The influence of thermal treatment on phase development in ZrO2-Fe2O3 and HfO2-Fe2O3 systems

Article abstract of DOI:10.1016/S0925-8388(01)01401-3

The effects of thermal treatment (500, 600, 800 and 1100°C) of the amorphous precursors of ZrO₂-Fe₂O₃ and HfO₂-Fe₂O₃ systems, coprecipitated from aqueous solutions of the corresponding salts, were studied at room temperature (RT) by X-ray powder diffraction (XRD) and, in some cases, by laser Raman spectroscopy. The incorporation of Fe³⁺ ions partially stabilized the high temperature polymorphs of zirconia (ZrO₂) and hafnia (HfO₂). The stabilization of the cubic polymorphs of ZrO₂ and HfO₂ occurred in the solid solutions with an Fe₂O₃ content of ≥10 mol%. The presence of Fe³⁺ ions above their solid solubility limit caused the transition of metastable polymorphs of ZrO₂ and HfO₂ to monoclinic polymorphs. The terminal solid solubility limit of Fe₂O₃ in ZrO₂, as estimated at RT, was ～33 mol% after calcination at 600°C, ～9 mol% after calcination at 800°C and ～2 mol% after calcination at 1100°C. After the same thermal treatment, the terminal solid solubility limits of Fe₂O₃ in HfO₂ proved to be very similar only a little smaller. The unit-cell volume of the cubic polymorphs of ZrO₂ and HfO₂, determined by powder-pattern-fitting methods, decreased linearly with the increase in iron content, but the rate of decrease was greater in the ZrO₂ polymorph. The thermal behavior of the amorphous precursors was also followed by differential thermal analysis and thermogravimetric analysis. The crystallization temperature of both systems increased with an increase in the Fe₂O₃ content, from 405 to 730°C in the ZrO₂-Fe₂O₃ system, and from 520 to 720°C in the HfO₂-Fe₂O₃ system. The results indicated that the amorphous precursors with more than 30 mol% of Fe₂O₃ were single co-gels. The status of the Fe³⁺ ions in the ZrO₂-type lattice and the process of their segregation into the α-Fe₂O₃-phase were examined by⁵⁷Fe Mo?ssbauer spectroscopy. Significantly lower values of the hyperfine magnetic fields compared to that of α-Fe₂O₃ indicated the incorporation of Zr⁴⁺ ions into the α-Fe₂O₃ lattice, the amount of which decreased with the increasing temperature of treatment.

Full text of DOI:10.1016/S0925-8388(01)01401-3

Journal of Alloys and Compounds 327 (2001) 151–160
L
www.elsevier.com/locate/jallcom
The influence of thermal treatment on phase development in ZrO2 –Fe2O3
and HfO2 –Fe2O3 systems
a ,
a
b
a
a
ˇ
*
´
ˇ
´
G. Stefanic , B. Grzeta , K. Nomura , R. Trojko , S. Music
a
ˇ
Ruder Boskovic Institute, P.O. Box 180, 10002 Zagreb, Croatia
bSchool of Engineering, University of Tokyo, Hongo 7-3-1, Tokyo, 113 Japan
Received 9 November 2000; accepted 10 April 2001
Abstract
The effects of thermal treatment (500, 600, 800 and 11008C) of the amorphous precursors of ZrO2 –Fe2O3 and HfO2 –Fe2O3 systems,
coprecipitated from aqueous solutions of the corresponding salts, were studied at room temperature (RT) by X-ray powder diffraction
(XRD) and, in some cases, by laser Raman spectroscopy. The incorporation of Fe31 ions partially stabilized the high temperature
polymorphs of zirconia (ZrO2) and hafnia (HfO2). The stabilization of the cubic polymorphs of ZrO2 and HfO2 occurred in the solid
solutions with an Fe2O3 content of $10 mol%. The presence of Fe31 ions above their solid solubility limit caused the transition of
metastable polymorphs of ZrO2 and HfO2 to monoclinic polymorphs. The terminal solid solubility limit of Fe2O3 in ZrO2, as estimated at
RT, was |33 mol% after calcination at 6008C, |9 mol% after calcination at 8008C and |2 mol% after calcination at 11008C. After the
same thermal treatment, the terminal solid solubility limits of Fe2O3 in HfO2 proved to be very similar only a little smaller. The unit-cell
volume of the cubic polymorphs of ZrO2 and HfO2, determined by powder-pattern-fitting methods, decreased linearly with the increase in
iron content, but the rate of decrease was greater in the ZrO2 polymorph. The thermal behavior of the amorphous precursors was also
followed by differential thermal analysis and thermogravimetric analysis. The crystallization temperature of both systems increased with
an increase in the Fe2O3 content, from 405 to 7308C in the ZrO2 –Fe2O3 system, and from 520 to 7208C in the HfO2 –Fe2O3 system. The
results indicated that the amorphous precursors with more than 30 mol% of Fe2O3 were single co-gels. The status of the Fe31 ions in the
57
¨
ZrO2-type lattice and the process of their segregation into the a-Fe2O3-phase were examined by Fe Mossbauer spectroscopy.
Significantly lower values of the hyperfine magnetic fields compared to that of a-Fe2O3 indicated the incorporation of Zr41 ions into the
a-Fe2O3 lattice, the amount of which decreased with the increasing temperature of treatment.
2001 Elsevier Science B.V. All rights
reserved.
¨
Keywords: ZrO2; HfO2; Fe2O3; Chemical synthesis; Thermal analysis; X-ray diffraction; Mossbauer spectroscopy
1. Introduction
thermal decomposition of zirconium salts [1,2] or as the
crystallization product of hydrous zirconia [3,4]. 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
[5–7]. Cubic polymorphs of ZrO2 and HfO2 can be
stabilized at RT by the incorporation of suitable metal
cations, such as Ca21, Y31, Gd31, etc. [8–15].
Pure zirconia and hafnia have very similar physical and
chemical properties. According to the temperature, ZrO2
and HfO2 appear in monoclinic (m-ZrO2, m-HfO2), tetra-
gonal (t-ZrO2, t-HfO2) and cubic (c-ZrO2, c-HfO2) poly-
morphs of which only the monoclinic is thermodynamical-
ly stable at room temperature (RT) and standard pressure.
High temperature polymorphs of ZrO2 become stable at
above 11708C (t-ZrO2) and 23708C (c-ZrO2), while the
stabilization of t- and c-HfO2 require significantly higher
temperatures, 1720 and 26008C, respectively. Metastable
t-ZrO2 can be easily obtained at RT as a product of the
The incorporation of Fe31 cations can also stabilize high
temperature polymorphs of zirconia, but this stabilization
is not very efficient [15–21]. The stabilization of the cubic
polymorph of zirconia depends on both the amount of
incorporated iron cations and the preparation conditions
used [19,20]. Li et al. [15] investigated zirconia solid
solutions with Fe2O3, Ga2O3, Y2O3 and Gd2O3. The
authors found that after calcination at 13008C, only
oversized dopants (Gd31, Y31) can stabilize c-ZrO2-type
structures at RT. This stabilization was attributed to the
*
Corresponding author. Tel.: 1385-1-456-1111; fax: 1385-1-46-
80084.
E-mail address: stefanic@rudjer.irb.hr (G. Stefanic).
ˇ
´
0925-8388/01/$ – see front matter
PII: S0925-8388(01)01401-3
2001 Elsevier Science B.V. All rights reserved.

ˇ
´
G. Stefanic et al. / Journal of Alloys and Compounds 327 (2001) 151–160
152
decrease of the Zr coordination number by the introduction
of oxygen vacancies associated with the smaller Zr cation.
The incorporation of undersized dopants (Fe31, Ga31) also
introduced oxygen vacancies, but the decrease of the Zr
coordination number was smaller due to the association of
Philips diffractometer (model MPD 1880) with mono-
chromatized CuKa radiation. Corundum, a-Al2O3
(99.999%, Merck), was used as internal standard (space
¯
group R3c, unit-cell parameters in terms of hexagonal axes
˚
˚
at 258C: a54.7584(2) A, c512.9903(5) A [22]). XRD
patterns were scanned in steps of 0.028 (2u ), in the 2u
range from 20 to 1108, 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.
ˇ
the vacancies with the smaller dopant cation [15]. Stefanic
et al. [21] investigated the influence of oxygen on the
thermal behavior of the ZrO2 –Fe2O3 system. It was found
that in the presence of oxygen vacancies, introduced into
the ZrO2 lattice by calcination at very low pressure (|43
1023 Pa), Fe31 ions caused a transition of high-tempera-
ture polymorphs of zirconia into the m-ZrO2-type struc-
tures. These results showed that aliovalent undersized
dopants have both a stabilizing and destabilizing influence
on the high-temperature polymorphs of zirconia [21]. In
the present work we investigated and compared the
influence of thermal treatment on the phase development in
ZrO2 –Fe2O3 and HfO2 –Fe2O3 systems.
The thermal behavior of the amorphous samples was
also examined by differential thermal analysis (DTA) and
thermogravimetric (TG) analysis. The prepared samples
were heated in an oxygen atmosphere up to 8008C with a
rate of 108C min21, using the standard instrumentation.
The status of the Fe31 ions in the ZrO2 –Fe2O3 system
57
¨
was monitored by Fe Mossbauer spectroscopy. The
¨
Mossbauer spectra were recorded at room temperature
using standard instrumentation. A 57Co source in the
¨
rhodium matrix was used. Mossbauer spectra were de-
2. Experimental
convoluted using the MossWinn program.
Amorphous precursors of ZrO2 –Fe2O3 and HfO2 –
Fe2O3 systems at the ZrO2- and HfO2-rich side of the
concentration range were coprecipitated from an aqueous
solution of ZrO(NO3)2?2H2O and Fe(NO3)3?9H2O salts
and an aqueous solution of HfCl4 and FeCl3?6H2O salts,
respectively. The coprecipitation was performed by the
addition of 25% NH3(aq.) up to pH|10.4. All the chemi-
cals used were of analytical-grade. The solid phase was
separated from the corresponding liquid using an ultra-
speed centrifuge (operational range up to 20 000 rpm), then
washed repeatedly in doubly distilled water and dried at
708C for 24 h. The samples were calcined at 500, 600, 800
and 11008C for 2 h and the phase compositions of the
obtained crystallization products were analyzed at RT
using X-ray powder diffraction (XRD). Laser Raman
(DILOR Z24) spectroscopy was also used in the cases
when the presence of the cubic or tetragonal phase in the
ZrO2 –Fe2O3 system could not be clearly distinguished by
XRD. XRD patterns were recorded using an automated
3. Results and discussion
3.1. Phase analysis
The initial molar compositions and the results of phase
analysis obtained after the calcination and cooling of the
amorphous precursors of the ZrO2 –Fe2O3 (ZF) and
HfO2 –Fe2O3 (HF) systems are given in Tables 1 and 2.
After calcination at 5008C, the XRD patterns of the ZF
crystallization products with an Fe2O3 content of up to 3
mol% contained diffraction lines typical of t- or c-ZrO2
and m-ZrO2 (Fig. 1). The corresponding HF crystallization
products contained only the Hm-phase, closely structurally
related to m-HfO2. The Hc-phase, closely structurally
related to c-HfO2, appeared in the HF crystallization
product with 10 mol% of Fe2O3. The products of both
systems with an Fe2O3 content of $20 mol% are still
amorphous.
Table 1
The notation of the samples from the ZrO2 –Fe2O3 system and the results of phase analysis
Sample
x
Phase composition (volume fraction)
ZrO2
Fe2O3
5008C
6008C
8008C
11008C
ZF0
ZF1
ZF2
ZF3
ZF4
ZF5
ZF5A
ZF6
1
0
Zt(0.62)1Zm(0.38)
Zt(0.99)1Zm(0.01)
Zt
Zc
Am1Zc
Am
–
Zm(0.85)1Zt(0.15)
Zm(0.67)1Zt(0.33)
Zt
Zc
Zc
Zc
Zc 1F
Zc 1F
Zm(0.95)1Zt(0.05)
Zm(0.95)1Zt(0.05)
Zm(0.84)1Zt(0.16)
Zt 1F
Zt 1F1Zm
Zt 1F1Zm
Zm
–
0.99
0.97
0.90
0.80
0.70
0.65
0.50
0.01
0.03
0.10
0.20
0.30
0.35
0.50
Zm 1F
Zm 1F
–
Zm 1F
–
–
Am1F
Zt 1F1Zm
Zm 1F
Zm 5phase structurally similar to m-ZrO2, Zt 5phase structurally similar to t-ZrO2; Zc 5phase structurally similar to c-ZrO2, Am5amorphous phase;
F5phase structurally similar to a-Fe2O3.

ˇ
´
G. Stefanic et al. / Journal of Alloys and Compounds 327 (2001) 151–160
153
Table 2
The notation of the samples from the HfO2 –Fe2O3 system and the results of phase analysis
Sample
x
Phase composition
HfO2
Fe2O3
5008C
6008C
8008C
11008C
HF0
HF1
HF2
HF3
HF4
HF5
HF6
1
–
m-HfO2
Hm
Hm 1Hc 1Am
Am
Am
Am
–
m-HfO2
Hm
Hm
Hc 1Hm
Hc
Hc 1F
Hc 1F1Hm
m-HfO2
Hm
Hm
Hm 1F1Hc
Hm 1F
Hm 1F
Hm 1F
m-HfO2
Hm 1F
Hm 1F
Hm 1F
Hm 1F
Hm 1F
–
0.99
0.97
0.90
0.80
0.70
0.50
0.01
0.03
0.10
0.20
0.30
0.50
Hm 5phase structurally similar to m-HfO2, Hc 5phase structurally similar to c-HfO2; Am5amorphous phase, F5phase structurally similar to a-Fe2O3
After calcination at 6008C all the samples crystallized.
The phase compositions of the crystallization products
with an Fe2O3 content of up to 10 mol% remain approxi-
mately the same, but the content of the Zm-phase, closely
structurally related to m-ZrO2, in the crystallization prod-
ucts of the samples ZF0 and ZF1 proved to be higher. In
both systems the increase in the Fe2O3 content caused the
increase of the tetragonal or cubic phases followed with
the decrease of the Zm and Hm-phases. The Zm and
Hm-phases disappeared from the ZF and HF crystallization
products with 3 and 20 mol% of Fe2O3, respectively. The
Hm-phase reappeared in the crystallization product of the
HF6 sample (50 mol% of Fe2O3). The F-phase, closely
structurally related to a-Fe2O3, appeared in the HF and ZF
crystallization products with 30 and 35 mol% of Fe2O3,
respectively.
The laser Raman spectra of the corresponding ZF
crystallization products with an Fe2O3 content of up to 3
mol% contained the bands at 267 and 148 cm21 typical of
t-ZrO2 [4]. The spectrum of the crystallization product
with 10 mol% of Fe2O3 contained only a broad band at
|610 cm21 typical of c-ZrO2 [23,24]. These results
indicated that the amorphous precursors with an Fe2O3
content of $10 mol% crystallized into the Zc-phase,
closely structurally related to c-ZrO2, while the smaller
amount of incorporated Fe2O3 stabilized the Zt-phase,
closely structurally related to t-ZrO2.
The volume fractions of the Zm and Zt-phases in the ZF
crystallization products with an Fe2O3 content of up to 3
mol% were estimated from the integrated intensities of the
¯
diffraction lines and 111 and 111 of m-ZrO2, and the line
101 of t-ZrO2 following the procedure proposed by Toraya
et al. [25]. The volume fractions of Zm (nm) and Zt (nt) are
given by
1.311x
1 1 0.311x
nm 5
(1)
(2)
nt 5 1 2 nm
where x is
¯
Im(111) 1 Im(111)
]
x 5
(3)
¯
Im(111) 1 Im(111) 1 It(101)
The volume fraction of the Zt-phase increased with the
increase in the Fe2O3 content and decreased with the
increase of the treatment temperature (Fig. 2). The in-
corporation of 3 mol% of Fe2O3 stabilized the Zt-phase
after calcination at 500 and 6008C, but after calcination at
8008C, the Zm-phase was still the dominant one (nm 5
0.84). The Zm-phase disappeared in the crystallization
product with 10 mol% of Fe2O3 and reappeared in the
crystallization products of the samples with an Fe2O3
content of $20 mol% (Fig. 3a). The presence of diffrac-
tion lines typical of t-ZrO2 and a-Fe2O3 in the XRD
Fig. 1. Characteristic part of XRD patterns of the samples from the
ZrO2 –Fe2O3 system obtained after calcination at 5008C.

ˇ
´
G. Stefanic et al. / Journal of Alloys and Compounds 327 (2001) 151–160
154
patterns of the ZF crystallization products with Fe2O3
content of $10 mol% indicate that the calcination at
8008C caused a significant decrease in the iron solubility
followed by the Zc →Zt transition. The HF crystallization
products with an Fe2O3 content of up to 3 mol%, obtained
after calcination at 8008C, showed only diffraction lines of
the Hm-phase. Diffraction lines of the Hc and F-phases
appeared in the XRD pattern of the product with 10 mol%
of Fe2O3 (Fig. 3b). Further increases in the Fe2O3 content
caused the disappearance of the Hc diffraction lines and
the increase of the diffraction lines of the F-phase, which
indicate a decrease in iron solubility followed by the
Hc →Hm transition.
The results show that the presence of Fe2O3 has both a
stabilizing and destabilizing influence on the high tempera-
ture polymorphs of ZrO2 and HfO2. The incorporation of
aliovalent cations, such as Fe31, introduces oxygen vac-
ancies that stabilize high-temperature polymorphs of ZrO2
and HfO2 [26]. However, once introduced oxygen vac-
ancies tend to associate with smaller cations [15], so the
presence of Fe31 ions destabilizes the high-temperature
polymorph of ZrO2 and HfO2. In our previous inves-
tigation, oxygen vacancies were introduced into the ZrO2
Fig. 2. Influence of the initial Fe2O3 content on the volume fraction of
the Zt-phase in the total content of ZrO2-type structures.
Fig. 3. Characteristic part of XRD patterns of the samples from the ZrO2 –Fe2O3 system (a) and HfO2 –Fe2O3 system (b), obtained after calcination at
8008C.

ˇ
´
G. Stefanic et al. / Journal of Alloys and Compounds 327 (2001) 151–160
155
lattice by calcination at very low pressure, which resulted
in the stabilization of c-ZrO2 [21]. The incorporation of
Fe31 ions caused transition of this cubic-phase into the
phase closely structurally related to m-ZrO2 [21]. The
presence of the F-phase in the ZF and HF products with
Fe2O3 content of $10 mol% indicate that after calcination
at 8008C the solubility limits of Fe2O3 decrease to below
10 mol%. The content of iron above the solubility limits
has no influence on the amount of oxygen vacancies
introduced and therefore could only destabilize the high-
temperature polymorphs of ZrO2 and HfO2.
After calcination at 11008C the Zm and Hm-phases were
dominant in the crystallization products of all the samples.
Besides these phases, the F-phase appeared in the crys-
tallization products of the samples HF1 (1 mol% of
Fe2O3) and ZF2 (3 mol% of Fe2O3) and its fraction
increased with further increases in the Fe2O3 content.
Fig. 4. Influence of the calcination temperature on the terminal solid
solubility limits of Fe2O3 in the ZrO2 and HfO2 lattices.
3.2. The solubility of Fe2O3
Solid solubility limits of Fe2O3 in the ZrO2 and HfO2
lattice were estimated at RT from the dependence of the
diffraction line intensities of the F-phase on the initial
content of Fe2O3 and by extrapolation to the zero intensity.
The results show that the solubility of Fe2O3 in the ZrO2
and HfO2 lattices decreased with an increase in the
treatment temperature in a similar way. The terminal solid
solubility limits of Fe2O3 in ZrO2 were |33 mol% after
calcination at 6008C, |9 mol% after calcination at 8008C
and |2 mol% after calcination at 11008C. The solid
solubility limits of Fe2O3 in the HfO2 lattice were
constantly a little smaller (Fig. 4), estimated to be |29
mol% after calcination at 6008C, |8 mol% after calcina-
tion at 8008C and |1 mol% after calcination at 11008C.
Rp 50.043 and Rwp 50.0586 for sample ZF0, Rp 50.041
and Rwp 50.0565 for sample ZF2, Rp 50.041 and Rwp 5
0.0592 for sample ZF3, Rp 50.035 and Rwp 50.0480 for
sample ZF4, Rp 50.063 and Rwp 50.0915 for sample ZF5
and Rp 50.048 and Rwp 50.0644 for sample ZF5A. Fig. 5
shows the observed and calculated powder patterns of
some investigated samples.
The same procedure was used to determine the unit-cell
parameters of the Hc-phase in the samples HF3, HF4 and
HF5 calcinated at 6008C (Fig. 6). The fitting procedure
converged to the discrepancy factors Rp 50.062 and Rwp 5
0.0855 (HF3), Rp 50.050 and Rwp 50.0883 (HF4) and
Rp 50.042 and Rwp 50.0585 (HF5).
The refined unit-cell parameters of both the ZF and HF
products are listed in Table 3. The results show that the
unit-cell volume of both the Zc and Hc-phases decreased
linearly with an increase in Fe31 content (Fig. 7). The
difference in the unit-cell volumes of the crystallization
products depends mostly on the ionic radii of their cations.
3.3. Lattice parameters
The incorporation of Fe31 ions shifts the diffraction
lines of the Zt and Zc-phases to the higher Bragg angle,
indicating a decrease in the lattice parameters. The unit-
cell parameters of the crystallization products of the
samples ZF0 (5008C), ZF2 (6008C), ZF3 (6008C), ZF4
(6008C), ZF5 (6008C) and ZF5A (6008C) were first
determined using the method proposed by Toraya [27],
employing the UNITCELL computer program. The Bragg
angle position, 2u, of the diffraction lines of the samples
and the internal standard, a-Al2O3, were determined by
the individual profile fitting method (PROFIT program [28])
and taken as input data for the UNITCELL program. A
polynomial model of the peak shift correction function was
applied. Unit-cell parameters were then refined by the
whole-powder-pattern decomposition method (WPPF pro-
gram [28]). The fitting was performed in the scanned 2u
range, from 208 to 1108, using the split-type pseudo-Voigt
profile function and the polynomial background model.
The fitting procedure converged to the discrepancy factors
The ionic radius of Zr41 is a little larger (0.84 A) than that
˚
of Hf41 (0.83 A) [29], which resulted in a somewhat larger
˚
unit-cell of c-ZrO2 compared to c-HfO2. The incorporation
of an undersized dopant, such as Fe31 (0.65 A [29]),
˚
caused a decrease in both Zc and Hc unit-cells. The rate of
the decrease in the ZF system was higher, probably
because of the greater difference between the ionic radii of
Zr41 and Fe31. In the products with 30 mol% of Fe2O3,
the estimated unit-cell of the Zc-phase was a little smaller
than that of the Hc-phase, which can be attributed to the
incomplete incorporation of Fe31 ions into the HfO2 lattice
(the presence of the F-phase). The stabilization of the cubic
polymorphs of ZrO2 and HfO2 occurred only in the
crystallization products with an Fe2O3 content of $10
mol%. In the case of the ZF system a smaller amount of
Fe2O3 stabilized the Zt-phase. However, the unit-cell
volume of the Zt-phase, present in the products of the

ˇ
´
G. Stefanic et al. / Journal of Alloys and Compounds 327 (2001) 151–160
156
Fig. 6. XRD patterns of the HF3, HF4 and HF5 samples. Observed data
are shown by squares, the refined pattern by a full line. The positions of
hafnia diffraction lines are shown by short bars and those of the internal
standard, a-Al2O3, by long bars. The difference, D, between the observed
and refined patterns is also shown in the same scale. Radiation5CuKa
(40 kV, 30 mA).
Fig. 5. XRD patterns of the ZF2, ZF3, ZF4 and ZF5A samples. Observed
data are shown by squares, the refined pattern by a full line. The positions
of zirconia diffraction lines are shown by short bars and those of the
internal standard, a-Al2O3, by long bars. The difference, D, between the
observed and refined patterns is also shown in the same scale. Radiation5
CuKa (40 kV, 30 mA).
Table 3
Refined values of unit-cell parameters of the Zt, Zc and Hc-phases from
the product of the sample ZF0 calcinated at 5008C and the samples ZF2,
ZF3, ZF4, ZF5, HF3, HF4 and HF5 calcinated at 6008C
3
samples ZF0 and ZF2, proved to be on the same line (Fig.
7). This indicates that the transition from a tetragonal to a
cubic polymorph has a very small influence on the unit-cell
volume, probably because the coordination number of the
Zr41 ion does not change in this transition.
Kim et al. [9,10] proposed empirical equations that
relate lattice constants of the c-ZrO2- and c-HfO2-type
solid solutions with the type (ionic radius, valency) and
concentration of dopant cations. For the ZrO2 –RO1.5 and
HfO2 –RO1.5 systems these equations can be expressed as
follows:
˚
˚
˚
Sample
Phase
a (A)
c (A)
V (A )
ZF0
t-ZrO2
3.6057(6)
5.0992a
5.1475(14)
5.1596(8)
66.923
133.846a
66.568
ZF2
t-Zr0.94Fe0.06O1.97
3.5919(3)
5.0797a
133.136a
129.885
126.038
121.958
120.407
128.382
125.158
122.327
ZF3
ZF4
ZF5
ZF5A
HF3
HF4
HF5
c-Zr0.82Fe0.18O1.91
c-Zr0.67Fe0.33O1.84
c-Zr0.54Fe0.46O1.77
c-Zr0.50Fe0.50O1.75
c-Hf0.82Fe0.18O1.91
c-Hf0.67Fe0.33O1.84
c-Hf0.55Fe0.45O1.78
5.0643(3)
5.0138(4)
4.9591(7)
4.9380(8)
5.0447(2)
5.0021(9)
4.9641(8)
b
aZr (nm) 5 0.5120 1 (0.0203 Dr 2 0.00022)m
(1)
b
a Related to fluorite type lattice.
b Composition estimated from the solid solubility limits of Fe2O3 in
HfO2.
aHf (nm) 5 0.5098 1 (0.0203 Dr 2 0.00022)m
(2)
where aZr and aHf are lattice constants of the c-ZrO2 and

ˇ
´
G. Stefanic et al. / Journal of Alloys and Compounds 327 (2001) 151–160
157
ZrO2 –RO1.5 and HfO2 –RO1.5 systems in which the dop-
ant R31 ion has a similar or greater ionic radius compared
to the Zr41 or Hf41 ions. Our results show that these
equations could not be used for the ZrO2 –FeO1.5 and
HfO2 –FeO1.5 systems, probably because the solid solu-
tions with significantly smaller dopant cations, such as
Fe31, have different local structure around vacancies [15].
3.4. Thermal behavior
The results of TG analysis showed that during heating to
8008C the amorphous precursors of the HF system lost
between 13.3 and 20.5% of their initial mass, while the
amorphous precursors of the ZF system lost between 20.7
and 25.9% of their initial mass. The most significant
weight loss occurred during heating in the temperature
range between 50 and 2508C, which corresponds to the
endothermic peaks of dehydration in the DTA curves.
The results of the DTA showed that the crystallization
temperature of the amorphous samples of the ZF and HF
systems increased with an increase in the Fe2O3 content
(Fig. 8). However, the rate of the increase was sig-
nificantly higher in the ZF system. Pure HfO2 crystallizes
at a 1158C higher temperature (5208C) than pure ZrO2, but
the crystallization temperature of the sample HF6 (50
mol% of Fe2O3) proved to be 108C lower (7208C) than
that of the sample ZF6. The DTA curve of the sample ZF6
contains a small exothermic peak at |3008C, which can be
attributed to the appearance of the F-phase (Fig. 1). Pure
amorphous iron(III)hydroxide crystallized into a-Fe2O3 at
2458C, so the absence of the F-phase in the crystallization
products obtained after treatment of the samples with
Fe2O3 content of $30 mol% inside the DTA (8008C)
indicated that these precursors are in the form of a co-gel.
Calcination of the amorphous precursors caused both
crystallization and segregation of the initial single phase.
However, the activation energy required for the crys-
tallization of the amorphous precursors is smaller than the
activation energy required for the segregation of the
starting single phase. Due to this fact, the first crys-
tallization product of the samples is a metastable phase
with an extended capability for the formation of solid
solutions.
Fig. 7. Influence of the Fe2O3 content on the unit-cell volumes of Zt (h),
Zc (j) and Hc (d) phases.
c-HfO2 type solid solutions, respectively; Dr is the differ-
ence in ionic radius of dopant cation (R31) and the host
cation (Zr41 or Hf41); and m is the concentration of the
dopant cation (mol%) in the form of RO1.5
.
When R5Fe Eqs. (1) and (2) become:
aZr (nm) 5 0.5120 2 6.33 3 1024 m
aHf (nm) 5 0.5098 2 5.85 3 1024 m
(3)
(4)
The Eqs. (3) and (4) show in agreement with our results
that the incorporation of Fe31 ions caused a linear decrease
of the lattice parameters of c-ZrO2 and c-HfO2 and that the
rate of the decrease is greater for ZrO2-type solid solu-
tions. However, the rate of the decrease obtained from our
data showed to be about four times smaller than it was
expected by these equations. The results of regression
analysis of our data give the following relation between the
lattice constants (aZr and aHf) and the m value:
A similar effect was also observed in the system Er2O3 –
Fe2O3 [30]. For the stoichiometric ratio used for garnet
formation (Er3Fe5O12), the samples were amorphous up to
600–6508C, while a mixture of crystalline oxide phases
appeared at 7508C.
¨
3.5. Mossbauer spectra
aZr (nm) 5 0.5099 2 1.65 3 1024 m
aHf (nm) 5 0.5076 2 1.44 3 1024 m
(5)
(6)
¨
Mossbauer spectra of the products obtained after the
calcination of the samples ZF2, ZF3, ZF5 and ZF6 at 500,
600 and 8008C are shown in Figs. 9–11. The 57Fe
¨
Mossbauer parameters of the selected products are given in
The data used to obtain Eqs. (1)–(4) are taken from the
Table 4. All the products showed the superposition of two

ˇ
´
G. Stefanic et al. / Journal of Alloys and Compounds 327 (2001) 151–160
158
Fig. 8. Effect of the iron content on the temperature of the crystallization
of the amorphous hydroxides of the ZrO2 –Fe2O3 (ZF) and HfO2 –Fe2O3
(HF) systems.
quadrupole doublets, with the splitting values higher than
for superparamagnetic a-Fe2O3 (|0.50 mms21 or less),
thus indicating the presence of two structural environments
for iron ions. These two quadrupole doublets do not
automatically mean two defined structural positions of
iron. They can be due to the presence of two nonequivalent
distributions of iron in the solid solutions.
57
Hyperfine magnetic splitting, indicating the presence of
¨
Fig. 9. Fe Mossbauer spectra of ZF2, ZF3, ZF5 and ZF6 samples
¨
the F-phase, appeared in the Mossbauer spectrum of the
calcinated at 5008C and recorded at room temperature.
sample ZF6 (50 mol% Fe2O3) calcined at 6008C (Fig. 10).
After calcination at 8008C, hyperfine magnetic splitting
¨
appeared in the Mossbauer spectra of the products with an
Fe2O3 content of $10 mol% (Fig. 11), while after
calcination at 11008C it is present in the spectra of the
products with an Fe2O3 content of $3 mol% [18].
Hyperfine magnetic fields (HMFs) increase with an in-
crease in temperature, but the obtained values were
significantly lower than that of pure a-Fe2O3 (HMF5517
kOe). For example, the HMF value of the sample ZF6 was
497 kOe after calcination at 6008C and increased to 502
kOe after calcination at 8008C. This effect indicates that,
as in the case of the ZrO2-type solid solution, the
amorphous precursors of the a-Fe2O3-type solid solutions
crystallized into the metastable phase with extended capa-
bility for the incorporation of Zr41 ions. The increase in
the temperature caused a release of Zr41 ions from the
a-Fe2O3 lattice that caused a gradual increase in the HMF
value.

ˇ
´
G. Stefanic et al. / Journal of Alloys and Compounds 327 (2001) 151–160
159
57
¨
Fig. 11. Fe Mossbauer spectra of ZF2, ZF3, ZF5 and ZF6 samples
57
¨
Fig. 10. Fe Mossbauer spectra of ZF2, ZF3, ZF5 and ZF6 samples
calcinated at 8008C and recorded at room temperature.
calcinated at 6008C and recorded at room temperature.
4. Conclusion
limit of Fe2O3 in ZrO2 decreased with the increase in
thermal treatment from |33 mol% at 6008C to the |2.0
mol% at 11008C. The same decrease occurred in the
HfO2 –Fe2O3 system, but the solid solubility of Fe2O3 was
The crystallization temperature of the ZrO2 –Fe2O3 and
HfO2 –Fe2O3 systems increased with an increase in the
Fe2O3 content, indicating that their amorphous precursors
with Fe2O3 contents of up to 30 mol% of are single
co-gels. These co-gels crystallized into ZrO2- and HfO2-
type solid solutions with extended capability for the
incorporation of Fe31 ions. The terminal solid solubility
¨
constantly a little smaller. The results of Mossbauer
spectroscopy indicated that on the other side of the
concentration range, the amorphous precursors of the a-
Fe2O3-type solid solutions crystallized into the metastable
phase with extended capability for the incorporation of

ˇ
´
G. Stefanic et al. / Journal of Alloys and Compounds 327 (2001) 151–160
160
Table 4
57
¨
Fe Mossbauer parameters at RT of selected samples prepared in the system ZrO2 –Fe2O3
Sample
Temperature
Lines
d*
D or Eq
HMF
(kOe)
G
A
%
(mm s21
)
(mm s21
)
(mm s21
)
ZF2
5008C
6008C
8008C
Q1
Q2
Q1
Q2
Q1
Q2
0.36
0.41
0.33
0.41
0.33
0.36
0.81
1.51
0.71
1.41
0.75
1.30
0.52
0.52
0.46
0.46
0.45
0.45
71
29
59
41
58
42
ZF3
5008C
6008C
8008C
Q1
Q2
Q1
Q2
Q1
Q2
0.37
0.38
0.37
0.38
0.37
0.27
0.87
0.51
0.87
1.51
1.15
1.08
0.44
0.44
0.44
0.44
0.57
0.57
66
34
66
34
14
31
S
504
0.28
55
d5Isomer shift relative to a-Fe; D and Eq5quadrupole splitting of doublet; HMF5hyperfine magnetic field; G5line-width; A5area under the peaks.
Errors: d, D, Eq560.01 mm s21; HMF562 kOe.
Zr41 ions. The incorporation of 10 mol% of Fe2O3
[11] 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.
[12] M.F. Trubelja, V.S. Stubican, Solid State Ionics 49 (1991) 89.
stabilized the cubic polymorphs of ZrO2 and HfO2, but the
presence of Fe31 ions above their solid solubility limit
caused their transition to monoclinic polymorphs of ZrO2
and HfO2. These results indicated that Fe31 ions have both
a stabilizing and destabilizing influence on the high-tem-
perature polymorphs of ZrO2 and HfO2. Precise determi-
nation of unit-cell parameters indicated that the transition
from a tetragonal to a cubic polymorph of ZrO2 has a very
small influence on their unit-cell volume. The unit-cell
volume of t- or c-ZrO2-type and c-HfO2-type solid solu-
tions decreased linearly with an increase in iron content,
but the rate of decrease was higher in the ZrO2 –Fe2O3
system.
[13] V.V. Kharton, A.A. Yaremchenko, E.N. Naumovich, F.M.B. Mar-
ques, J. Solid State Electrochem. 4 (2000) 243.
[14] C.J. Howard, R.J. Hill, B.E. Reichert, Acta Crystallogr. Sect. B 44
(1988) 116.
[15] P. Li, I.-W. Chen, J.E. Penner-Hahn, J. Am. Ceram. Soc. 77 (1994)
118.
[16] S. Davison, R. Kershaw, K. Dwight, A. Wold, J. Solid State Chem.
73 (1988) 47.
[17] F.J. Berry, M.H. Loretto, M.R. Smith, J. Solid State Chem. 83
(1989) 91.
[18] I.B. Inwang, F. Chyad, I.J. McColm, J. Mater. Chem. 5 (1995)
1209.
ˇ
´
ˇ
´
´
´
[19] S. Popovic, B. Grzeta, G. Stefanic, I. Czako-Nagy, S. Music, J.
Alloys Comp. 241 (1996) 10.
ˇ
´
´
´
[20] G. Stefanic, S. Music, S. Popovic, K. Nomura, J. Mol. Struct.
480/481 (1999) 627.
References
ˇ
´
ˇ
´
´
[21] G. Stefanic, B. Grzeta, S. Music, Mater. Chem. Phys. 65 (2000)
216.
[1] I. Clark, D.H. Reynolds, Ind. Eng. Chem. 29 (1937) 711.
[22] Powder Diffraction File, International Center for Diffraction Data,
Newtown Square, PA, 19073-3273, USA.
ˇ
´
´
´
´
[2] G. Stefanic, S. Music, S. Popovic, K. Furic, Croat. Chem. Acta 69
(1996) 223.
ˇ
´
´
´
[23] G. Stefanic, S. Popovic, S. Music, Thermochim. Acta. 303 (1997)
ˇ
´
´
´
[3] G. Stefanic, S. Popovic, S. Music, Thermochim. Acta 259 (1995)
31.
225.
ˇ
ˇ
´
´
´
[24] G. Stefanic, I.I. Stefanic, S. Music, Mater. Chem. Phys. 65 (2000)
ˇ
´
´
´
[4] G. Stefanic, S. Music, A. Sekulic, Thermochim. Acta 273 (1996)
197.
119.
[25] H. Toraya, M. Yoshimura, S. Somiya, J. Am. Ceram. Soc. 67 (1984)
C119.
[26] S.M. Ho, Mater. Sci. Eng. 54 (1982) 23.
[27] H. Toraya, J. Appl. Cryst. 26 (1993) 583.
[28] H. Toraya, J. Appl. Cryst. 19 (1986) 440.
[29] F. Liebau, Structure Chemistry of Silicates, Springer, Berlin, 1985.
[5] A. Larbot, T. Hours, P. Bergez, J. Charpin, L. Cot, J. Non-Cryst.
Solids 147–148 (1992) 85.
[6] L.-M. Tau, R. Srinivasan, R.J. De Angelis, T. Pinder, B.H. Davis, J.
Mater. Res. 3 (3) (1988) 561.
[7] R. Srinivasan, B.H. Davis, J. Am. Ceram. Soc. 73 (6) (1990) 1780.
[8] P. Duran, C. Pascual, J. Mater. Sci. 19 (1984) 1178.
[9] D.-J. Kim, J. Am. Ceram. Soc. 72 (8) (1989) 1415.
[10] D.-J. Kim, S.-H. Hyun, S.-G. Kim, M. Yashima, J. Am. Ceram. Soc.
77 (2) (1994) 597.
´
´
´
´
´
[30] M. Ristic, S. Popovic, S. Music, I. Czako-Nagy, A. Vertes, M.
Maiorov, A. Petrov, J. Alloys Comp. 256 (1997) 27.