Structural evolution during calcination of sol-gel synthesized alumina and alumina-8 vol % zirconia composite

Article abstract of DOI:10.1023/A:1018631324141

Alumina and alumina-8 vol % zirconia (3 mol % yttria) composite powders were synthesized via a sol-gel process using metal chlorides as precursors. Sol-gel synthesized alumina and alumina-zirconia composite powders were initially amorphous and contained structural water. Calcination was necessary to drive off the water and prompt crystallization of the powder. Calcination of synthesized powder was performed at various temperatures from 400 1300°C. Powder morphology and phase evolution during calcination were established using electron microscopy, X-ray diffraction, and thermogravimetric analysis.

Full text of DOI:10.1023/A:1018631324141

JOURNAL OF MATERIALS SCIENCE 32 (1997) 2687 — 2692
Structural evolution during calcination
of sol–gel synthesized alumina and
alumina–8 vol % zirconia composite
J . KU O , D. L. BO U RELL
Center for Materials Science and Engineering, C2201, The University of Texas at Austin,
Austin, TX 78712, USA
Alum ina and alum ina—8 vol % zirconia (3 m ol % yttria) com posite powders were synthesized
via a sol—gel process using m etal chlorides as precursors. Sol—gel synthesized alum ina and
alum ina—zirconia com posite powders were initially am orphous and contained structural
water. Calcination was necessary to drive off the water and prom pt crystallization of the
powder. Calcination of synthesized powder was perform ed at various tem peratures from
4
00—1300 °C. Powder m orphology and phase evolution during calcination were established
using electron m icroscopy, X-ray diffraction, and therm ogravim etric analysis.
1
. Introduction
demonstrated excellent sinterability of the chloride-
synthesized Y O -doped ZrO powder [16]. In the
With increasing demand in the quality of advanced
ceramics, much research is being devoted to produ-
cing nanocrystalline ceramic powder. Nanocrystalline
ceramic powder promises high purity with homogen-
eity at the molecular level and submicrometre grain
size after consolidation. Those advantages may elim-
inate time-consuming milling and mixing steps in con-
ventional ceramic powder preparation, which makes
nanocrystalline ceramic powder very attractive for the
fabrication of high-quality advanced ceramics. A num-
ꢀ
ꢁ
ꢀ
present study, that technique was used again to syn-
thesize alumina and alumina—8 vol % zirconia com-
posite powder. Because the synthesized product was
in the form of hydroxides, subsequent calcination
was required to drive off the structural water and
transform the hydroxides to transition aluminas and,
finally, alpha alumina. This paper presents the micro-
structural development and phase evolution of syn-
thesized powder during calcination at different
temperatures.
ber of nanocrystalline ceramic powders, such as TiO
ꢀ
ꢀ
[
[
1—5], Al O [5—12], ZrO [5, 7, 13—20], and SiO
6], have been produced by various methods. The
ꢀ
ꢁ
ꢀ
sintering temperature of those nanocrystalline pow-
ders is lower than that of conventional powder [7, 9,
2. Experimental procedure
The precursor used for making alumina was AlCl ·
ꢁ
6H O. For alumina—8 vol % zirconia (3 mol % yttria),
ꢀ
1
1, 15—17, 21—23]. This is attributed to the shorter
diffusion length and higher surface curvature of
nanosized powder. Both help to lower the sintering
temperature and minimize coarsening. Submic-
rometre final grain size not only makes the con-
solidated parts stronger [9, 24, 25], but also makes
ceramics candidates for superplastic processing
YCl · 6H O and ZrCl were used as well. The precur-
ꢁ
ꢀ
ꢂ
sors were weighed and dissolved in 0.4 M HCl. While
the solution was stirred with a 37 mm propeller,
NH OH was added until the pH reached 11. The gel
ꢂ
was then washed four times with decreasing amounts
of ammonia in aqueous solution. Finally, the gel was
washed three times with ethanol to remove free water.
The gel was dried in air at 120 °C for 16 h. The powder
thus obtained was ground using an agate mortar.
Calcination of the powder was carried out in air and
in batches at temperatures between 400 and 1300 °C
for 2 h.
Phases present in the powder at different stages and
the crystallite sizes were determined by X-ray diffrac-
tion (XRD) using a Philips X-ray diffractometer,
[26—30], because fine structure is a prerequisite for
superplasticity [31, 32].
Sol—gel synthesis has several advantages over other
methods for producing nanocrystalline ceramic pow-
der. Sol—gel synthesis involves minimal capital invest-
ment, produces an acceptable amount of output, can
be easily scaled up, and can accommodate multicom-
ponent compounds. Winnubst et al. [14] synthesized
nanocrystalline Y O -doped ZrO powder with 8 nm
ꢀ
ꢁ
ꢀ
crystallite size with two different precursors, metal
chlorides and metal alkoxides. They found that the
microstructure of the powder synthesized from chlor-
ides was preferable to the powder synthesized from
alkoxides, because the former did not aggregate. We
previously followed the chloride synthesis route and
Model PW 1729, with CuK radiation. The scanning
a
rate was 2° min\ꢃ. ‘‘JADE’’ software, developed by
Materials Data Inc., was used to analyse the XRD
data.
Microstructural evolution of the powder was ob-
served using transmission electron microscopy (TEM,
0
022—2461 ( 1997 Chapman & Hall
2687

Jeol 200CX) and scanning electron microscopy (SEM,
Jeol JSM-35CX). Powder for TEM observation was
dispersed in acetone and dripped on to a 3 mm copper
grid covered with holey carbon film. The powder for
SEM observation was spread on to a conductive tape
and sputtered with a thin layer of a Au—Pd alloy to
prevent charging.
A-120, AZ-120 showed no crystalline phase at all.
Winnubst et al. [34] synthesized a ZTA (85 wt %
Al O ) powder with a method similar to ours. They
observed crystalline bayerite in their 120 °C-dried
ZTA powder; however, no crystalline bayerite was
observed in our powder.
ꢀ
ꢁ
In AZ-400, no crystalline phase was observed. In
both AZ-550 and AZ-800, gamma alumina was pres-
ent. In AZ-900 and AZ-1000, peaks corresponding to
theta alumina and tetragonal zirconia appeared, while
those from gamma alumina still remained. The
gamma phase finally disappeared in the powder cal-
cined at 1100 °C. In AZ-1200, theta alumina and tet-
ragonal zirconia were the only phases present. In
AZ-1300, alpha alumina replaced theta alumina and
coexisted with tetragonal zirconia. The disappearance
of gamma alumina and the appearance of theta
alumina occurred at a higher temperature in the com-
posite powder than in pure alumina. In our previous
study [16], yttria-stabilized tetragonal zirconia was
well crystallized when calcined at 550 °C, but in the
current study, crystallization did not start until 900 °C
and sharp peaks were not obtained until 1300 °C.
Apparently, the coexistence of alumina and zirconia
inhibits crystallization of both. The implication that
alumina and zirconia were intimately mixed was later
confirmed by microstructural analysis of our sample.
A summary of the phases present in the powders
after calcination is given in Table I. Generally, boeh-
mite and bayerite undergo different transformation
routes to the final phase, alpha alumina [33]. These
routes are:
Thermogravimetric analysis (TGA) was conducted
in bottled air on pure alumina gel immediately after
the gel was dried at 120 °C for 16 h. A Perkin—Elmer
Series 7 thermal analysis system was used. The tem-
perature range was from 50—1000 °C at a heating rate
of 5 °C min\ꢃ.
3
. Results and discussion
3
.1. X-ray diffraction
X-ray diffraction (XRD) patterns of the calcined
alumina powder are shown in Fig. 1 as a function of
calcination temperature. Powder A-120 (gel dried at
1
20°C for 16 h) showed both sharp peaks and broad
peaks. The sharp peaks belong to bayerite (Al(OH) ),
ꢁ
which is well crystallized. The broad peaks were iden-
tified as pseudoboehmite, or poorly crystallized boeh-
mite (c-AlOOH). Pseudoboehmite is a solid whose
XRD pattern shows broad lines that coincide with the
major reflections of well-crystallized boehmite [33].
At 400 °C (A-400; A-alumina, AZ-composite, and
the drying/calcination temperatures are given), both
pseudoboehmite and bayerite were dissociated, all the
peaks disappeared and the powder appeared to be
amorphous. At 550 °C (A-550), gamma alumina
(
and/or eta alumina) peaks appeared and remained
boehmitePgammaPdeltaPthetaPalpha
through 800 °C calcination (A-800). X-ray diffraction
could not clearly distinguish between gamma alumina
and eta alumina; therefore, gamma is used to denote
both phases in the following text. At 900 °C (A-900),
gamma alumina peaks disappeared and those corres-
ponding to theta alumina appeared. Theta alumina
was the only phase present in A-1000 and A-1100. The
final phase, alpha, emerged in A-1200, where it coexis-
ted with theta. Theta alumina disappeared when the
powder was calcined at 1300 °C (A-1300), and alpha
alumina was the only remaining phase in the powder.
Similar XRD patterns of the alumina—8 vol % zir-
conia composite powder are shown in Fig. 2. Unlike
(delta alumina does not appear in the decomposition
sequence of pseudoboehmite [35]), and
bayeritePetaPthetaPalpha
According to Table I, the powder generally followed
those transformation routes. The structure of gamma,
eta, delta, and theta are all deformed spinels with
aluminium cations occupying either tetrahedral or
octahedral sites in an oxygen anion sublattice. Alpha
alumina (corundum) consists of aluminium cations oc-
cupying octahedral sites in a hexagonally close-packed
oxygen sublattice. Therefore, the transformations
Figure 2 Phase evolution of alumina—8 vol % zirconia powder with
increasing calcination temperature; powder was calcined for 2 h in
air after drying for 16 h at 120 °C. (;) Alpha alumina, (᭺ ) tetrag-
onal zirconia.
Figure 1 Phase evolution of pure alumina powder with increasing
calcination temperature; powder was calcined for 2 h in air after
drying for 16 h at 120°C. (;) Alpha alumina.
2
688

TABLE I Phases present and crystalllite size in the alumina and alumina—8 vol % zirconia powder after calcination at various temperatures
Alumina
Alumina—zirconia composite
Alumina
Zirconia
¹
(°C)
Phase!
Size(nm)
Phase!
Size(nm)
Phase!
Size(nm)
1
20 (drying)
Ba#P
—
—
5
5
13
14
16
27
35
—
—
—
—
4
6
—
—
—
—
T
T
T
T
T
—
—
—
—
6
8
17
21
33
450
550
800
900
—
"
"
"
"
h
h
h
h, a
a
h,"
h,"
h
h
a
6
1
1
1
1
000
100
200
300
13
20
32
47
!
Ba, Bayerite; P, Pseudoboehmite; h, theta alumina; a, alpha alumina; T, tetragonal zirconia; —, amorphous state/size not available.
c and/or g alumina.
"
among transition aluminas are of the reordering type,
while the transformation between theta alumina and
alpha alumina is of the massive type.
In the present study, the crystallite size calcula-
ted from X-ray peak broadening remained almost
unchanged after calcination below 800 °C. As the
calcination temperature was increased above 800 °C,
crystallites coarsened, and they did so dramatically
above 1300 °C. Wefers and Misra [33] reported signif-
icant sintering and three-dimensional fusion of
transition alumina above 827 °C. Their explanation
for enhanced sintering and increased cation mobility
in association with conversion to a-alumina, is cor-
roborated by our observation of particle coarsening.
Figure 3 TGA result for alumina powder after drying for 16 h at
20°C. The initial powder weight was 46.486 g.
1
3
.2. Therm al analysis
The TGA result for the alumina powder is shown in
Fig. 3. The sample weight decreased abruptly between
tral region of Fig. 4a). Those particles have a mottled
appearance which is consistent with a pore network
described by Wefers and Misera [33]. Such porosity is
associated with the decomposition of aluminium hy-
droxides during calcination. The pores are described
in transition alumina as an intersecting pore system
with 1 nm size parallel pores located in the cleavage
plane (0 0 1). The parallel pores separate the solid into
lamellae that are of the order of 2—3 nm thick, while
a network of irregularly shaped slits extends perpen-
dicular to the pores parallel to (0 0 1). The result is an
intersecting pore system that divides the lamellae into
interconnected, irregularly shaped domains of solid.
The selected-area diffraction (SAD) pattern yielded
spots indicating the existence of a crystalline structure.
This spot pattern was indexed as eta alumina. How-
ever, when a micro-electron beam was used to exam-
ine the powder, the diffraction pattern obtained was
that of bayerite single crystals, as shown in the insert
of Fig. 4a, and not of eta alumina. Because eta
alumina is the first transition alumina into which
bayerite transforms when it is heated, we believe the
heat generated by electron-beam bombardment is re-
sponsible for the transformation from bayerite to eta
alumina. Also apparent in Fig. 4a is an agglomerated
cluster of particles approximately 4 nm in size. The
diffraction pattern from this region consisted only of
2
9
10 and 250 °C, and weight loss continued up to
00°C. The total weight loss after heating to 1000 °C
was 25%. XRD analysis showed pseudoboehmite and
bayerite were the two phases present in the 120 °C
treated powder. The weight loss of boehmite and
bayerite, upon dehydration, was 15% and 35%, re-
spectively [36]. Therefore, the weight percentage of
pseudoboehmite and bayerite in the dried powder
might be estimated to be approximately 50% of each.
Pseudoboehmite converts to crystalline bayerite, the
phase thermodynamically stable in air, when aged in
an aqueous solution at temperatures under 77 °C [33].
In our case, it is not known whether both bayerite and
pseudoboehmite were the hydroxides that formed dur-
ing synthesis, or whether pseudoboehmite was first
formed and then partially converted to bayerite.
3
.3. Electron m icroscopy observation
Microstructures of the pure, calcined nanocrystalline
alumina powders, A-120, A-550, A-900, A-1100, and
A-1300, were observed using TEM and are shown in
Fig. 4a—e respectively. The microstructure of A-120
(Fig. 4a) consists of two distinctively different features.
One is faceted particles of nominal size 250 nm (cen-
2
689

Figure 4 Transmission electron micrographs of pure alumina pow-
der after drying at (a) 120 °C for 16 h (insert showing the SAD of
bayerite), and calcination at (b) 550 °C, (c) 900 °C, (d) 1100 °C, and
(e) 1300°C for 2 h.
agrees with Wilson’s [37] report that the transforma-
tion of transition aluminas is a nucleation-and-growth
process. Crystallites in A-1300 have grown to approx-
imately 40 nm, an indication of significant fusion.
Fig. 5a—d shows the microstructure of the calcined
alumina—zirconia composite, AZ-550, AZ-900, AZ-
1
100, and AZ-1300, respectively. The composite pow-
der generally shows the same characteristics of pure
alumina powder. However, some of the crystallites in
AZ-900, AZ-1100, and AZ-1300 appear darker than
pure alumina powder. The darker crystallites are
most likely zirconia, because zirconia would
appear darker due to mass contrast. The zirconia
crystallites, as shown in AZ-1300, are smaller than the
alumina crystallites. This result agrees with the size
measured from X-ray peak broadening. Our earlier
hypothesis that zirconia crystallites form an intimate
mixture with alumina crystallite is also confirmed by
AZ-1300.
faint rings, consistent with a nanocrystalline assem-
blage. We believe that this cluster is the pseudoboeh-
mite phase whose presence was indicated by XRD.
Fig. 4b shows the microstructure of A-550. Traces
of Moir e´ fringes can barely be seen in pseudoboeh-
mite. The pores in bayerite have slightly enlarged. No
other changes were observed in A-550. Fig. 4c, d and
e show the microstructure of A-900, A-1100, and A-
1
300, respectively. The pores observed in hydroxides
Fig. 6 is a scanning electron micrograph of AZ-550.
Almost all the powder is agglomerated to various sizes
from less than 1 lm to more than 10 lm. No unag-
glomerated powder was observed in the micrograph.
Aggregates or agglomerates can be disadvantageous if
they are difficult to break up during compaction of the
powder. The large voids between the agglomerates
will prolong the sintering time and the final grain size
enlarged progressively with increasing calcination
temperature. Individual crystallites are visible in
Fig. 4c—e, although they have retained the ‘‘intercon-
nected’’ feature from the hydroxides and are heavily
aggregated. In A-900, typical crystallite size is 15 nm
but with some variation, and the aggregate size is
about 100 nm. The inhomogeneity of crystallite size
2
690

Figure 5 Transmission electron micrographs of alumina—8 vol % zirconia powder after calcination for 2 h at (a) 550 °C, (b) 900 °C, (c) 1100 °C,
and (d) 1300 °C.
ides were calcined at different temperatures from
4
00—1300 °C. With increasing calcination temper-
ature, the hydroxides dehydrated and transformed to
transition alumina, eventually reaching a final phase
alpha alumina. In the composite powder, tetragonal
zirconia was the final phase. The coexistence of zirco-
nia and alumina in the composite did not affect the
transformation route of the hydroxides. However, the
temperature at which the transformation was com-
pleted was affected by the presence of zirconia. Sim-
ilarly, the crystallization of zirconia in alumina was
delayed to 900 °C, rather than 550 °C where it has
previously been observed to crystallize.
TEM observation of the powder showed that ag-
glomerated clusters of pseudoboehmite and bayerite
single crystals both are porous. Pore size increased as
the calcination temperature increased. The crystallite
size also increased with calcination temperature, al-
though the precise crystallite size may not be assessed
directly from transmission electron micrographs due
to particle clustering. The crystallite size increased
dramatically after conversion to alpha alumina was
complete.
Figure 6 Scanning electron micrograph of alumina—8 vol % zirco-
nia powder after calcination at 550 °C for 2 h, showing the variation
of agglomerate size.
will be larger than that of unagglomerated powder.
The major benefit of nanosized powder is thus lost.
4
. Conclusion
Sol—gel synthesized zirconia particles using metal
chlorides as the precursors, are spherical, and each
particle is discrete [14, 16]. However, our alumina and
alumina—zirconia composite powders had an irregular
shape and formed aggregates, although the crystallite
size was still within the nanometre range.
Al O and Al O —8 vol % ZrO were synthesized via
ꢀ
ꢁ
ꢀ ꢁ
ꢀ
a sol—gel method. Bayerite crystals and pseudoboeh-
mite were produced. The proportion of each hydrox-
ide in the powder was estimated from weight loss
analysis to be approximately 50% each. The hydrox-
2
691

References
20. D. HUANG, K . R . V E N K A T A C H A R I and G. C.
STANGLE, ibid. 10 (1995) 762.
1
.
H. HAHN, J . L O G A S and R. S. AVERBACK, J. Mater. Res.
(1990) 609.
2
1. M. M. R. BOUTZ, L. WINNUBST, A. J. BURGGRAAF, M.
NAUER and C. CARRY, J. Am. Ceram. Soc. 78 (1995) 121.
2. M. KUMAGAI and G. L. MESSING, ibid. 67 (1984) C-230.
3. Y. J. HE, A. J. A. WINNUBST, H. VERWEIJ and A. J.
BURGGRAAF, J. Mater. Sci. 29 (1994) 6505.
5
2
3
.
.
J. A. EASTMAN, J. Appl. Phys. 75 (1994) 770.
2
2
E. HARO-PONIATOWSKI, R. RODRIGUEZ-TALAV-
ERA, M. DELA CRUZ HEREDIA, O. CANO-CORONA
and R. ARROYO-MURILLO, J. Mater. Res. 9 (1994) 2102.
F. W. YAN and W. W. RHODES, Mater. Sci. Eng. 61 (1983)
2
4. H. HAHN and R. S. AVERBACK, J. Nanostr. Mater. 1 (1992)
4
.
.
9
5.
5
9.
D. VOLLATH and K. E. SICKAFUS, J. Nanostr. Mater.
(1992) 427.
®
B. GUNTHER and A. KUMPMANN, ibid. 1 (1992) 27.
2
5. M. YOSHINAKA, K. HIROTA, O. YAMAGUCHI, H.
KUME, S. INAMURA, H. MIYAMOTO, N. SHIOKAWA
and R. SHIKATA, Br. Ceram. ¹rans. 93 (1994) 234.
5
1
6
7
8
.
.
.
2
2
2
2
3
3
6. K. R. VENKATACHARI and R. RAJ, J. Am. Ceram. Soc. 69
R. CHAIM, ibid. 1 (1992) 479.
J. A. EASTMAN, L . J . T H O M P S O N and D. J. MAR-
SHALL, ibid. 2 (1993) 377.
(
1986) 135.
7. T. G. NIEH and J. WADSWORTH, Acta Metall. Mater. 39
1991) 3037.
8. D. M. OWEN and A. H. CHOKSHI, J. Nanostr. Mater.
(1993) 181.
9. I-WEI CHEN and L. A. XUE, J. Am. Ceram. Soc. 73 (1990)
585.
(
9
.
S. RAJENDRAN, J. Mater. Sci. 29 (1994) 5664.
1
0. Y. MIZUSHIMA and M. HORI, J. Non-Cryst. Solids 167
1994) 1.
2
(
1
1. G. P. JOHNSTON, R. MUENCHAUSEN, D. M. SMITH,
2
W. FAHRENHOLTZ and S. FOLTYN, J. Am. Ceram. Soc. 75
0. K. KAJIHARA, Y . Y O S H I Z A W A and T. SAKUMA, Acta
Metall. Mater. 43 (1995) 1235.
1. Y. MAEHARA and T. G. LANGDON, J. Mater. Sci. 25 (1990)
(
1992) 3293.
2. S. RAJENDRAN and J. L. WOOLFREY, Key Eng. Mater.
3–55 (1991) 462.
3. T. OKUBO and H. NAGAMOTO, J. Mater. Sci. 30 (1995)
49.
4. A. J. A. WINNUBST, W. F. M. GROOT ZEVERT,
G. S. A. M. THEUNISSEN and A. J. BURGGRAAF, Mater.
Sci. Eng. A109 (1989) 215.
1
1
1
5
2
275.
3
2. T. G. LANGDON, JOM 42 (1990) 8.
7
3
3. K. WEFERS and C. MISERA, ‘‘Oxides and Hydroxides of
Aluminum’’, ALCOA Technical Paper No. 19, Revised (Alu-
minum Company of America, ALCOA Center, PA, 1987).
4. P. D. EXTER, L. WINNUBST, T. H. P. LEUWERINK and
A. J. BURGGRAAF, J. Am. Ceram. Soc. 77 (1994) 2376.
5. B. C. LIPPENS and J. J. STEGGERDA, in ‘‘Physical and
Chemical Aspects of Adsorbents and Catalysts’’, edited by
B. G. Linsen (Academic Press, London, 1970) pp. 171—209.
6. B. E. YOLDAS, J. Appl. Chem. Biotechnol. 23 (1973) 803.
7. S. J. WILSON and J. D. C. McCONNELL, J. Solid State
Chem. 34 (1980) 315.
3
3
1
1
1
1
5. D. L. BOURELL, P A R I M A L and W. KAYSSER, J. Am.
Ceram. Soc. 76 (1993) 705.
6. G. B. PRABHU and D. L. BOURELL, J. Nanostr. Mater.
6
(Nano ’94 Proceedings, Part I) (1995) 361.
7. PARIMAL, Master Thesis, University of Texas at Austin
1992).
3
3
(
8. K. R. VENKATACHARI, D. HUNG, S. P. OSTRANDER,
W. A. SCHULZE and G. C. STANGLE, J. Mater. Res. 10
(1995) 748.
1
9. K. R. VENKATACHARI, D. HUNG, S. P. OSTRANDER,
Received 19 January
W. A. SCHULZE and G. C. STANGLE, ibid. 10 (1995) 756.
and accepted 17 September 1996
.
2
692