The influence of thermal treatment on the phase development in HfO 2-Al2O3 and ZrO2-Al 2O3 systems

Article abstract of DOI:10.1016/j.jallcom.2004.07.004

Amorphous precursors of HfO₂-AlO_1.5 and ZrO ₂-AlO_1.5 systems covering the whole concentration range were co-precipitated from aqueous solutions of the corresponding salts. The thermal behaviour of the amorphous

Full text of DOI:10.1016/j.jallcom.2004.07.004

Journal of Alloys and Compounds 388 (2005) 126–137
The influence of thermal treatment on the phase development
in HfO2–Al2O3 and ZrO2–Al2O3 systems
∗
ˇ
´
´
G. Stefanic , S. Music, R. Trojko
−
Ruder Bosˇkovic´ Institute, P.O. Box 180, HR-10002 Zagreb, Croatia
Received 10 May 2004; received in revised form 13 July 2004; accepted 13 July 2004
Abstract
Amorphous precursors of HfO2–AlO1.5 and ZrO2–AlO1.5 systems covering the whole concentration range were co-precipitated from
aqueous solutions of the corresponding salts. The thermal behaviour of the amorphous precursors was examined by differential thermal
analysis, X-ray powder diffraction (XRD), laser Raman spectroscopy and scanning electron microscopy. The crystallization temperature of
both systems increased with increase in the AlO1.5 content, from 530 to 940 ◦C in the HfO2–AlO1.5 system, and from 405 to 915 ◦C in the
ZrO2–AlO1.5 system. The results of phase analysis indicate an extended capability for the incorporation of Al3+ ions in the metastable HfO2-
and ZrO2-type solid solutions obtained after crystallization of amorphous co-gels. Precise determination of lattice parameters, performed using
whole-powder-pattern decomposition method, showed that the axial ratio cf /af in the ZrO2- and HfO2-type solid solutions with 10 mol% or
more of Al3+ approach 1. The tetragonal symmetry of these samples, as determined by laser Raman spectroscopy, was attributed to the
displacement of the oxygen sublattice from the ideal fluorite positions. It was found that the lattice parameters of the ZrO2-type solid solutions
decreased with increasing Al3+ content up to ∼10 mol%, whereas above 10 mol%, further increase of the Al3+ content has very small influence
on the unit-cell volume of both HfO2- and ZrO2-type solid solutions. The reason for such behaviour was discussed. The solubility of Hf4+
and Zr4+ ions in the aluminium oxides lattice appeared to be negligible.
© 2004 Elsevier B.V. All rights reserved.
Keywords: ZrO2–AlO1.5; HfO2–AlO1.5; XRD; Raman spectroscopy; DTA; SEM/EDX
1. Introduction
[1,2] or as the crystallization product of hydrous zirconia [3].
In the case of hafnia, the formation of metastable t- or c-HfO2
(ratio c/a is nearly 1) at RT is much harder to obtain [4–8]. Cu-
bic polymorphs of HfO2 and ZrO2 can be stabilized at RT by
the incorporation of aliovalent oversized metal cations, viz.
Ca2+, Sc3+, Y3+, etc. [9–17]. Stabilization of these high-
temperature polymorphs was attributed to the decrease in the
Hf and Zr coordination number caused by the introduction of
oxygen vacancies [2,9,17].
The incorporation of aliovalent undersized dopant
cations, such as Al3+, Fe3+ or Cr3+, also introduces oxygen
vacancies into the HfO2 and ZrO2 lattice. However, in the
thermodynamically stable systems the amount of the incorpo-
rated dopants is too small to stabilize their high-temperature
polymorphs [17–20]. The results of phase analysis of the
ZrO2–CrO1.5 system showed that, regardless of the negli-
gible solubility of the Cr3+ ions, partial stabilization of a
Pure hafnia and zirconia exhibit similar physical and
chemical properties, which made them very interesting high-
temperature materials. In dependence on the temperature,
HfO2 and ZrO2 appear as monoclinic (m-HfO2, m-ZrO2),
tetragonal (t-HfO2, t-ZrO2)and cubic (c-HfO2, c-ZrO2) poly-
morphs of which only the monoclinic is thermodynamically
stable at RT and standard pressure. High temperature poly-
morphs of ZrO2 become stable at above 1170 ◦C (t-ZrO2) and
2300 ◦C (c-ZrO2), whereas the stabilization of t- and c-HfO2
require significantly higher temperatures, 1720 and 2600 ◦C,
respectively. Metastable t-ZrO2 can be easily obtained at RT
as a product of the thermal decomposition of zirconium salts
∗
Corresponding author. Tel.: +385-1-456-1111; fax: +385-1-46-80098.
ˇ
E-mail address: stefanic@rudjer.irb.hr (G. Stefanic´).
0925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2004.07.004

ˇ
G. Stefanic´ et al. / Journal of Alloys and Compounds 388 (2005) 126–137
127
metastable t-ZrO2 occurred [20]. The observed stabilization
was attributed to surface interaction between Cr2O3 and
ZrO2. Thisinteractionpreventedthediffusionofoxygenfrom
the atmosphere into the ZrO2 lattice, which otherwise trig-
gered the transformation t-ZrO2 → m-ZrO2 on cooling [20].
The solubility of undersized dopant cations in the
metastable HfO2- and ZrO2-type solid solutions obtained af-
ter crystallization of amorphous precursors is significantly
higher. High temperature cubic and tetragonal polymorphs
of HfO2 and ZrO2 could appear in such metastable solid
solutions. However, such stabilization is not very efficient
[7,8,21–24]. In our previous investigations, we examined
the thermal behaviour of the amorphous precursors to the
HfO2–FeO1.5 and ZrO2–FeO1.5 systems [7,8,24]. It was
found that the incorporation of Fe3+ ions has both a sta-
bilizing and destabilizing influence on the high-temperature
polymorphs of HfO2 and ZrO2. The destabilizing influence
significantly increased if the thermal treatments were per-
formed at small air pressure (∼4 × 10−3 Pa), which was at-
tributed to the association of the introduced oxygen vacancies
with smaller Fe3+cations [7,8].
lected temperatures up to 1300 ◦C. The phase compositions
of the obtained crystallization products were analysed at RT
using X-ray powder diffraction (XRD). XRD patterns were
recorded using an automated Philips diffractometer (model
MPD 1880) with monochromatized Cu K␣ radiation. Silicon,
¯
␣-Si, was used as int◦ernal standard (space group Fd3m, unit-
˚
cell parameter at 25 C: a = 5.43088 A, [36]). XRD patterns
were scanned in steps of 0.03◦ (2θ), in the 2θ range from 20◦
to 110◦, with a fixed counting time (5 or 10 s), for the purpose
of the accurate determination of the unit-cell parameters of
the studied samples by means of the powder-pattern-fitting
methods.
In case where monoclinic and tetragonal polymorphs of
hafnia or zirconia co-exist their volume fractions (νt and νm)
were estimated from the integral intensities of the monoclinic
¯
diffraction lines (1 1 1) and (1 1 1) and tetragonal diffraction
line (1 0 1), following a procedure proposed by Toraya et al.
[37]. The volume fractions are given by following equations:
¯
Im(1 1 1) + Im(1 1 1)
x =
,
(1)
(2)
(3)
¯
Im(1 1 1) + Im(1 1 1) + It(1 0 1)
The capability of Al3+ ions to stabilize the cubic poly-
morph of zirconia in the metastable solid solutions is still a
matter of discussion [25–35]. Several authors reported that
the amorphous precursors to the ZrO2–Al2O3 system crys-
tallized after heating between 500 and 900 ◦C to c-ZrO2-type
solid solution containing up to 40 mol% Al2O3 [25–29]. At
higher temperature c-ZrO2 converted to t-ZrO2 and finally to
m-ZrO2 that was followed by a decrease in Al2O3 solubility.
On the other hand, other reports [31–35] stated that the in-
corporation of Al3+ ions into the ZrO2 lattice stabilized only
the tetragonal polymorph of zirconia.
νt = 1 − νm,
1.311x
νm
=
,
1 + 0.311x
Integrated intensities of the diffraction lines were deter-
mined using the individual profile-fitting method (computer
program PRO-FIT) [38].
Laser Raman (DILOR Z24) spectroscopy was also used
in the cases when the presence of the cubic or tetragonal
polymorphs of hafnia and zirconia could not be clearly dis-
tinguished by XRD.
Differently from the ZrO2–AlO1.5 system, the stabiliz-
ing influence of Al3+ ions in the HfO2–AlO1.5 system was
much less investigated. In the present work, we focused on
the thermal behaviour of the amorphous precursors to the
HfO2–AlO1.5 and ZrO2–AlO1.5 systems, covering the whole
concentration region. The capability of Al3+ ions to incorpo-
rate into the HfO2 and ZrO2 lattice and stabilize their high-
temperature polymorphs was discussed.
The first crystallization products of some samples at
the ZrO2-rich side of the concentration range were also
analysed with a scanning electron microscope (JSM-5800
JEOL) equipped with an energy dispersive X-ray spectrom-
eter (LINK ISIS 300 Oxford Instruments). Quantitative bulk
analysis was performed using SEMQuant program.
The thermal behaviour of the amorphous precursors from
the HA and ZA systems was also examined by differential
thermal analysis (DTA). The prepared samples were heated
in the air atmosphere up to 950 ◦C with a rate of 10 ◦C min−1
using the standard instrumentation.
,
2. Experimental
Amorphous precursors of the HfO2–AlO1.5 (HA) and
ZrO2–AlO1.5 (ZA) systems covering the whole concentra-
tion range were co-precipitated from aqueous solutions of
HfOCl2·8H2O and AlCl3·6H2O salts and aqueous solutions
of the ZrO(NO3)2·2H2O and Al(NO3)3·9H2O salts, respec-
tively. The co-precipitations were performed by the addition
of 25% NH3·aq. up to pH ∼7.5. All the chemicals used were
of analytical grade. The solid phase was separated from the
liquid using an ultra-speed centrifuge (operational speed up
to 20,000 rpm), then washed repeatedly in doubly distilled
water and dried at 70 ◦C for 24 h. The obtained solid phases
were ground in an agate mortar and calcined for 3 h at the se-
3. Results and discussion
3.1. Thermal behaviour
Fig. 1 shows DTA curves of the samples from the HA and
ZA systems with AlO1.5 content between 0 and 60 mol%.
All curves are characterized by endothermic peaks resulting
from dehydration and exothermic peaks resulting from crys-
tallization. The position of the exothermic peak shifted to the
higher temperature with increasing Al3+ content. The crys-
tallization temperature increased from 530 to 940 ◦C for the

ˇ
G. Stefanic´ et al. / Journal of Alloys and Compounds 388 (2005) 126–137
128
Fig. 1. DTA curves of the samples from HfO2–AlO1.5 and ZrO2–AlO1.5 systems. Letters a, b, c, d, e, f and g stand for 0, 10, 20, 30, 40, 50 and 60 mol% of
AlO1.5, respectively.
HA system and from 405 to 915 ◦C for ZA system. Similar
results were observed for the HfO2–FeO1.5 and ZrO2–FeO1.5
systems [8], but in these cases, the rate of the increase was
significantly smaller (Fig. 2). The observed increase indi-
cated that amorphous precursors are single co-gels, which
crystallized into metastable crystalline phases with extended
capability for the formation of the solid solutions.
treatment at 1000 ◦C, only traces of the metastable phase are
still present (Fig. 3). The stability of the high-temperature
polymorph of zirconia with regard to the influence of tem-
perature increased with increasing of aluminium content and
showed to be significantly bigger compared to the stabil-
ity of corresponding high-temperature polymorphs of hafnia
(Fig. 4). In some cases, it was hard to distinguished between
the tetragonal and cubic polymorph of zirconia and almost
impossible to distinguished between the tetragonal and cubic
polymorph of hafnia, for which the cf/af ratios approach 1
(af and cf are related to fluorite-type lattice), on the basis of
XRD analysis alone. In a series of papers, Yoshimura et al.
[39–42] proved the existence of the tetragonal polymorph of
hafnia with an axial ratio cf/af = 1, having a displacement of
the oxygen sublattice from the ideal fluorite positions. Due to
the significantly smaller atomic scattering factor of oxygen
compared to hafnium, X-ray diffraction is not very sensitive
to structural changes caused by oxygen displacement. Laser
Raman spectroscopy showed to be the most powerful tech-
nique in such cases [2,3,8,39–42]. Group theory allows six
Raman active modes of vibration (A1g + 2B1g + 3Eg) for the
tetragonal polymorph of zirconia and hafnia and only one
active modes of vibration for their cubic polymorph (F2g).
All those six Raman active modes of vibration are present in
the laser Raman spectra of the HA crystallization products
with 20, 30 and 40 mol% of Al3+ ions and the spectra of
the ZA crystallization products with 10, 40 and 50 mol% of
Al3+ ions. The spectrum of the HA crystallization product
with 40 mol% of Al3+ ions contained also small bands typ-
3.2. Phase analysis
Phase development of the amorphous precursors of the
HfO2–AlO1.5 and ZrO2–AlO1.5 systems after calcinations
(3 h) and cooling in the presence of air at atmospheric pres-
sure are given in Table 1 . Amorphous precursors to the pure
hafnia crystallized to m-HfO2. The incorporation of Al3+
ions caused the appearance of the high-temperature tetrag-
onal polymorph, which in the crystallization product of the
samples with 20 and 30 mol% became the only phase present.
However, the stabilization is not very efficient. Additional
temperature treatment caused transition to the monoclinic
polymorph and the appearance of aluminium oxide-type solid
solutions (Table 1).
Differently form hafnia, the crystallization product ob-
tained after calcinations of the amorphous precursor to pure
zirconia contained t-ZrO2 as the dominant phase with only
traces of the thermodynamically stable m-ZrO2 phase. Such
a high content of metastable tetragonal polymorph can be at-
tributed to the influence of grinding [3]. Further temperature
treatment caused t-ZrO2 → m-ZrO2 transition, so that after

ˇ
G. Stefanic´ et al. / Journal of Alloys and Compounds 388 (2005) 126–137
129
Table 1
The initial molar compositions and the results of phase analysis obtained after the calcination (3 h) and cooling of the amorphous precursors of the HfO2–AlO1.5
and ZrO2–AlO1.5 systems in the presence of air at atmospheric pressure
Molar fraction of AlO1.5
T (◦C)
Phase composition (relative volume fraction)
HfO2–AlO1.5 system
ZrO2–AlO1.5 system
–
400
600
800
1000
1300
Amorphous
m-HfO2
m-HfO2
m-HfO2
m-HfO2
t-ZrO2 (0.98) + m-ZrO2 (0.02)
m-ZrO2 (0.63) + t-ZrO2 (0.37)
m-ZrO2 (0.85) + t-ZrO2 (0.15)
m-ZrO2 (0.97) + t-ZrO2 (0.03)
m-ZrO2
0.05
400
500
Amorphous
Hm (0.87) + Ht (0.13)
Amorphous
Zt
600
Hm (0.92) + Ht (0.08)
Zt
800
Hm
Hm
Hm
Hm + A␣
Zt (0.84) + Zm (0.16)
Zm (0.82) + Zt (0.18)
Zm (0.98) + Zt (0.02)
Zm + A␣
1000
1100
1300
0.10
0.20
400
600
800
1000
1100
1300
Amorphous
Ht
Hm (0.70) + Ht (0.30)
Hm (0.99) + Ht (0.01)
Hm
Amorphous
Zt
Zt (0.96) + Zm (0.04)
Zt (0.57) + Zm (0.43)
Zm (0.94) + Zt (0.06)
Zm + A␣
Hm + A␣
500
600
700
Amorphous
Amorphous
Amorphous + Ht
Ht
Amorphous
Amorphous + Zt
Zt
Zt
800
1000
1100
1300
Hm (0.98) + Ht (0.02)
Hm + A␣
Hm + A␣
Zt
Zm (0.92) + Zt (0.08) + A␦ + A␣
Zm + A␣ + Zt
0.30
0.40
700
800
1000
1100
1300
Amorphous
Ht
Hm (0.98) + Ht (0.02)
Hm + A␣
–
Zt
Zt
Zm (0.85) + Z t(0.15) + A␦ + A␣
Zm (0.98) + Zt (0.02) + A␣
Hm + A␣ + Ht
700
800
900
–
Amorphous
Zt
–
Amorphous + Ht
Ht + Hm
1000
1100
1300
Hm + Ht + A␦
Hm + Ht + A␣ + A␦
Hm + A␣ + Ht
Zt
Zt + Zm + A␦ + A␣
Zm + Zt + A␣
0.50
0.60
700
800
900
1000
1100
1300
–
Amorphous
Zt
–
Zt + A␦
Zt + A␦ + A␣
Zm + Zt + A␣
Amorphous + Ht
Ht + Hm + A␦
Hm + Ht + A␦ + A␣
Hm + Ht + A␣ + A␦
Hm + Ht + A␣ + A␦
800
900
Amorphous + Ht
–
Amorphous + Zt
Zt + A␦
1000
1100
1300
Ht + Hm + A␦
Hm + Ht + A␦ + A␣
A␣ + Hm
Zt + A␦
Zt + A␦ + A␣
A␣ + Zm + Zt
0.70
0.80
800
1000
1100
1300
Amorphous
Amorphous + Zt
A␦ + Zt
A␦ + A␣ + Zt
A␣ + Zm + Zt
Ht + Hm + A␦
A␦ + A␣ + Hm + Ht
A␣ + Hm + Ht
800
900
Amorphous
–
Amorphous + Zt
A␥ + Zt
1000
1100
1300
Ht + Hm + A␦
A␦ + A␣ + Hm
A␣ + Hm
A␦ + Zt
A␣ + Aa + Zt
A␣ + Zm + Zt

ˇ
G. Stefanic´ et al. / Journal of Alloys and Compounds 388 (2005) 126–137
130
Table 1 (Continued )
Molar fraction of AlO1.5
T (◦C)
Phase composition (relative volume fraction)
HfO2–AlO1.5 system
ZrO2–AlO1.5 system
0.90
0.95
–
800
1000
1100
1300
–
Amorphous
Amorphous + A␥ or A␦
A␦ + Zt
A␦ + A␣ + Zt + Zm
A␣ + Zt + Zm
Amorphous
A␦ + A␣ + Ht + Hm
A␦ + A␣ + Hm + Ht
A␣ + A␦ + Hm + Ht
–
400
600
1000
1100
1300
Amorphous
–
–
Amorphous + A␦ + Hm + Ht
Zt + A
Zt + A
A␥ + A␦
Amorphous + A␥
Amorphous + A␥
Amorphous + A␦ + Zt
A␣ + A␦ + Zt + Zm
A␣ + Zm + Zt
1.00
–
600
Amorphous
–
AB
A␥
800
A␥
A␦
1000
1100
1300
A␦ + A␣
A␣ + A␦
A␣
A␦ + A␣
A␣ + A␦
A␦
Description: The symbols Zm, Zt, Hm, Ht, AB, AG, A␣, A␥, A␦ stand for the phases structurally similar to m-ZrO2, t-ZrO2, m-HfO2, t-HfO2, beohmite, gibbsite,
␣-Al2O3, ␥-Al2O3 and ␦-Al2O3, respectively.
ical of monoclinic polymorph (Fig. 5). These results show
that regardless to the amount of Al3+ ions only tetragonal
polymorph of HfO2 and ZrO2 could be stabilized.
On the Al-rich side of the concentration range, besides
the ZrO2 and HfO2-type solid solutions, the crystallization
products contained aluminium oxide-type solid solutions.
The solubility of Zr4+ or Hf4+ ions in the lattice of the
aluminium oxides appeared to be very small (less than
5 mol%). Due to that, phase developments during the
temperature treatment of aluminium oxide-type solid
solutions are not very different from the phase development
of pure aluminium oxide. However, the phase transitions on
calcinations occurred at somewhat higher temperatures. The
starting material obtained after precipitation from the aque-
ous solution of Al(NO3)3 was characterized as ␥-AlOOH,
which, on calcination, underwent phase transitions, first to
␥-Al2O3, then to ␦-Al2O3 and finally to ␣-Al2O3. The amor-
phous phase, obtained after precipitation from an aqueous
solution of AlCl3, crystallized on calcination to ␥-Al2O3,
which underwent phase transitions to ␦-Al2O3 and finally to
␣-Al2O3.
3.3. Lattice parameters
Precise determinations of the unit-cell parameters of the
HfO2- and ZrO2-type solid solutions were performed us-
ing the whole-powder-pattern decomposition method (WPPF
program, [38]), following the procedure described previously
[8]. The fitting was performed in the scanned 2θ range, from
20◦ to110◦, usingthesplit-typepseudo-Voigtprofilefunction
and the polynomial background model. Figs. 6 and 7 show the
observed and calculated powder patterns of the crystalliza-
Fig. 2. The influence of the molar fraction of AlO1.5 and FeO1.5 on the
crystallization temperatures of the amorphous precursors of HfO2–AlO1.5
tion products from the HA and ZA series, respectively. Al-
though laser Raman spectroscopy indicated that, regardless
(full square), HfO2–FeO1.5 (empty square), ZrO2–AlO1.5 (full triangle) and
ZrO2–FeO1.5 (empty triangle) systems.

ˇ
G. Stefanic´ et al. / Journal of Alloys and Compounds 388 (2005) 126–137
131
Fig. 3. Individual profile-fitting results of the crystallization products from ZrO2–AlO1.5 system with 0 mol% of AlO1.5 (a), 5 mol% of AlO1.5 (b) and 10 mol%
of AlO1.5 (c). Observed data are represented by squares, and the calculated one by full line. Long and short vertical bars marked the Cu K␣1 and Cu K␣2
positions of the reflection, respectively. Differences between the observed and calculated intensities are plotted in the same scale below.
Fig. 4. Volume fraction of tetragonal polymorphs of HfO2 (full line) and ZrO2 (broken line) in the product obtained after 3 h of calcinations of the amorphous
precursors of the HfO2–AlO1.5 and ZrO2–AlO1.5 systems at the selected temperatures.

ˇ
132
G. Stefanic´ et al. / Journal of Alloys and Compounds 388 (2005) 126–137
Fig. 5. Laser Raman spectra of the crystallization products from the HfO2–AlO1.5 system (left) and ZrO2–AlO1.5 system (right).
the amount of the incorporated Al3+ ions, only the tetrago-
(Hf4+ or Zr4+), and m is the concentration of the dopant
cation (mol%) in the form of RO1.5. These empirical relations
were developed from the lattice parameters obtained from the
HfO2- and ZrO2-type solid solutions with divalent or triva-
lent oversized dopant cations. The results presented here, as
well as our previous results obtained for the HfO2–FeO1.5
and ZrO2–FeO1.5 systems [8], showed that these equations
could not be used in cases of HfO2 and ZrO2 solid solutions
nal polymorph of hafnia and zirconia can be stabilized, the
results of WPPF showed that the axial ratios cf/af in the crys-
tallization products with 10 or more mol% of Al3+ approach
1 (Table 3). The tetragonal symmetry of these samples can be
explained by the displacement of oxygen sublattice [39–42].
In case of the system ZA, the obtained results indicate a de-
crease of the lattice parameters with increasing Al3+ content
up to ∼10 mol%. An additional increase of the Al3+ content
has a very small influence on the lattice parameters. A similar
result was obtained for the HA system with 10–40 mol% of
Al3+ (Table 3).
In accordance with Vegard’s law Kim et al. [10,11] pro-
posed empirical equations that relate the lattice constants
of the fluorite-type HfO2 and ZrO2 solid solutions with the
type (ionic radius, valency) and concentration of the dopant
cations. For the HfO2–RO1.5 and ZrO2–RO1.5 systems, these
equations can be expressed as follows:
with dopant cations significantly smaller than Hf4+ (0.83 A,
˚
[43]) or Zr4+ (0.84 A, [43]) ions. The incorporation of Fe
3+
˚
˚
ions (0.65 A, [43]) caused a linear decrease of the lattice pa-
rameters of HfO2- and ZrO2-type solid solution, but the rate
of the decrease is about two times smaller than it should be
according to the relations above (Fig. 8). In the case of the
ZrO2–FeO1.5 system, a smaller amount of Fe3+ ions stabi-
lized the tetragonal polymorph. However, the unit-cell vol-
ume of these samples showed to be on the same line, which
indicates that the transition from a tetragonal to a cubic poly-
morph has a very small influence on the unit-cell volume. In
case of the ZrO2–AlO1.5 system the increase of the Al3+ ion
aHf
= 0.5098 + (0.0203 r − 0.00022)m
= 0.5120 + (0.0212 r − 0.00023)m
(4)
(5)
nm
aZr
nm
˚
(0.54 A, [43]) content up to 10 mol% caused a decrease of the
unit-cell volume. The rate of the decrease is bigger than for
theZrO2–FeO1.5 system, butsmallerthanitshouldbeaccord-
ing to the relations proposed by Kim et al. [10,11]. However,
in the concentration range between 10 and 50 mol% of Al3+
ions, the increase of the Al3+ content had very small influence
where aHf and aZr are lattice constants of the c-HfO2- and c-
ZrO2-type solid solutions, respectively, r is the difference
in ionic radius of dopant cation (R3+) and the host cation

ˇ
G. Stefanic´ et al. / Journal of Alloys and Compounds 388 (2005) 126–137
133
Fig. 6. Whole-powder-pattern decomposition results for the crystallization products from the HfO2–AlO1.5 system and Si as internal standard. Observed data
are shown by squares, the calculated one by full line. Short and long vertical bars marked the Bragg reflection positions of the crystallization products and
standard, respectively. The differences between the observed and calculated intensities are plotted in the same scale below each pattern.
on the unit-cell volume of the t-HfO2- and t-ZrO2-type solid
solutions (Fig. 8). These results cannot be explained by stan-
dard model of cation substitution and random distribution of
oxygen vacancies.
product, obtained after calcination of the amorphous precur-
sors from ZA system with 10, 40 and 50 mol% of Al3+ ions
at 800 ◦C, show only one type of grains (Fig. 9). The results
of quantitative analysis (Table 2) showed to be in good
agreement with initial molar fraction of the samples. The alu-
minium distribution showed to be uniform in the whole area
of the image, which also indicated the formation of the solid
solution. However, precise measurement of unit-cell param-
eters shows that an increase of the aluminium fraction above
10 mol% has very small influence on the unit-cell volume of
the HfO2- or ZrO2-type solid solutions (Fig. 8). Such results
are hard to explain when we consider the significantly smaller
3.4. The solubility of Al3+ ions
No sign of the diffraction lines of aluminium oxide phases
in the XRD patterns of HA and ZA crystallization products
with 50 mol% of aluminium indicates very high solubility of
Al3+ ions. High solubility was also indicated by the results
of SEM/EDX analysis. SEM images of the crystallization

ˇ
134
G. Stefanic´ et al. / Journal of Alloys and Compounds 388 (2005) 126–137
Table 2
Refined values of unit-cell parameters of the t-HfO2- and t-ZrO2-type solid
solutions with different amount of incorporated Al3+ ions
3
˚
˚
˚
Phase
a (A)
c (A)
V (A )
Rwp
t-Hf0.90Al0.10O1.95 3.5682 (8)
5.0451 (14)
5.0360 (13)
5.0388 (14)
5.0293 (14)
5.1475 (14)
5.0918 (9)
5.0691 (8)
5.0643 (6)
5.0672 (9)
5.0637 (8)
5.0643 (5)
64.23 (5) 0.0855
128.47 (8)a
5.0462 (11)a
t-Hf0.80Al0.20O1.90 3.5643 (8)
63.98 (5) 0.0823
127.96 (7)a
5.0407 (11)a
t-Hf0.70Al0.30O1.85 3.5661 (8)
64.08 (5) 0.0838
128.18 (8)a
5.0441 (11)a
t-Hf0.60Al0.40O1.80 3.5634 (8)
63.87 (5) 0.0979
127.74 (8)a
5.0394 (11)a
t-ZrO2
3.6057 (6)
5.0992 (8)a
66.92 (4) 0.0586
133.85 (7)a
t-Zr0.95Al0.05O1.98
t-Zr0.90Al0.10O1.95
t-Zr0.80Al0.20O1.90
t-Zr0.70Al0.30O1.85
t-Zr0.60Al0.40O1.80
3.5974 (5)
5.0875 (8)a
65.89 (3) 0.0726
131.79 (5)a
3.5858 (5)
5.0711 (7)a
65.18 (3) 0.0828
130.36 (5)a
3.5832 (4)
5.0674 (6)a
65.00 (2) 0.0675
129.99 (4)a
3.5904 (6)
5.0776 (8)a
65.32 (3) 0.0765
130.64 (5)a
3.5835 (5)
5.0678 (7)a
65.03 (3) 0.0618
130.05 (5)a
t-Zr0.50Al0.50O1.75
3.5816 (4)
5.0651a
64.964
129.93a
0.0683
a
Related to fluorite-type lattice.
4+
radius of Al3+ ion (0.54 A) compared to the radii of Hf
˚
˚
4+
˚
(0.83 A) and Zr (0.84 A) ions [43]. Here, we would like
to pay attention to the publication of Balmer et al. [34]. The
authors used 27Al NMR to investigate local environment of
aluminium atoms in a series of metastable Zr1−xAlxO2−x/2
crystalline materials with Al3+ content up to 57 mol%. The
obtained results showed that the products with Al2O3 content
above 10 mol% contain Al3+ ions in four-fold coordination,
which cannot be explained by the standard model of cation
substitution in the solid solutions. The authors proposed
two possible explanations of these results. The first one
associated the four-fold coordinated aluminium with a
Table 3
The results of EDX analysis (program SEMQuant) of the crystallization
product from ZrO2–AlO1.5 system with 10, 40 and 50 mol% of Al3+ ions
calcinated at 800 ◦C
Initial Al/Zr ratio
0.11
Element
At.%
Obtained Al/Zr ratio
0.09
Fig. 7. Whole-powder-pattern decomposition results for the crystallization
products from the ZrO2–AlO1.5 system and Si as internal standard. Ob-
served data are shown by squares, the calculated one by full line. Short
and long vertical bars marked the Bragg reflection positions of the crys-
tallization products and standard, respectively. Differences between the ob-
served and calculated intensities are plotted in the same scale below each
pattern.
O
Al
Zr
73.01
2.30
24.69
0.67
1.00
O
Al
Zr
70.88
11.41
17.71
0.64
0.99
O
Al
Zr
69.90
15.02
15.08

ˇ
G. Stefanic´ et al. / Journal of Alloys and Compounds 388 (2005) 126–137
135
Fig. 8. Influence of the dopant cations (Al3+ or Fe3+) content on the unit cell-volume of the corresponding fluorite-type HfO2 and ZrO2 solid solutions. Dash
lines represent theoretical values calculated from the empirical relations proposed by Kim et al. [10,11].
residual amorphous phase, present in the samples due to the
incomplete crystallization. The second explanation assumed
the existence of aluminium-rich clusters as disordered
regions within the individual crystallites or on their surfaces.
The authors concluded that solid solutions with random
distributions of oxygen vacancies exist only in the products
with molar fractions of aluminium below 10 mol% [34]. This
conclusion is in good agreement with our results obtained
Fig. 9. SEM image of the crystallization products from the ZrO2–AlO1.5 system with 50 mol% of AlO1.5, obtained after 3 h of calcination at 800 ◦C in the air
atmosphere and cooling to room temperature.

ˇ
G. Stefanic´ et al. / Journal of Alloys and Compounds 388 (2005) 126–137
136
from the measurements of the unit-cell volume. In case of the
products with molar fraction of aluminium above 10 mol%,
the second explanation proposed by Balmer et al. [34] seems
to be more realistic. SEM images of the crystallization
product with a ratio Zr/Al = 1 indicate the presence of only
one type of grains. No sign of the significant amount of
amorphous phase, required to compensate the difference in
the aluminium content between 10 and 50 mol% could be
found.
References
[1] I. Clark, D.H. Reynolds, Ind. Eng. Chem. 29 (1937) 711–715.
ˇ
[2] G. Stefanic´, S. Music´, Croat. Chem. Acta 75 (2002) 727–767.
ˇ
[3] G. Stefanic´, S. Music´, S. Sekulic´, Thermochim. Acta 273 (1996)
119–133.
[4] A. Larbot, T. Hours, P. Bergez, J. Charpin, L. Cot, J. Non-Cryst.
Solids 147/148 (1992) 85–91.
[5] L.-M. Tau, R. Srinivasan, R.J. De Angelis, T. Pinder, B.H. Davis, J.
Mater. Res. 3 (3) (1988) 561–562.
[6] R. Srinivasan, B.H. Davis, J. Am. Ceram. Soc. 73 (6) (1990)
1780–1782.
ˇ
[7] G. Stefanic´, S. Music´, Thermochim. Acta 373 (2001) 59–67.
ˇ
[8] G. Stefanic´, B. Grzˇeta, K. Nomura, R. Trojko, S. Music´, J. Alloys
Compd. 327 (2001) 151–160.
4. Conclusion
[9] S.M. Ho, Mater. Sci. Eng. 54 (1982) 23–29.
[10] D.-J. Kim, J. Am. Ceram. Soc. 72 (1989) 1415–1421.
[11] D.-J. Kim, S.-H. Hyun, S.-G. Kim, M. Yashima, J. Am. Ceram. Soc.
77 (1994) 597–599.
[12] M. Yashima, H. Takahashi, K. Ohtake, T. Hirose, M. Kakihana, H.
Arashi, Y. Ikuma, Y. Suzuki, M. Yoshimura, J. Phys. Chem. Solids
157 (1996) 289–295.
[13] M. Yashima, N. Ishizawa, M. Yoshimura, J. Am. Ceram. Soc. 75
(1992) 1541.
[14] M.F. Trubelja, V.S. Stubican, Solid State Ionics 49 (1991) 89–
97.
The results of the DTA showed that the crystallization tem-
perature of the amorphous precursors of the HfO2–AlO1.5
and ZrO2–AlO1.5 systems increased with increasing AlO1.5
content, from 530 ◦C (0 mol% AlO1.5) to 940 ◦C (60 mol%
AlO1.5) and from 405 ◦C (0 mol% AlO1.5 to 915 ◦C (60 mol%
AlO1.5), respectively. These results indicated that in this con-
centration range amorphous precursors exist in the form of
a co-gel. The results of phase analysis obtained using XRD,
laser Raman spectroscopy and SEM/EDX showed only the
presence of HfO2- and ZrO2-type solid solutions in the crys-
tallization products with AlO1.5 contents ≤50 mol%. These
results suggested an extended capability for the formation of
solid solutions in these metastable phases. However, precise
measurement of unit-cell parameters shows that an increase
of the aluminium content above 10 mol% has a very small
influence on the unit-cell volume of the HfO2- or ZrO2-type
solid solutions. Such results could not be explained by the
standard model of cation substitution and random distribution
of oxygen vacancies due to the significantly smaller radius of
[15] V.V. Kharton, A.A. Yaremchenko, E.N. Naumovich, F.M.B. Marques,
J. Solid State Electrochem. 4 (2000) 243–266.
[16] C.J. Howard, R.J. Hill, B.E. Reichert, Acta Crystallogr. Sect. B 44
(1988) 116–120.
[17] P. Li, I.-W. Chen, J.E. Penner-Hahn, J. Am. Ceram. Soc. 77 (1994)
118–128.
ˇ
[18] S. Popovic´, B. Grzˇeta, G. Stefanic´, I. Cako´-Nagy, S. Music´, J. Alloys
Compd. 241 (1996) 10–15.
ˇ
[19] S. Popovic´, G. Stefanic´, S. Music´, Mater. Lett. 31 (1997) 19–
22.
ˇ
[20] G. Stefanic´, S. Popovic´, S. Music´, Mater. Lett. 36 (1998) 240–
244.
[21] S. Davison, R. Kershaw, K. Dwight, A. Wold, J. Solid State Chem.
73 (1988) 47–51.
Al3+ ion (0.54 A) compared to the radii of Hf (0.83 A) and
4+
˚
˚
[22] F.J. Berry, M.H. Loretto, M.R. Smith, J. Solid State Chem. 83 (1989)
Zr4+ (0.84 A) ions. It was concluded that solid solutions with
˚
91–99.
random distributions of oxygen vacancies exist only in the
products with molar fractions of aluminium below 10 mol%.
In case of the products with molar fractions of aluminium
above 10 mol%, the model proposed by Balmer et al. [34],
which assume the existence of aluminium-rich clusters as dis-
ordered regions within the individual crystallites or on their
surfaces, seems to be the most realistic. The appearance of
such clusters in the products with aluminium contents above
10 mol% is probably the reason why only the tetragonal poly-
morph of hafnia and zirconia could be stabilized. Although
the results of whole-powder-pattern decomposition indicated
stabilization of the cubic polymorphs of HfO2 and ZrO2 in
the products with AlO1.5 content ≥10 mol% (axial ratio cf/af
approach (1), Raman spectra of those crystallization products
contained all six Raman active modes of vibration, typical for
the tetragonal polymorphs of HfO2 and ZrO2. The tetragonal
symmetry of these samples can be attributed to the displace-
ment of oxygen sub-lattice from the ideal fluorite positions.
On the other side of the concentration range the solubility of
Zr4+ and Hf4+ ions in the aluminium oxides lattice appeared
to be negligible.
[23] I.B. Inwang, F. Chyad, I.J. McColm, J. Mater. Chem. 5 (1995)
1209–1213.
[24] G. Stefanic´, B. Grzˇeta, S. Music´, Mater. Chem. Phys. 65 (2000)
ˇ
216.
[25] O. Yamaguchi, M. Shirai, M. Yoshinaka, J. Am. Ceram. Soc. 71
(1988) C-510–C-512.
[26] K. Ishida, K. Hirota, O. Yamaguchi, H. Kume, S. Inamura, H.
Miyamoto, J. Am. Ceram. Soc. (1994) 77.
[27] W. Zhang, E.E. lachowski, F.P. Glasser, J. Mater. Sci. 28 (1993)
6222–6232.
[28] J.S. Hong, S.D. De la Torre, K. Miyamoto, H. Miyamoto, L. Gao,
Mater. Lett. 37 (1998) 6–9.
[29] A. Mondal, S. Ram, Solid State Ionics 160 (2003) 169–181.
[30] S. Kikkawa, A. Kijima, K. Hirota, O. Yamamoto, J. Am. Ceram.
Soc. 85 (2002) 721–723.
[31] S. Moreau, M. Gervais, A. Douy, Solid State Ionics 101–103 (1997)
625–631.
[32] M.L. Balmer, F.F. Lange, C.G. Levi, J. Am. Ceram. Soc. 77 (1994)
2069–2075.
[33] M.L. Balmer, F.F. Lange, V. Jayaram, C.G. Levi, J. Am. Ceram. Soc.
78 (1995) 1489–1494.
[34] M.L. Balmer, H. Eckert, N. Das, F.F. Lange, J. Am. Ceram. Soc. 79
(1996) 321–326.
[35] M. Yoshimura, S.-T. Oh, M. Sando, K. Niihara, J. Alloys Compd.
290 (1999) 284–289.

ˇ
G. Stefanic´ et al. / Journal of Alloys and Compounds 388 (2005) 126–137
137
[36] Powder Diffraction File, International Center for Diffraction Data,
Newtown Square, PA 19073-3273, USA.
[37] H. Toraya, M. Yoshimura, S. Somiya, J. Am. Ceram. Soc. 67 (1984)
C119–C121.
[38] H. Toraya, J. Appl. Cryst. 19 (1986) 440–447.
[39] M. Yashima, T. Hirose, M. Kakihana, Y. Suzuki, M. Yoshimura, J.
Ceram. Soc. Jpn. Int. Ed. 103 (1995) 622–628.
[40] H. Fujimori, M. Yashima, S. Sasaki, M. Kakihana, T. Mori, M.
Tanaka, M. Yoshimura, Chem. Phys. Lett. 346 (2001) 217–223.
[41] H. Fujimori, M. Yashima, S. Sasaki, M. Kakihana, T. Mori, M.
Tanaka, M. Yoshimura, Phys. Rev. B 64 (13) (2001) 41041–41045.
[42] H. Fujimori, M. Yashima, M. Kakihana, M. Yoshimura, J. Am. Ce-
ram. Soc. 84 (2001) 663–665.
[43] R.D. Shannon, Acta Cryst. A 32 (1976) 751–767.