Reactions within the system, 2SrCO3-Fe2O3: Part I. CO2 atmosphere

Article abstract of DOI:10.1007/s10973-005-0639-0

Simultaneous TG/DSC and high temperature X-ray diffraction studies were performed on the system SrCO₃-Fe₂O₃ in an atmosphere of CO₂ from room temperature to 1300°C. SrCO ₃ decomposes and reacts simultaneously with the Fe₂O ₃ beginning around 850°C. A transient iron rich phase is formed initially, which soon diminishes to yield a mixture of perovskite SrFeO _3-x and the iron rich phase Sr₄Fe₆O ₁₃. Upon cooling the perovskite phase predominantly orders into the brownmillerite structure Sr₂Fe₂O₅. There does not appear to be a Hedvall effect associated with the first order phase transformation in SrCO₃ at 927°C.

Full text of DOI:10.1007/s10973-005-0639-0

Journal of Thermal Analysis and Calorimetry, Vol. 80 (2005) 217–223
REACTIONS WITHIN THE SYSTEM, 2SrCO3–Fe2O3
Part I. CO2 atmosphere
P. K. Gallagher1*, J. P. Sanders2, P. M. Woodward 3 and I. N. Lokuhewa3
1Emeritus Prof., Depts. of Chemistry and MS&E, The Ohio State University, Columbus, OH, 43210, USA
Adjunct Prof. Depts. of Chemistry and Ceramic and Materials Eng., Clemson University Clemson, SC, 29632, USA
2National Brick Research Center, 100 Clemson Research Park., Anderson, SC, 29625, USA
Adjunct Prof. School of Ceramic and Materials Eng., Clemson University
3Dept. of Chemistry, The Ohio State University, Columbus, OH, 43210, USA
Simultaneous TG/DSC and high temperature X-ray diffraction studies were performed on the system SrCO3–Fe2O3 in an
atmosphere of CO2 from room temperature to 1300°C. SrCO3 decomposes and reacts simultaneously with the Fe2O3 beginning
around 850°C. A transient iron rich phase is formed initially, which soon diminishes to yield a mixture of perovskite SrFeO3–x and
the iron rich phase Sr4Fe6O13. Upon cooling the perovskite phase predominantly orders into the brownmillerite structure Sr2Fe2O5.
There does not appear to be a Hedvall effect associated with the first order phase transformation in SrCO3 at 927°C.
Keywords: simultaneous TG/DSC, solid-state reactivity, X-ray diffraction
Introduction
been shown that for samples that are slow cooled the de-
fect concentration at room temperature is 0.09≥x≥0.12
when prepared in pure oxygen, 0.20≥x≥0.16 when pre-
pared in air, and 0.43 when prepared in an atmosphere
The system, 2SrCO3–Fe2O3 offers several interesting
opportunities to study solid-state reactivity.
with pO =0.003 atm [6]. The brownmillerite structure
• There is a wide range of stoichiometry and structural
modifications available for the products of the thermal
decomposition because of the variability in the valence
of iron and the tendency for the perovskite structure to
tolerate large concentrations of vacancies and various
ordered patterns of these defects [1].
2
can be derived from the perovskite structure by the or-
dering of oxygen vacancies in a pattern that leads to dis-
torted tetrahedral coordination for 50% of the iron sites
and distorted octahedral coordination for the remaining
iron sites [7]. As the temperature is increased, the condi-
tions become more reducing and further oxygen vacan-
cies are created, i.e., the value of x increases. As the
value of x approaches 0.5 and the oxidation state of iron
approaches +3, the perovskite structure is no longer sta-
ble and the brownmillerite structure forms as shown in
Eq. (2). Temperature also affects the competition be-
tween brownmillerite and perovskite in a direct manner.
If brownmillerite, Sr2Fe2O5, is heated in an inert atmo-
sphere so that iron remains in the +3 oxidation state over
the entire temperature range, the oxygen ion mobility in-
creases with temperature. Eventually the oxygen vacan-
cies become disordered and orthorhombic Sr2Fe2O5
brownmillerite transforms to cubic SrFeO2.5 defect
perovskite. Campbell and Schmidt have shown this to
occur at 875°C in an inert atmosphere [8]. In principle,
in the absence of available O2 the brownmillerite phase
can be formed directly from SrCO3 and Fe2O3 (Eq. (3)).
• The decomposition of SrCO3 may be catalyzed by
the presence of Fe2O3 [2].
• The decomposition of SrCO3 is reversible and
depends upon the partial pressure of CO2 [3].
• There is the possibility of a ‘Hedvall effect’, because
of the crystallographic transformation in SrCO3
within the temperature range of the reactions [4].
Consequently, the nature and rates of the reac-
tions are investigated as a function of the partial pres-
sures of O2 and CO2 using simultaneous TG/DSC and
variable temperature X-ray powder diffraction.
When the Sr/Fe ratio is unity the most important
ternary phases are the cubic perovskite phase, SrFeO3–x,
and the orthorhombic brownmillerite phase, Sr2Fe2O5.
In the presence of significant amounts of O2 the
perovskite phase SrFeO3–x is formed as shown in
Eq. (1). In the absence of oxygen vacancies (x=0) the
perovskite structure stabilizes the unusual oxidation
state of +4 for iron [5]. However, high partial pressures
of O2 are needed to prepare defect free SrFeO3. It has
2SrCO3+Fe2O3+(0.5–x)O2↔2SrFeO3–x+2CO2 (1)
2SrFeO3–x↔Sr2Fe2O5+(0.5–x)O2
(2)
*
Author for correspondence: p_gallag@bellsouth.net
1388–6150/$20.00
Akadémiai Kiadó, Budapest, Hungary
Springer, Dordrecht, The Netherlands
© 2005 Akadémiai Kiadó, Budapest

GALLAGHER et al.
Results and discussion
2SrCO3+Fe2O3↔Sr2Fe2O5+2CO2
(3)
Part I of this study concentrates on the decompo-
sition and subsequent reactions in a CO2 atmosphere.
Part II will deal with the effects of oxidation by com-
paring the results in O2 and Ar atmospheres.
The effect of CO2 on the reversible decomposition is in-
dicated in Fig. 1 which presents TG curves, collected at
8 K min–1, for the reactions in O2, Ar and CO2. Clearly
the reversible decomposition and simultaneous reaction
of SrCO3 is shifted about 200°C to higher temperatures
in CO2. This has the effect of essentially separating the
phase transformation in SrCO3, around 927°C, from the
subsequent reactions. The decomposition and simulta-
neous reactions, however, are clearly influenced by
higher temperature transitions of the reactants and/or
products in the case of CO2. The onset of the decompo-
sition for the mixture is at a slightly lower temperature
than for pure SrCO3 itself. The overall mass loss in O2 is
slightly less by virtue of its greater oxidation strength
and consequently higher average valence for the iron
species. The theoretical mass loss for Eq. (1) with x=0 is
15.83%, for Eq. (2) is 3.52%, and for Eq. (3) is 19.35%.
Experimental
The sources of SrCO3 and Fe2O3 were Aldrich
(99.995% purity) and Alfa Aesar (Puratronic Grade,
99.998% purity), respectively. The sample was pre-
pared by thoroughly mixing the two materials in the
molar proportion of 2SrCO3 with Fe2O3, i.e. Sr/Fe
equals unity, in methanol using a mortar and pestle.
After mixing, the methanol was ignited and the resul-
tant mixture was dried at 110°C for 24 h and then
stored in a desiccator prior to use.
Simultaneous
thermogravimetric/differential
scanning calorimetry (TG/DSC) measurements were
performed on a Netzsch 449C simultaneous thermal
analyzer using open alumina crucibles. The nominal
sample size used in this work was 20.3 mg ( 0.3 mg).
A flowing atmosphere at 100 mL min–1 was used for all
measurements. The three atmospheres used were Ar,
O2, and CO2. Nominal heating rates of 1, 2, 4, 8, 16
and 32 K min–1 were used in this study. Baselines and
sensitivity calibrations were obtained for each heating
rate. An empty alumina crucible was used as the refer-
ence for the DSC measurements. The sensitivity cali-
bration calculations for the DSC measurements were
made using measurements on sapphire disks.
Data analysis was performed using Netzsch’s
Proteus and Thermokinetic 2 software packages [9].
In the least squares fitting of the TG data no weighing
factors were applied. For the TG data set, single step
linear regression was performed as well as multi-step
non-linear regression analysis using consecutive
reactions. Kinetic analysis was not performed on the
DSC curves because of the complications due to the
multiple phase transformations.
Variable temperature X-ray powder diffraction
(XRPD) data sets were collected with a Bruker D8
diffractometer. The diffractometer is equipped with an
incident beam Ge 111 monochromator that selects only
CuKα1 radiation, a Braun linear position sensitive detec-
tor, and an Anton Parr HTK 1200 high temperature
stage. The HTK 1200 temperature stage employs ambi-
ent heating to minimize temperature gradients. At each
temperature the data collection time was 45–60 min be-
fore ramping to the next temperature and holding for the
next run. The data were collected in flowing atmo-
spheres of Ar, O2 and CO2. The data were analyzed us-
ing the Rietveld method [10], as implemented in the
TOPAS software package [11].
Fig. 1 TG data for 2SrCO3⋅Fe2O3 mixture in O2, Ar and CO2
at 8 K min–1
DSC results of the simultaneous TG/DSC experi-
ments in CO2 at various heating rates are presented in
Figs 2 and 3. Figure 2 shows the DSC curves during
heating and those during cooling are shown in Fig. 3.
The TG curves are seen later in the discussion of the ki-
netic analyses. Table 1 summarizes the extensive X-ray
diffraction study indicating the mass fractions of each
solid phase as a function of temperature, and the overall
Sr:Fe ratio that would result from such a mixture. As no
constraints were placed on the phase fractions during
analysis of the X-ray data the deviation of the Sr:Fe ratio
from unity acts as an independent indicator of the accu-
racy of the results. The differences between the time–
temperature programs applied to the thermoanalytical
and X-ray diffraction experiments preclude a direct
comparison. Nevertheless, valuable insights can be ob-
tained through a judicious comparison between the two
sets of results.
The general nature of the overall sequence of reac-
tions is described using the intermediate heating rate
of 8 K min–1 as an example. Figures 4 and 5 show ex-
218
J. Therm. Anal. Cal., 80, 2005

2SrCO3–Fe2O3
Table 1 Change in mass% on cooling and heating as determined by Rietveld analysis of the X-ray powder diffraction dataa
SrCO3
R3m
Fe2O3
R3c
SrFeO3–x Sr4Fe6O13
Icc21
SrFe12O19
P63/mmc
SrCO3
Pmcn
Sr4Fe2.6O8.2(CO3)0.4
14/mmm
Sr2Fe2O5 Mole ratio
Ibm2
T/°C
Sr/Fe
Pm3m
25
800
66(3)
66(1)
63(2)
45(1)
34(2)
34(1)
14(1)
8(0.1)
3(0.5)
1.03
1.04
1.13
0.97
0.93
0.93
0.91
0.88
0.94
0.87
0.84
0.86
0.97
0.91
0.91
0.90
0.90
0.89
0.89
0.88
0.89
0.89
0.88
0.91
0.91
0.95
0.95
0.98
0.95
0.97
0.97
0.98
1.00
0.99
850
5(0.5)
17(1)
18(1)
24(1)
23(1)
21(1)
18(1)
11(1)
14(1)
10(1)
7(0.9)
900
7(0.4)
43(1)
38(1)
34(1)
28(1)
24(1)
19(1)
15(1)
12(1)
9(0.5)
5(0.4)
4(0.6)
3(0.5)
3(0.3)
1(0.4)
2(0.3)
1(0.4)
1(0.4)
1(0.4)
2(0.4)
5(0.5)
5(0.5)
2(0.5)
3(0.4)
8(0.4)
13(1)
20(1)
25(1)
32(1)
32(1)
38(1)
42(1)
46(1)
45(1)
47(2)
48(1)
49(2)
52(2)
52(2)
55(2)
54(2)
54(2)
60(2)
60(2)
61(2)
58(2)
59(2)
59(2)
46(2)
37(1)
32(1)
28(1)
21(1)
16(1)
18(1)
22(1)
25(1)
25(1)
30(1)
33(1)
34(1)
37(1)
40(1)
40(1)
43(1)
42(1)
41(1)
40(1)
38(1)
39(1)
38(1)
38(1)
37(1)
37(1)
36(1)
36(1)
34(1)
32(1)
32(1)
31(1)
30(1)
29(1)
28(1)
28(1)
25(1)
925
950
975
1000
1025
1050
1075
1100
1125
1150
1150
1150
1150
1150
1150
1100
1050
1000
950
6(1)
8(1)
7(1)
8(1)
8(1)
8(1)
7(1)
6(1)
7(1)
8(1)
900
850
800
5(0.8)
9(0.08)
9(0.8)
6(0.6)
8(0.7)
7(0.7)
8(0.7)
750
700
650
17(1)
25(1)
32(1)
37(1)
32(1)
51(2)
600
550
500
300 10(1)
25 8(0.8)
aEstimated standard deviations are enclosed in parentheses.
panded sections of the DTG and DSC curves for the
heating and cooling portions, respectively. At about
838°C there are small irregularities noted during heating
in the DTG and DSC curves. These are believed to be
associated with the very initial onset of the reaction or
with an as yet unknown event occurring in SrCO3 which
has been observed previously and is the subject of fur-
ther work [3]. The first order orthorhombic to rhombo-
hedral phase transformation is clearly evident as a rela-
tively sharp endothermic peak in the DSC curve having
an extrapolated onset temperature of 927.1°C at this
heating rate. The X-ray results in Table 1 confirm the
transition. There is no detectable change in either the
TG or DTG curves, ruling out any significant enhance-
ment of the rate of the reaction as a result of the phase
transformation, i.e., ‘Hedvall effect’.
The XRPD results shown in Table 1 provide
valuable information about the details of this reaction.
At 850°C ternary reaction products begin to appear.
This corresponds rather closely with the first sign of
J. Therm. Anal. Cal., 80, 2005
219

GALLAGHER et al.
mass loss in the TG data and the small features in the
DTG and DSC curves noted above. The simultaneous
decomposition of SrCO3 and reaction with Fe2O3 is
substantiated by the failure to detect SrO in the X-ray
patterns. This is in contrast to XRPD patterns col-
lected in argon where SrO formation was clearly seen,
as will be discussed in detail in part II. The begin-
nings of the transformation of SrCO3 from ortho-
rhombic to rhombohedral is first seen in the XRPD
data set collected at 900°C and the transformation ap-
pears complete at 925°C in reasonably good agree-
ment with the DSC analysis.
The first ternary phases to form are the Fe-rich
phases SrFe12O19 and Sr4Fe6O13. The former compound
has the magnetoplumbite structure with hexagonal sym-
metry and structurally related to Fe2O3, while the latter
phase has a complex orthorhombic structure that is re-
lated to perovskite. The concentration of SrFe12O19 rises
sharply as the SrCO3 and Fe2O3 begin to react and then
decreases with further heating, disappearing upon heat-
ing above 1100°C. The transient appearance of this
phase is most likely a result of diffusion limited interfa-
cial reactions. The concentration of Sr4Fe6O13 climbs to
a mass fraction of roughly 40% and then decreases very
slowly upon cooling, finally coming to a value of
25 mass% upon cooling all the way back to room tem-
perature. Apparently once this phase forms its decompo-
sition is very sluggish. Thus it acts as a kinetic trap that
must be overcome in order to prepare either cubic
perovskite or brownmillerite. Significant concentrations
of this phase were also observed in air and argon atmo-
spheres, but in air the concentration of this phase de-
creased more rapidly upon cooling. Apparently
Sr4Fe6O13 does not easily accommodate additional oxy-
gen that must accompany oxidation of iron, and condi-
tions that favor Fe4+ are needed in order to stabilize the
cubic perovskite phase in preference to Sr4Fe6O13. A re-
cent study of phase equilibria in the SrO–Fe2O3 system
shows that in the temperature range 800–1200°C an
equilibrium mixture of SrFeO3 and Sr4Fe6O13 is ex-
pected when the Fe:Sr ratio is between 1.0 and 1.5 [12].
Therefore, given the presence of unreacted SrCO3 and
the Sr-rich oxycarbonate phase Sr4Fe2.6O8.2(CO3)0.4 the
presence of Sr4Fe6O13 is to be expected. The existence
of both of these phases is based on the appearance a few
weaks peaks in the X-ray patterns. Consequently, it is
difficult to accurately determine their phase fractions.
The presence of an oxycarbonate phase,
Sr4Fe2.6O8.2(CO3)0.4, at high temperature was not antic-
ipated in advance, nor could it have been anticipated
from the thermal analysis data. The structure of this
phase is closely related to the Ruddlesden–Pop-
per (RP) series of compounds, in particular Sr4Fe3O10
[13]. In order to confirm and better understand the for-
mation of this phase the RP phase Sr3Fe2O7–x was pre-
Fig. 2 DSC curves for 2SrCO3⋅Fe2O3 mixture on heating in CO2
Fig. 3 DSC curves for 2SrCO3⋅Fe2O3 mixture on cooling in CO2
Fig. 4 TG, DTG and DSC curves for 2SrCO3⋅Fe2O3 mixture
on heating in CO2 at 8 K min–1
Fig. 5 TG, DTG and DSC curves for 2SrCO3⋅Fe2O3 mixture
on cooling in CO2 at 8 K min–1
220
J. Therm. Anal. Cal., 80, 2005

2SrCO3–Fe2O3
pared and its stability was studied in a CO2 atmosphere
quite close to the value of 3.894  reported for
SrFeO2.81 at 300°C [14]. From this observation it is
concluded that some oxidation occurred in the high
temperature X-ray stage that prevented complete
transformation to brownmillerite upon cooling. We
suspect adventitious oxygen that was not completely
flushed from the system.
by variable temperature X-ray diffraction. Upon heat-
ing above 700°C the RP starts to decompose to form
SrCO3 and SrFeO3–x. Upon heating above 1100°C the
oxycarbonate phase forms. This phase appears only to
be stable at high temperatures because on cooling be-
low 1000°C it decomposes to form SrCO3 and
SrFeO3–x. These temperatures are in good agreement
with the temperature range where the oxycarbonate
phase is observed in Table 1. From this behavior we
can also infer that the increase in the quantity of SrCO3
that is seen in Table 1 upon cooling can be attributed to
the decomposition of the oxycarbonate phase. The TG
data suggests that if the X-ray furnace could access
higher temperatures both the oxycarbonate and the
SrCO3 phases would have decomposed and reacted
with the iron rich Sr4Fe6O13 phase to form pure
SrFeO3–x perovskite.
At the highest temperatures the dominant phases
are Sr4Fe6O13 and the cubic perovskite phase,
SrFeO3–x. While it may initially seem surprising to
form cubic perovskite instead of Sr2Fe2O5 brown-
millerite at high temperature in a CO2 atmosphere,
bear in mind that the cubic perovskite structure can
support considerable oxygen non-stoichiometry at
high temperature. As stated in the introduction pure
Sr2Fe2O5 has been shown to undergo a phase transi-
tion from the orthorhombic brownmillerite structure
to an oxygen deficient cubic perovskite structure
upon heating above 875°C in an argon atmosphere
[8]. At high temperature in a CO2 atmosphere the
oxidation state of iron should be close to +3, and
order–disorder transition should occur at a similar
temperature. Schmidt and Campbell found the lattice
parameter of the disordered cubic form of Sr2Fe2O5
(SrFeO2.5) to vary from 3.976  at 875°C to 3.982 
at 950°C in argon [8, 14]. They also studied SrFeO2.81
upon heating in air and found that at 1000°C the cubic
perovskite phase had a composition of SrFeO2.56 and a
unit cell parameter of 3.985 . By comparison the
unit cell parameter for the cubic perovskite phase in
this experiment was determined to be 3.9792(3) 
at 950°C upon cooling. Therefore, it can be concluded
that the stoichiometry of the high temperature cubic
perovskite phase seen in the XRPD patterns is fairly
close to SrFeO2.5. The pre-dominance of Fe in the +3
oxidation state in both the cubic perovskite and the
Sr4Fe6O13 phases is consistent with the TG results.
Upon cooling below 850°C the cubic perovskite
phase starts transforming back into the orthorhombic
brownmillerite phase as the oxygen vacancies order.
Surprisingly, roughly 16% of the cubic perovskite
phase was retained upon cooling back to room
temperature. The lattice parameter of the cubic phase
that remains at room temperature, 3.8948(9) , is
The decomposition and reactions proceed until
completion around 1280°C. There are four general
areas of enhanced rate particularly evident in the
DTG curve peaking at 1160, 1196, 1253 and 1259°C.
There are obvious endotherms in the DSC curve
associated with them. Unfortunately all are right at or
above the upper temperature limit of the X-ray
diffraction study. These are attributed to unknown
solid1, solid2 or solid–liquid phase transformations.
The thermoanalytical curves on cooling, Fig. 5
show less features, as would be expected. There is a
prominent sharp exotherm, presumably resulting from
solidification of the highest temperature melt. There is
a change of slope, and perhaps a small peak (depend-
ing on one’s choice of baseline, around 750°C consis-
tent with oxygen vacancy ordering in Sr2Fe2O5 ob-
served in the X-ray study. The TG results show no sig-
nificant gain in mass upon cooling in CO2 implying no
re-carbonation during cooling.
Earlier work had indicated superior results were
achieved using the TG rather than the DSC data for
kinetic analysis of simultaneous TG/DSC results [15].
In addition, the multiple phase transformations
obvious in Fig. 3 would invalidate such a DSC based
analysis. The TG data were analyzed using two
approaches. Only the data below a fraction reacted, α,
of 0.7 was used because of the irregularities imposed
by the melting at high temperatures.
The traditional kinetic analyses of thermo-
analytical data has justifiably been the subject of much
criticism in recent times, e.g., the extensive review by
Galwey [16]. The analyses herein are an attempt to ac-
cess the utilization of conventional commercial pro-
gramming to gain insights into an obviously very com-
plex reaction system. Recognizing that such analyses
represent a compromise between achieving the best fit
and minimizing the number of variable parameters, a
single step mechanism was used initially [15]. Model-
free kinetic analysis using the Ozawa–Flynn analysis
was applied first. Figure 6 summarizes the results. In
the region 0.2<α<0.6 the Arrhenius parameters are rea-
sonably constant. A value of about 500 kJ mol–1 is ob-
tained for the apparent activation energy, E, and
about 16 for the log of the corresponding pre-exponen-
tial term, logA. At either extreme of α, the values rise
abruptly and, as usual, an apparent correlation effect is
observed between E and logA.
J. Therm. Anal. Cal., 80, 2005
221

GALLAGHER et al.
Table 2 Kinetic parameters for single step kinetic modeling
Model type
Avrami–Erofeev
fit range/α
logA/s–1
0.1–0.7
15
E/kJ mol–1
reaction dimension
R2
478
0.27
0.996
Table 3 Kinetic parameters for multi-step kinetic modeling
First step
Second step
Model type
nth order
Avrami–Erofeev
Fig. 6 Ozawa–Flynn model-free analysis of TG data for
2SrCO3⋅Fe2O3 mixture (heating in CO2), α=0.1 to 0.7
fit range/α
0.1–0.7
logA/s–1
56
1276
21
14
E/kJ mol–1
444
reaction order
reaction dimension
following reaction
R2
0.32
0.16
0.999
The two-step process invoking a change in mech-
anism after about 16% of the decomposition correlates
well with the observation of the transient phase,
SrFe12O19. It is proposed that the initial rate corre-
sponds to the formation of the transient phase and then
transforms to that primarily associated with the forma-
tion of the final products. An alternative explanation
invokes the first order phase transformation of SrCO3..
The change in mechanism appears very close to the
transition temperature. This suggests that, although no
momentary enhancement of the rate is associated with
the transition, there may be a significant difference in
the rate and/or mechanism occurring between the two
crystallographic phases.
Fig. 7 Single step data fit of TG data for 2SrCO3⋅Fe2O3 mix-
ture (heating in CO2), α=0.1 to 0.7
Allowing the kinetics software package to
achieve the best fit by single step linear regression
produced the result summarized in Table 2. Figure 7
indicates the fit to the experimental data. The Arr-
henius parameters from the Avrami–Erofeev mecha-
nism are in fair agreement with those suggested by the
Ozawa–Flynn analysis.
In Figs 1 and 4, however, it is obvious the first
several mass percent of the reaction appear to have a
significantly different reaction rate. Consequently, it
is reasonable to introduce a two-step process into the
analysis. The optimum results using a two-step
consecutive reaction mechanism are summarized in
Table 3. The much improved fit is illustrated in Fig. 8.
The Avrami–Erofeev is still the mechanism of choice
for the major portion of the reaction, above α=0.16
and the Arrhenius parameters are very similar to those
for the single-step process. An nth order reaction,
however, is selected for the first portion of the
reaction. The Arrhenius parameters derived and the
order of this step are unreasonably high, however.
Fig. 8 Multi-step data fit of TG data for 2SrCO3⋅Fe2O3 mix-
ture (heating in CO2), α=0.1 to 0.7
222
J. Therm. Anal. Cal., 80, 2005

2SrCO3–Fe2O3
Conclusions
the underlying assumptions inherent in traditional
kinetic analysis and the complexity of the reactions.
• A two-step mechanism could be supported either
by the formation of the transient iron rich interme-
diate at the heating rates used or by a difference in
reactivity of the two crystallographic forms of the
reactant SrCO3.
• Because of the reversible nature of the thermal
decomposition of SrCO3, the decomposition and
simultaneous reaction with Fe2O3 does not begin
until approximately 850°C.
• Initially the iron rich phases, SrFe12O19 and Sr4Fe6O13
are formed followed closely by the cubic perovskite
SrFeO3–x. While SrFe12O19 is a transient phase that
decomposes on further heating, Sr4Fe6O13 does not References
decompose easily once it is formed. Above 1100°C
1 P. K. Gallagher, in Application of Mössbauer Spectroscopy,
R. L. Cohen (Ed.), Academic Press, New York 1976,
Chap. 7 and references therein.
the oxycarbonate phase Sr4Fe2.6O8.2(CO3)0.4 forms.
There are indications of melting and further reaction
in the DSC curves above about 1150°C.
2 P. K. Gallagher and D. W. Johnson, Jr., J. Am. Cer. Soc.,
59 (1976) 171.
• Upon cooling below 1000°C the oxycarbonate
phase decomposes to form SrFeO3–x and SrCO3.
Upon cooling somewhere in the vicinity of
700–750°C the oxygen vacancies in the cubic
perovskite phase (SrFeO2.5) order and a transforma-
tion to the orthorhombic brownmillerite phase
(Sr2Fe2O5) is observed.
• There does not appear an obvious enhancement in
the reaction rate, Hedvall effect, associated with
first order phase transformation in SrCO3 at 927°C,
even at the slowest heating rates.
• Utilizing traditional software, the reaction appears
best modeled using TG data with a two-step process
involving an initial nth order reaction followed by
an Avrami–Erofeev type reaction after an α of
about 0.16. Only data below α=0.7 were fit because
the higher temperature reactions markedly influ-
enced the rate of mass loss. DSC data were not ana-
lyzed kinetically because of these complications.
Little faith, however, can be placed in the direct im-
plications of these simplified analyses. This in no
way reflects on the software used but is the result of
3 S. A. Robbins, R. G. Rupard, B. J. Weddle, T. R. Maull and
P. K. Gallagher, Thermochim. Acta, 269/270 (1995) 43.
4 J. A. Hedvall, Chem. Rev., 15 (1934) 139.
5 P. K. Gallagher, J. B. MacChesney and D. N. E. Buchanan,
J. Chem. Phys., 41 (1964) 2429.
6 M. Schmidt, J. Phys. Chem. Solids, 61 (2000) 1363.
7 R. H. Mitchell, ‘Perovskites: Modern and Ancient’
Almaz Press, Thunder Bay, Ontario, Canada 2002.
8 M. Schmidt and S. J. Cambell, J. Solid State Chem.,
156 (2001) 292.
9 J. Opffermann, J. Therm. Anal. Cal., 60 (2000) 641.
10 R. A. Young, ‘The Rietveld Method’ Oxford University
Press, London 1995, Chap. 1.
11 R. W. Cheary and A. A. Coelho, J. Appl. Cryst.,
25 (1992) 109.
12 A. Fossdal, M. A. Einarsrud and R. Grande,
J. Solid State Chem., 177 (2004) 2933.
13 Y. Bréard, C. Michel, M. Hervieu and B. Raveau,
J. Mater. Chem., 10 (2000) 1043.
14 M. Schmidt and S. J. Cambell, J. Phys. Chem. Solids,
63 (2002) 2085.
15 J. P. Sanders and P. K. Gallagher, Thermochim. Acta,
388 (2002) 115.
16 A. K. Galwey, Thermochim. Acta, 423 (2004) 139.
J. Therm. Anal. Cal., 80, 2005
223