ARTICLE IN PRESS
A. Fossdal et al. / Journal of Solid State Chemistry 177 (2004) 2933–2942
2940
hence kinetics largely govern the decomposition reac-
tions. Due to the similarity of the structures, disordered
intergrowths between the different phases are expected
to form, as seen for the SrO–TiO2 system [18,19].
Synthesis of the RP-phases Sr2FeO4ꢁd and
Sr4Fe3O10ꢁd by the solid-state reaction method is
complicated by low cation diffusion rates below the
decomposition temperatures and the stability of SrCO3
at low temperatures. The latter can to some extent be
circumvented by using Sr(NO3)2 rather than SrCO3.
Sr(NO3)2 will first melt, followed by decomposition and
formation of SrO. The Sr:Fe ratio after annealing was
observed to be lower than the initial ratio when
Sr(NO3)2 was used as a reactant. The proposed
explanation for this discrepancy is evaporation of
Sr(NO3)2. The apparent evaporation of a Sr-containing
compound has also been observed previously [3]. A slow
heating rate was necessary to obtain a controlled
decomposition of the nitrate, but slow heating was also
observed to increase the apparent evaporation.
The synthesis of Sr4Fe3O10ꢁd proved to be most
challenging. The nucleation and growth of Sr4Fe3O10ꢁd
appears to be sensitive to the homogeneity of the
starting powder mixture, as different batches of the
same nominal composition could contain different
phases after being subjected to the same firing sequence.
As an example, firing mixtures of SrCO3 and Fe2O3 in
stoichiometric amounts at 800ꢀC yielded a mixture of
Sr3Fe2O7ꢁd and SrO in one case, whereas another,
similar batch gave close to phase pure Sr4Fe3O10ꢁd
after firing. Attempts to reverse reaction (2) at 700ꢀC
were not successful even after two weeks annealing
time.
4.3. Perovskite
The absence of phase transitions in Sr0.95FeO3ꢁd and
SrFeO3ꢁd materials, calcined only at 750ꢀC, implies that
the oxygen vacancies are disordered even at low
temperatures in these materials. Increasing the calcina-
tion temperature to 800ꢀC causes precipitation of
entropy-stabilized Sr4Fe6O137d in these materials, as
observed by Kleveland et al. [4]. The solid solubility
limit on the Sr-deficient side thus decreases abruptly at
this temperature.
Both the tetragonal to cubic and the orthorhombic to
cubic phase transition temperatures of SrFeO3ꢁd in-
crease with increasing nominal Sr-content in the
perovskite, when comparing materials with the same
thermal history. The differing phase transition tempera-
tures infers a different Sr:Fe ratio in the SrFeO3ꢁd
phase. The tetragonal to cubic phase transition tem-
peratures correspond well with the literature values
[6,11,12]. The orthorhombic to cubic phase transition
temperatures correspond reasonably with the phase
transition temperature of 325ꢀC measured by Takeda
et al. [6], whereas the phase transition temperatures
reported by Haavik et al. [11] and Fournes et al. [12]
are considerably higher. The increase in phase transition
temperature with nominal Sr-content is likely to
be linked with the observation that an increasing
Sr-content (or decreasing Fe-content) in the
perovskite stabilizes a higher average oxidation state
of Fe (Fig. 1).
The apparent solid solubility on either side of the
perovskite composition may be due to either solid
solution, i.e. cation vacancies on the Sr and Fe sites, or
due to disordered intergrowths of the perovskite and the
coexistent phase. On the Fe-deficient side, the
RP-phases Sr4Fe3O10ꢁd (at low temperatures) or
Sr3Fe2O7ꢁd (at high temperatures) can be considered
as ordered intergrowths between SrO of rock salt
structure and perovskite. It is therefore likely that
disordered intergrowths between the RP-phases and
the perovskite can occur. On the Sr-deficient
side, the perovskite phase is coexistent with the
magnetoplumbite-type phase SrFe12O19 below 775ꢀC.
The magnetoplumbite structure is not perovskite-related
[23], hence intergrowth between SrFeO3ꢁd and
SrFe12O19 is not expected. Above 775ꢀC, the perovs-
kite-related [2] Sr4Fe6O137d phase is stabilized.
Bredesen et al. [24] have documented nano-scale inter-
The high temperatures in the combustion step [20] of
the glycine-nitrate synthesis were seen to cause the
formation of a non-equilibrium mixture of binary and
ternary oxides. In case of Sr3Fe2O7ꢁd, a calcination
temperature at or above 900ꢀC was sufficient to produce
a single-phase material, whereas single-phase Sr2FeO4ꢁd
and Sr4Fe3O10ꢁd could not be obtained by this method
due to the slow reversal rate of reactions (1) and (2) at
ambient pressures. Reversal of reaction (1) has however
previously been obtained by Dann et al. [6] at 750ꢀC
under an oxygen pressure of 200 atm.
4.2. Sr4Fe6O137d
The phase Sr4Fe6O13 is entropy stabilized relative to
SrFe12O19 and Sr1ꢁxFeO3ꢁd, as inferred from reaction (3).
The reaction is rationalized by the stability of tetravalent
Fe in Sr1ꢁxFeO3ꢁd at low temperatures. An endothermic
enthalpy for reaction (3) is expected due to the high
content of tetravalent iron in Sr1ꢁxFeO3ꢁd at low
temperature and only a moderate content of tetravalent
in metastable Sr4Fe6O13+d at low temperature [21,22].
growths of La1ꢁxSrxFeO3ꢁd in La-doped Sr4Fe6O137d
,
which suggests that intergrowth of SrFeO3ꢁd
and Sr4Fe6O137d is possible. The oxygen vacancy
ordering observed in both Sr- and Fe-deficient perovs-
kites infers ordering of trivalent and tetravalent Fe. We
therefore interpret the observed variations in Sr:Fe
ratio to reflect solid solution rather than disordered
intergrowths.