. ceramics, A. composites, C. thermogravimetric analysis
(TGA), C. X-ray diffraction, D. microstructure
INTRODUCTION
There has been considerable interest in mullite and mullite-matrix composites in recent years
due to their unique properties, such as low thermal expansion, low dielectric constant, high
melting point, remarkable chemical resistance, high hardness, and excellent creep resistance
[1]. These properties have been optimized using a variety of methods for mullite synthesis,
for example, sol-gel [2], coprecipitation [3], hydrolysis [4], spray pyrolysis [5], chemical
vapor deposition [6], and thermal decomposition [7].
*To whom correspondence should be addressed.
†Undergraduate Research Scholar under exchange program.
1527

1528
B.G. RAVI et al.
Vol. 33, No. 10
Use of microwaves in the synthesis and processing of ceramics has attracted much
attention recently because it offers a novel synthetic pathway notable for features such as
rapid heating rates, short processing durations, low power requirements, and product uni-
formity [8–14]. Synthesis of ␤-SiC [15], AlN [16], MoO2 [17], and several other materials
[13] by the microwave route has already been reported by this laboratory. We also have
reported achieving very rapid sintering of ZrO2–CeO2 ceramics using microwave irradiation
[18]. Recently, aluminum-tri-sec butoxide and zirconium-iso-propoxide have been decom-
posed by microwaves to form secondary precursors for the preparation of composites [19].
Direct pyrolysis of metal organics by conventional heating is unsuitable for powder
production because it gives rise to nonuniform powder morphology. Several other techno-
logical problems have been reported to be associated with combustion methods [20].
However, when metal-organic liquids are subjected to microwave irradiation, very rapid and
volumetric heating occurs, which leads to their decomposition. This is expected to result in
the formation of products of good structural and, eventually, of morphological uniformity.
Microwave pyrolysis is superior to conventional heating since, in most instances, it behaves as a
self-regulating process (microwave absorption characteristics change drastically once the suscep-
tor is either chemically or physically altered) and, therefore, products having uniform particle size
distribution can be expected to form. Different precursors in a mixture exhibit different suscep-
tibilities and, therefore, lead to different thermal histories at a microscopic level. This can be
exploited in the preparation of ceramic composites with unique morphologies.
Very little is known in the literature with regard to the use of microwaves in the
preparation of ceramic powders and composites starting from organic precursors. In this
paper we report our investigations on the use of organic precursors in the microwave-assisted
preparation and sintering of mullite and mullite-zirconia composites.
Another important aspect of this work is our investigation into microwave-assisted sin-
tering of mullite and mullite–zirconia composites. Pure mullite is difficult to densify at
temperatures below 1973 K, due to the slowness of diffusion of aluminum and silicon in the
product [21]. However, it is well known that the addition of ZrO2 promotes densification and
retards grain growth of the mullite phase [22]. Different processing routes for obtaining
mullite–zirconia ceramics of high densities have been reported in the literature, such proce-
dures were reviewed in 1994 by Schneider et al. [1].
Mullite and mullite–zirconia mixtures are generally microwave inert at low temperatures, but
they have been found to sinter rapidly in microwaves when irradiated at higher temperatures.
Higher temperatures are achieved by a technique based on using a secondary heater in the
microwave field [18]. In this “hitch-hiking” approach, materials which are inert to microwaves at
low temperatures are embedded in another, strongly microwave-active material (secondary
heater) and then subjected to microwave irradiation. The critical requirement is that the material
of the secondary heater be chemically unreactive with the material being sintered [18].
EXPERIMENTAL DETAILS
For the microwave heating experiment, we employed a microwave oven (Batliboy Eddy,
India) operating at a frequency of 2.45 GHz and maximum output power of 980 W. We
selected, as the organic precursors, aluminum-tri-sec butoxide (C4H9O)3Al, tetraethyl ortho-
silicate (C2H5O)4Si (TEOS), and zirconium-iso-propoxide (C3H7O)4Zr. For the preparation
of mullite, a calculated quantity of aluminum-tri-sec butoxide and TEOS were mixed with

Vol. 33, No. 10
MULLITE ZIRCONIA
TABLE 1
1529
Microwave Preparation Conditions and the Compounds Formed
Exposure
Microwave duration
Starting material
power (W)
(min)
Compound formed
Al2O3 (A)
(C4H9O)3⅐Al
(C3H7O)4⅐Zr
560
560
560
560
28
14
30
30
ZrO2 (Z)
(C4H9O)3⅐Al ϩ (C2H5O)4Si ϩ CH3COOC2H5
(C4H9O)3⅐Al ϩ (C2H5O)4Si ϩ (C3H7O)4⅐Zr ϩ
CH3COOC2H5
3Al2O3⅐2SiO2 (M)
(3Al2O3⅐2SiO)⅐ZrO2 (MZ)
a
aFour samples were prepared with different proportions of zirconia (up to 40 mol%, in steps of 10
mol%).
ethyl acetate to obtain a clear solution. For the preparation of mullite–zirconia composites,
calculated quantities of zirconium-iso-propoxide were added to the ethyl acetate solution
containing aluminum-tri-sec butoxide and TEOS. These mixtures were put into glass beakers
and irradiated with microwaves for 15 min. Within a few minutes of initiating irradiation, a
rapid increase in the viscosity of the metal-organic precursors was observed, and foam-like
products were found to form in the beaker within a short duration of further irradiation. At
this stage, the microwave oven was switched off and the foam-like powder was removed by
gently tapping the beaker. To understand the decomposition characteristics of the precursor
components, they were also individually decomposed by microwave pyrolysis. TEOS could
not be pyrolyzed directly by microwaves because of its high volatility. Hence, TEOS solution
was pre-polymerized by adding water. The experimental details of the microwave pyrolysis
are given in Table 1.
Powders obtained from the microwave pyrolysis were characterized using X-ray diffrac-
tion (XRD) studies (Jeol), thermogravimetric analysis (TGA), and differential thermal anal-
ysis (DTA; Cahn Instruments), Fourier transform infrared spectroscopy (Biorad). Powders
were heated in a furnace (Thermolyne 46100) to various temperatures and then examined by
X-ray diffraction.
Sintering of mullite and mullite–zirconia composites was performed in the same microwave
oven. For the sintering experiments, mullite and mullite–zirconia composite powders were first
pelletized under 196 MN/m2 pressure to obtain 10 mm diameter pellets of 3 to 5 mm thickness.
Polyvinyl alcohol (1%) was used as a binder. The pellets were embedded in ␤-SiC powders in a
silica crucible (about 10 g of ␤-SiC was sufficient) and the crucible was kept inside the
microwave oven. With this set-up, about 25 min was required for complete sintering of the pellets.
The products were examined by transmission electron microscopy (TEM; Jeol 200CX) and
scanning electron microscopy (SEM; Cambridge Instruments, Stereoscan 360).
RESULTS AND DISCUSSION
Microwave-Assisted Preparation. We designated as the mullite first stage (MF) the
decomposition product obtained from the mixture of aluminum-tri-sec butoxide and TEOS in
ethyl acetate solution by microwave irradiation after 15 min. MF is a foam-like powder.
Figure 1 shows the combined DTA and TGA of MF. From TGA data we note that a large

1530
B.G. RAVI et al.
Vol. 33, No. 10
FIG. 1
Combined TGA and DTA plots for microwave-obtained (MF) powder.
weight loss (55%) occurred before 793 K. The DTA for the product exhibits a sharp
exothermic peak at 1248 K, which is associated with crystallization of mullite. The XRD
pattern (Fig. 2) shows that the microwave-decomposed powder (MF) is amorphous. Heating
to a temperature of about 1273 K resulted in a sharp and complete crystallization of mullite.
Heating above 1773 K produced very few changes in the XRD patterns. No other phases,
such as aluminosilicate spinel, ␣-alumina, or cristobalite, were detected in the calcined
samples at any temperature.
FIG. 2
X-ray diffractogram of MF calcined at various temperatures.

Vol. 33, No. 10
MULLITE ZIRCONIA
1531
FIG. 3
IR spectra of MF powder calcined at (a) 300 K, (b) 1249 K, (c) 1473 K, (d) 1573 K, (e) 1673
K, and (f) 1773 K.
There have been several reports on the preparation of mullite starting from aluminum and
silicon alkoxides [5,7,21,22]. DTA of these products exhibited a sharp exothermic peak at
Х1273 K. However, XRD patterns of the phases after the exothermic peaks indicated the
presence of a small amount of aluminosilicate spinel phase along with mullite. It is well
known that the crystallization behavior of mullite precursor depends on the scale of chemical
homogeneity of Al and Si atoms [23,24]. When the scale of chemical homogeneity of the
precursor is in the nanometer or higher range (diphasic gel), Al2O3 and SiO2 crystallize
before mullite formation [25]. On the other hand, mullite crystallizes directly at about 1273
K from the precursors if there is atomic-scale mixing of Al and Si atoms in them [4,26,27].
Since the present product exhibited a sharp exothermic peak at about 1273 K, due to
crystallization of mullite, the chemical homogeneity of MF can be regarded as high and of
atomic scale.
Figure 3 shows the IR spectra of microwave-obtained mullite powder heat-treated at

1532
B.G. RAVI et al.
Vol. 33, No. 10
FIG. 4
(a) TEM image of MF calcined at 1773 K for 3 h and (b) corresponding electron diffraction
pattern projected from (001).
different temperatures. The spectra of as-produced amorphous powders are different from
those of amorphous SiO2 powders and bohemite gel [4]. There is a recognizable similarity
between the spectra of amorphous powder and the crystallized mullite (1273 K). For
example, in Figure 3 the broad peak in the region of 600 cmϪ1 and the sharp peak at 1010
cmϪ1 can be correlated with 550 and 1150 cmϪ1 peaks in crystallized mullite. Therefore, the
as-produced mullite powders are unlikely to be simple mixtures of SiO2 and aluminum
hydroxide. They seem to be already close to the chemistry of mullite. Therefore, Si–O–Al
bonding has already occurred in the structure. The absorption band at 1041 cmϪ1 in the IR
spectrum of the as-produced powder corresponds to ␯
red-shifted absorption at 1100
Si-O
cmϪ1 normally present in the spectrum of SiO2. The lowering of the frequency results from
the introduction of Si–O–Al bonds [4] in place of stiffer Si–O–Si bonds. Structural changes
occurring in the particles during heating can be characterized by shifts of the ␯Si-O band. The
absorption band in the 500 to 800 cmϪ1 region is associated with both octahedral and
tetrahedral Al–O bonds; this agrees well with the calculated data on ideal mullite [28].
Figure 4 shows the TEM image (a) and the corresponding diffraction pattern (b) of mullite
powders heat-treated at 1773 K for 3 h. This pattern confirms the completion of mullite
formation and of the faceted submicron-size crystallites. The d values calculated from

Vol. 33, No. 10
MULLITE ZIRCONIA
1533
FIG. 5
X-ray diffractograms of microwave-obtained mullite–zirconia powders heat-treated at 1773
K for 3 h: (a) MZ1, (b) MZ2, (c) MZ3, and (d) MZ4 (see Table 1 for definition of MZ1, MZ2,
MZ3, and MZ4). M ϭ mullite; Zm ϭ monoclinic ZrO2; Zt ϭ tetragonal ZrO2.
electron microscopy (d100 ϭ 7.6 Å; d010 ϭ 7.7 Å and d001 ϭ 2.9 Å) agree well with the XRD
values obtained in the present work and from the literature [31,32].
Mullite–zirconia composites were obtained by the microwave decomposition of mixtures
of ASB, TEOS, and zirconium-iso-propoxide in ethyl acetate solution. This process yielded
a foam-like powder designated as mullite–zirconia foam (MZF). XRD results show that these
powders were amorphous. The foam-like powder was used as a second-stage precursor for
the preparation of mullite–zirconia composites. On calcination at 1773 K in a conventional
TABLE 2
Sintering of Mullite and Mullite-Zirconia Composites by Use of Microwave Irradiation
Sintering
duration
(min)
Green
density
(g/cm3)
Sintered
density
(g/cm3)
Relative
density
(%)
Microwave
power (W)
Temp.
(K)
Sample
Composition
M
Mullite
Mullite
90M ϩ 10ZrO2
80M ϩ 20ZrO2
70M ϩ 30ZrO2
60M ϩ 40ZrO2
980
-
980
980
980
980
35
-
35
35
35
35
1733
-
1733
1733
1733
1733
1.82
1.82
1.94
2.05
2.15
2.27
2.95
3.00
3.21
3.38
3.59
3.37
92
94
96
96
97
95
MCa
MZ1
MZ2
MZ3
MZ4
aMullite sintered at 1723 K for 5 h in a conventional furnace.

1534
B.G. RAVI et al.
Vol. 33, No. 10
FIG. 6
SEM images of freshly fractured surfaces of microwave-sintered pellets: (a) M, (b) MC, (c)
MZ1, (d) MZ2, (e) MZ3, and (f) MZ4.
furnace, MZF powders gave rise to mullite–zirconia composites. Four mullite–zirconia
composites were prepared with different percentage of zirconia (10–40 mol%). Figure 5
shows the XRD patterns of MZF powders containing different percentages of ZrO2 heat-
treated to 1773 K for 3 h. In all these composites, zirconia was found to be present mostly
in the tetragonal phase. In ZrO2-rich composites, there is also weak but definite evidence of
the presence of monoclinic ZrO2, and the peak is indicated in Figure 5c and d. Such a
tendency for ZrO2 to transform into monoclinic phase in mullitic phases has been reported
[30].

Vol. 33, No. 10
MULLITE ZIRCONIA
1535
Microwave-Assisted Sintering. ␤-SiC used as a secondary heater initially reached a
temperature of about 1073 K. As a consequence, the temperature of the pellets increased to
nearly 1000 K. Once the pellets were heated to this temperature, direct microwave absorption
by the pellets became significant. The direct coupling of the mullite–zirconia ceramics to
microwaves escalated the temperatures to 1773 K. Efforts to measure the temperature profile
continuously during sintering were unsuccessful, since the thermocouple could not be kept
inserted in the ␤-SiC cover during sintering. Only the microwave power of the oven was kept
constant during sintering. Temperatures of the pellets, however, were measured at various
stages, by pushing a thermocouple (Pt–Pt 13% Rh) to the surface of the pellets while the
microwave power was briefly switched off. The pellets attained their maximum temperatures
in approximately 15 min of exposure and these temperatures were recorded. The densities of
the pellets were measured after 35 min of sintering operation.
Pure mullite could be sintered to only 92% of theoretical density. When the same mullite
pellets were sintered at 1723 K in a conventional furnace for 5 h for comparison, they were
found to sinter to a density of 94%. The effect of adding zirconia on the microwave sintering
was evident in the final densities of the composites. All mullite–zirconia composites were
found to sinter to still higher densities for the same duration of 35 min of sintering. The
sample with 40% of zirconia sintered to a maximum of 97% of theoretical density. Micro-
wave power, duration of irradiation, and final densities of the products in these experiments
are summarized in the Table 2.
The sintered samples were examined using SEM. The micrographs of the fractured
surfaces of the pellets are shown in Figure 6. The microstructure of pure mullite (M) appears
flaky. But, in MZ4 the particles are round and smooth in appearance. A similar appearance
of particles was noticed when alumina and zirconia powders bounded by PVA were sintered
by conventional heating, but the densities produced by sintering were remarkably low [33].
The microstructures of MZ1, MZ2, and, MZ3 were intermediate in appearance. Also particles
were larger in size in sintered M, compared to those in MZ4 and other samples. Larger
particle size has the effect of slightly increasing the enclosed porosity (see densities in Table
2): that of M is larger than that of the composites. But there was no clear distinction in
particle contrasts in the micrographs, i.e., ZrO2 particles could not be distinguished from
those of Al2O3.
CONCLUSIONS
Mullite and mullite–zirconia composites were prepared by a microwave-assisted method
starting from metal-organic liquids. This method involves preparation of second-stage pre-
cursors, which are later heated in a conventional furnace to decompose and crystallize the
products. The products are sintered by microwave heating to high densities. ZrO2 is retained
largely in tetragonal form in the composites.
REFERENCES
1. H. Schneider, K. Okada, and J. Pask, in Mullite and Mullite Ceramics, ed. K. Okada, p. 146, John
Wiley and Sons. Ltd., Chichester, UK (1994).
2. H. Schneider, B. Saruuhan, L. Merwin, and A. Sebald, J. Eur. Ceram. Soc. 11, 87 (1993).
3. K. Okada and N. Otsuka, J. Am Ceram. Soc. 69, 652, (1986).

1536
B.G. RAVI et al.
Vol. 33, No. 10
4. Y. Hirata, K. Sakeda, Y. Matsushits, K. Shimada, and Y. Ishihara, J. Am. Ceram. Soc. 72, 995
(1989).
5. O. Sakurai, N. Mizutani, and M. Kato, J. Ceram. Soc. Jpn. 96, 639 (1988).
6. S. Hori and R. Kurita, Ceram. Trans. 6, 311 (1990).
7. M. Inoue, H. Kominami, and T. Inui, J. Am. Ceram. Soc. 79, 792 (1996).
8. D.R. Baghurst and D.M.P Mingos, J. Chem. Soc., Chem. Commun., 829 (1988).
9. W.H. Sutton, Am. Ceram. Soc. Bull. 68, 376 (1989).
10. C.C. Landry and A.R. Barron, Science 260, 1653 (1993).
11. S. Komarneni, E. Breval, and R. Roy, in Microwave Processing of Materials, ed. W.H. Sutton,
M.H. Brooks, and I.J. Chabinsky, Vol. 124, p. 235, Materials Research Society, Pittsburgh, PA
(1988).
12. T.T. Meek, C.E. Holcombe, and N. Dykes, J. Mater. Sci. Lett. 6, 1060 (1987).
13. K.J. Rao and P.D. Ramesh, Bull. Mater. Sci. 18, 447 (1995).
14. M. Willert-Porada, MRS Bull. 18, 51 (1993).
15. P.D. Ramesh, B. Vaidhyanathan, M. Ganguli, and K.J. Rao, J. Mater. Res. 9, 3025 (1994).
16. P.D. Ramesh, and K.J. Rao, Adv. Mater. 7, 177 (1993).
17. B. Vaidhyanathan, M. Ganguli, and K.J. Rao, J. Mater. Chem. 6, 391 (1996).
18. P.D. Ramesh, P. Sarin, S. Jeevan, and K.J. Rao, J. Mater. Synth. Processing 4, 163 (1996).
19. B.G. Ravi, P.D. Ramesh, N. Gupta, and K.J. Rao, J. Mater. Chem. 7, 2043 (1997).
20. D.W. Sproson, G.L. Messing, and T.J. Gardner, Ceram. Int. 12, 3 (1986).
21. M.D. Sacks and J.A. Pask, J. Am Ceram. Soc. 65, 65 (1982).
22. H. Suzuki and H. Saito, J. Mater. Sci. 25, 2253 (1990).
23. S. Prochazka and F.J. Klug, J. Am. Ceram. Soc. 66, 874 (1983).
24. B.E. Yoldas, in Mullite and Mullite Matrix Composites, ed. S. Somiya, R.F. Davis, and J.K. Pask,
Vol. 6, pp. 225, American Ceramic Society, Westerville, OH (1990).
25. K. Okada, N. Otsuka, and S. Somiya, Am. Ceram. Soc. Bull. 70, 1633 (1991).
26. I.A. Aksay, D.M. Dabbs, and M. Sarikaya, J. Am Ceram. Soc. 74, 2343 (1991).
27. K. Okada and N. Otsuka, J. Am. Ceram. Soc. 69, 652 (1986).
28. J.A. Pask, X.W. Zhang, A.P. Tomsia, and B.E. Yoldas, J. Am. Ceram. Soc. 70, 704 (1987).
29. J.C. Huling and G.L. Messing, J. Am. Ceram. Soc. 74, 2374 (1991).
30. K.J.D. Mackenzie, J. Am. Ceram. Soc. 55, 68 (1972).
31. M.D. Sacks, N. Bozkurt, and G.W.Scheiffer, J. Am. Ceram. Soc. 74, 2428 (1991).
32. W.E. Cameron, Am. Ceram. Soc. Bull. 56, 1003 (1977).
33. R. Misra and K.J. Rao, Ceram. Int., in press.