An investigation of the Fe and Mo oxidation states in Sr2Fe 2-xMoxO6 (0.25 ≤ x ≤ 1.0) double perovskites by X-ray absorption spectroscopy

Article abstract of DOI:10.1016/j.jallcom.2012.05.056

Sr₂FeMoO₆ double perovskite systems are widely studied because of their interesting and technologically relevant physical properties. Sr₂FeMoO₆ is just a single composition in the Sr₂Fe_2-xMo_xO₆ solid-solution, and it is important to understand how the composition impacts the transition-metal valence states. Variations in the lattice parameters of these materials were studied using powder X-ray diffraction and it was found that a large change in the lattice constant occurs between Sr₂Fe _1.50Mo_0.50O₆ and Sr₂Fe _1.35Mo_0.65O₆ that likely coincides with a transition from a cubic to a tetragonal unit cell, in agreement with previous studies. Fe K- and Mo K-edge X-ray absorption near-edge spectra were also collected to investigate how the oxidation state and coordination environment change with composition. When the Mo content is low, Fe adopts a 3+ oxidation state and Mo adopts a 6+ oxidation state. As the Mo content is increased, the Fe and Mo cations are both partially reduced, resulting in a mixture of Fe ³⁺ and Fe²⁺ and Mo⁵⁺ and Mo⁶⁺. The reduction of the metal centers apparently drives the change in unit cell. The influence of preparation method on the oxidation states of Fe and Mo was also investigated by annealing the materials under vacuum. The results reported here show that the oxidation states of Fe and Mo are strongly impacted by both composition and preparation method, which may account for the wide variety of oxidation state and magnetic properties that have been reported previously.

Full text of DOI:10.1016/j.jallcom.2012.05.056

Journal of Alloys and Compounds 537 (2012) 323–331
Contents lists available at SciVerse ScienceDirect
Journal of Alloys and Compounds
journal homepage: www.elsevier.com/locate/jalcom
An investigation of the Fe and Mo oxidation states in Sr2Fe2ꢀxMoxO6
(0.25 6 x 6 1.0) double perovskites by X-ray absorption spectroscopy
⇑
John R. Hayes, Andrew P. Grosvenor
Department of Chemistry, University of Saskatchewan, Saskatoon, SK, Canada S7N 5C9
a r t i c l e i n f o
a b s t r a c t
Article history:
Sr2FeMoO6 double perovskite systems are widely studied because of their interesting and technologically
relevant physical properties. Sr2FeMoO6 is just a single composition in the Sr2Fe2ꢀxMoxO6 solid-solution,
and it is important to understand how the composition impacts the transition-metal valence states. Vari-
ations in the lattice parameters of these materials were studied using powder X-ray diffraction and it was
found that a large change in the lattice constant occurs between Sr2Fe1.50Mo0.50O6 and Sr2Fe1.35Mo0.65O6
that likely coincides with a transition from a cubic to a tetragonal unit cell, in agreement with previous
studies. Fe K- and Mo K-edge X-ray absorption near-edge spectra were also collected to investigate how
the oxidation state and coordination environment change with composition. When the Mo content is low,
Fe adopts a 3+ oxidation state and Mo adopts a 6+ oxidation state. As the Mo content is increased, the Fe
and Mo cations are both partially reduced, resulting in a mixture of Fe3+ and Fe2+ and Mo5+ and Mo6+. The
reduction of the metal centers apparently drives the change in unit cell. The influence of preparation
method on the oxidation states of Fe and Mo was also investigated by annealing the materials under vac-
uum. The results reported here show that the oxidation states of Fe and Mo are strongly impacted by both
composition and preparation method, which may account for the wide variety of oxidation state and
magnetic properties that have been reported previously.
Received 14 February 2012
Accepted 15 May 2012
Available online 24 May 2012
Keywords:
XANES
Order-disorder effects
Electronic structure
Data storage materials
Synchrotron radiation
Oxide materials
Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction
from a cubic to a tetragonal lattice with variation in composition is
the Sr2Fe2ꢀxMoxO6 double perovskite solid-solution. While X-ray
Double perovskite compounds, having the general formula
A2BB0O6 (A = Alkaline-earth, B/B0 = transition metal), are well
known to exhibit a variety of interesting and technologically rele-
vant properties [1–4]. The basic double perovskite structure is an
extension of the cubic perovskite structure (ABO3) in which cor-
ner-sharing BO6 and B0O6 octahedra alternate, resulting in long-
range ordering (Fm3m; Fig. 1). When the size difference between
the B and B0 cations is large, the structure may become tetragonally
distorted, resulting in two distinct sites for the B and B0 ions (here,
B represents the larger cation) [6,7]. This distortion results in the
formation of one of two possible crystal lattices; one described
by the I4/mmm space group and one described by the I4/m space
group [8,9]. In the I4/mmm lattice, the B0 octahedra undergo com-
pression along the c-axis while the B octahedra are elongated along
the c-axis. In the I4/m unit cell, the B and B0 octahedra undergo
opposing rotations, causing the octahedra to buckle, which results
in expanded BO6 octahedra and compressed B0O6 octahedra [8,9].
The transition from the Fm3m structure to the I4/m structure is
shown in Fig. 1. An example of a system that undergoes a transition
diffraction studies have not conclusively shown whether SrFeMoO6
adopts the I4/m or I4/mmm space group, the results of these dis-
tortions are similar, and the I4/m structure is assumed in this study
[8,10,11]. This assumption is supported by a previous study which
used group theory to suggest that I4/m symmetry must be adopted
[12]. (This system is known to be oxygen-deficient, but it will be
referred to here as Sr2Fe2ꢀxMoxO6 for simplicity [13,14]).
Sr2FeMoO6 has been shown to exhibit a variety of interesting
magnetic properties [13,15–17]. For example, these materials are
ferrimagnetic, with a Curie temperature reported to be between
410 and 450 K [9]. Here, ferrimagnetism arises from a double ex-
change mechanism, which results from the ferromagnetic align-
ment of Fe 3d electrons and the antiferromagnetic alignment of
Mo 4d electrons [15,18,19]. This model assumes Fe adopts a 3+ oxi-
dation state and Mo adopts a 5+ oxidation state, and the theoretical
maximum saturation magnetization is 4 lB, though experimen-
tally, the saturation magnetization has always been observed to
be less [20–22]. This reduced experimental value has been linked
to the amount of antisite disorder present, in which Fe occupies
a Mo site (and vice versa), leading to antiferromagnetic coupling
between adjacent Fe ions [14,23–27]. In addition to being
ferrimagnetic, Sr2FeMoO6 is known to exhibit magnetoresistive
⇑
Corresponding author. Tel.: +1 306 966 4660; fax: +1 306 966 4730.
E-mail address: andrew.grosvenor@usask.ca (A.P. Grosvenor).
0925-8388/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jallcom.2012.05.056

324
J.R. Hayes, A.P. Grosvenor / Journal of Alloys and Compounds 537 (2012) 323–331
Fig. 1. The Sr2Fe2ꢀxMoxO6 compounds adopt one of two crystal structures. At low values of x, the undistorted, ordered double perovskite structure is adopted (space group
Fm3m, left,), consisting of alternating Fe (gray) and Mo (dark gray) octahedra with Sr atoms (light gray) occupying the interstitial sites. (The structure shown here represents a
perfect Sr2FeMoO6 double perovskite.) As the value of x is increased, the Fe and Mo octahedra rotate (indicated by the circular arrows), resulting in FeO6 octahedra that are
stretched and MoO6 octahedra that are compressed, and this results in a transition to a tetragonal unit cell (space group I4/m, right). The view along the c-axis (center right),
shows the rotation of the octahedra (O atoms are shown as gray spheres). The relationship between the lattice constants from the two space groups is shown below each
structure [5].
behavior, in which a decrease in the resistivity of the material is
observed upon applying a magnetic field [23,24,28]. This behavior
arises from the half-metallic character of Sr2FeMoO6, in which a
spin-polarized band of Mo t2g-O 2p-Fe eg states crosses the Fermi
level [16]. The magnetoresistive behavior has also been shown to
be impacted by the presence of insulting grain-boundaries (when
low magnetic fields are applied) and, because of this low-field
magnetoresistance, Sr2FeMoO6 has been investigated for use in
spintronic devices [23,29,30]. To better understand the origins of
the aforementioned properties, and to better utilize these proper-
ties in technological applications, it is necessary to know what oxi-
dation states are adopted by Fe and Mo in these compounds.
While there are a wide variety of studies on the valence states of
Fe and Mo, no consensus has been reached [11,15,26,31–35].
Although no definitive values have been reported, the oxidation
state of Fe is generally found to be between 2+ and 3+ while the
oxidation state of Mo is found to be between 5+ and 6+
[15,21,18,31–35]. This disagreement probably arises from the com-
plex nature of Sr2FeMoO6, with the material being sensitive toa wide
variety of variables. For example, Sr2FeMoO6 is not a stoichiometric
compound, but rather a member of the Sr2Fe2ꢀxMoxO6 (0 6 x 6 2)
solid solution. As such, the system is sensitive to both the composi-
tion and synthetic conditions employed [14,36]. Despite this, no
studies have directly investigated how the oxidation states of Fe
and Mo vary in Sr2Fe2ꢀxMoxO6 materials with changing Mo content.
X-ray absorption near-edge spectroscopy (XANES) is a tech-
nique perfectly suited for such a study. In this technique, X-rays
are used to excite core-level electrons to empty conduction states.
Transition-metal (TM) K-edge XANES spectra, which result from an
excitation of 1s core electrons, are especially sensitive to changes
in the coordination environment and oxidation state of the metal
center [37]. These spectra are typically divided into two regions;
the pre-edge region consisting of excitations to TM (n ꢀ 1)d states
and the main edge region, which arises from excitations to TM np
states. In general, the pre-edge region is weak because of the qua-
drapolar nature of these excitations and its lineshape is indicative
of the coordination environment surrounding the TM center [38–
42]. As the main-edge region results from dipolar transitions, the
intensity of this feature is typically strong [37]. The absorption
edge energy, defined as the inflection point of the absorption edge,
is sensitive to changes in the oxidation state of the metal center
[40,43,44]. In this study, Fe K- and Mo K-edge XANES spectra have
been collected from a series of Sr2Fe2ꢀxMoxO6 compounds pro-
duced by the reaction of oxides under H2(g) in order to determine
how the oxidation states of Fe and Mo depend on Mo content.
These experiments have also provided information about how
the oxygen deficiency (and coordination number) in these systems
change with varying Mo content. Additionally, to better under-
stand how preparation conditions might affect these oxidation
states, Fe K- and Mo K-edge XANES spectra have been collected
from samples that were annealed in a vacuum environment and
compared to those from the originally prepared materials.
2. Experimental
2.1. Synthesis
A series of Sr2Fe2ꢀxMoxO6 (x = 0.25, 0.50, 0.65, 0.85, 1.0) compounds were syn-
thesized via
a common solid-state synthesis route [11,14,36]. Stoichiometric
amounts of SrCO3 (Alfa Aesar, 99%), MoO3 (Acros Organics, >99%), and Fe2O3 (Alfa
Aesar, 99.945%) were mixed and ground using an agate mortar and pestle. In
materials with x P 0.65, the mixture was pressed into a pellet at 6 MPa and heated
to 900 °C over 8 h under flowing 5% H2 (balance N2), held at 900 °C for 6–8 h (higher
values of x required longer heating times), and cooled to room temperature over
ꢁ4.5 h. For materials with x 6 0.575, it was found that pressing the precursor mate-
rials into a pellet resulted in a phase-impure material. In these cases, the loose pow-
der was heated in air at 1100 °C for >12 h to decompose SrCO3 to SrO. The powdered
mixture was then heated to 900 °C over 8 h under flowing 5% H2 (balance N2), held
at 900 °C for 3–4 h (higher values of x required longer heating times), and then
cooled to room temperature over ꢁ4.5 h. While many other studies have reported
the use of higher temperatures (1000–1200 °C) to synthesize these materials, in this
study, an acceptable level of phase purity could not be achieved using temperatures
>900 °C when heated under 5% H2 (balance N2) [11,14,36]. In all cases, the gas flow
was monitored using a mineral oil bubbler and the flow was kept at 1–2 bubbles/s.
The flow rate was found to strongly impact the purity of the resultant product, as
too high or too low a flow rate resulted in the formation of phase-impure materials,
with Sr2MoO4 (too high a flow rate) and SrMoO4 (too low a flow rate) comprising
the principle impurity phases. The as-synthesized materials were also annealed
under vacuum to study the impact of ordering on the oxidation states of the metal
ions [42]. Aliquots of the as-synthesized samples were loaded into alumina
crucibles and sealed in fused-silica tubes under rough vacuum and the samples
were heated to 1050 °C from room temperature over ꢁ7 h, held at this temperature
for 12 h, and then slowly cooled to room temperature (ꢁ5 h).
Oxygen-deficient SrFeO2.75-perovskite, Sr2Fe2O5-brownmillerite, and FeO were
also synthesized to act as standards. To synthesize the SrFeO2.75 and Sr2Fe2O5 sam-
ples, stoichiometric amounts of SrCO3 (Alfa Aesar, 99%) and Fe2O3 (Alfa Aesar,
99.945%) were mixed, pelleted at 6 MPa, and heated in air at 1100 °C for 24 h. To
control the O-stoichiometry, the samples were then reground, and subjected to sep-
arate heat treatments. The SrFeO2.75 sample was heated at 1200 °C for 48 h, cooled
to 600 °C, and held at this temperature for 24 h before being slowly cooled to room
temperature. The Sr2Fe2O5 sample was heated at 1200 °C for 48 h and quench-
cooled in liquid nitrogen [45]. FeO was synthesized by mixing stoichiometric
amounts of Fe metal powder (Alfa Aesar 99.5%) and Fe2O3 (Alfa Aesar, 99.945%)
and pressing the resultant powder into a pellet at 6 MPa. The pellet was then heated
at 900 °C for 12 h under flowing N2(g) [46,47]. Phase purity of all materials synthe-
sized was determined using powder X-ray diffraction (XRD). Diffraction patterns
were collected using a PANalytical Empyrean X-ray diffractometer equipped with
a monochromatic Cu K (k = 1.5406 Å) X-ray source. Lattice parameters were
a
1
determined using the UnitCell software program by assuming a double perovskite
cubic unit cell for all values of x in the chemical formula [48].

J.R. Hayes, A.P. Grosvenor / Journal of Alloys and Compounds 537 (2012) 323–331
325
2.2. XANES
Fe K-edge and Mo K-edge spectra of the as-synthesized and annealed
Sr2Fe2ꢀxMoxO6 samples were collected using the Pacific Northwest Consortium/
X-ray Science Division Collaborative Access Team (PNC/XSD-CAT, Sector 20)
bending magnet beamline (20BM) located at the Advanced Photon Source (APS),
Argonne National Laboratory. A silicon (111) double crystal monochromator with
harmonic rejection was used, which at 7000 eV has a resolution of 1 eV and a
photon flux of ꢁ1011 photons/s [49]. Samples were finely ground and sandwiched
between layers of Kapton tape and the number of layers of tape was varied to
maximize absorption. Samples were mounted at a ꢁ45° angle to the incident beam
with transmission spectra being collected using N2-filled ionization chambers while
partial fluorescence yield spectra were collected using a Canberra 13-element Ge
detector. The spectra were collected using an energy step of either 0.15 eV (Fe K-
edge) or 0.30 eV (Mo K-edge) through the absorption edge. Spectra were calibrated
using either an Fe reference foil (having a known Fe K-edge absorption-edge energy
of 7112 eV) or a Mo reference foil (having a known Mo K-edge absorption energy of
20000 eV) [50]. The spectra from the reference foils were collected concurrently in
transmission mode with the spectra from each oxide studied. All spectra collected
were calibrated, normalized, and analyzed using the Athena software program [51].
As Sr2FeMoO6 materials have been shown to degrade into SrMoO4 and other prod-
ucts over time, to eliminate the possibility of sample age significantly affecting the
collected spectra, all spectra were collected within one month of sample prepara-
tion [52].
3. Results
3.1. Structure
To confirm phase purity, and to study how the structure of the
Sr2Fe2ꢀxMoxO6 compounds change with varying Mo content,
powder XRD patterns of the as-synthesized materials have been
collected and are presented in Fig. 2. With the exception of
Sr2FeMoO6, single phase Sr2Fe2ꢀxMoxO6 compounds were
prepared. In the case of SrFeMoO6, an 85–90% phase-pure material
was synthesized, with small amounts (<6% each) of SrMoO3,
Sr2MoO4, and intermetallic Fe1ꢀxMox impurities being detected.
(The exact composition of the intermetallic phase is not known
[13].) The (101) reflection (denoted by an asterisk in Fig. 2), which
corresponds to ordering of the Fe and Mo sites was only observed
in the Sr2FeMoO6 compound, indicating a high degree of disorder
in the other compounds synthesized [53]. It has been previously
shown that the intensity ratio of the (101) reflection to the sum
of the (200) and (112) reflections may be used as a measure of
the amount of antisite disorder experienced by the Fe/Mo sites,
with high values corresponding to less antisite-disorder [53]. For
the as-synthesized x = 1.0 sample, the (101):((200)+(112)) ratio
is 0.024, indicating that there is antisite disorder in this material.
(In the compounds with x < 1.0, the ratio is zero, indicating that
the amount of ordering of the transition metal sites in these sys-
tems is too low to be determined). Compared to previous studies,
where the reaction temperature was higher (ꢁ1200 °C), this ratio
was ꢁ0.046, which suggests that the lower reaction temperature
used in this study is not sufficient to achieve high levels of site
ordering [11,53].
Upon annealing under vacuum, the XRD pattern (Fig. 3) shows
that the ordering of Sr2FeMoO6 (x = 1.0) increases considerably
compared to the as-synthesized version. The powder XRD patterns
from the as-synthesized and vacuum annealed Sr2FeMoO6 samples
(Fig. 3) show that in the as-synthesized materials, the
(101):((200)+(112)) ratio is 0.024 while in the vacuum annealed
sample, the ratio is 0.033. This observation implies that the mate-
rial becomes more ordered upon annealing (i.e., the amount of
antisite disorder is reduced) [53]. Additionally, the phase-purity
of Sr2FeMoO6 increased dramatically with vacuum annealing,
which was observed by the complete disappearance of peaks from
the Sr2MoO4 and Fe1ꢀxMox phases in the powder XRD pattern
(Fig. 3). In the other annealed Sr2Fe2ꢀxMoxO6 compounds
(x < 1.0), no superstructural reflections were observed in the
Fig. 2. Powder XRD patterns from as-synthesized Sr2Fe2ꢀxMoxO6 materials.
Sr2FeMoO6 (x = 1.0) showed the following impurities: (1) SrMoO3; (2) Fe1ꢀxMox
intermetallic; and (3) Sr2MoO4. In addition, superstructural peaks from the
tetragonal structure (denoted by asterisks) are observed in the Sr2FeMoO6 sample.
All reflections were observed to shift to lower 2h angles with increasing x. The broad
feature found at ꢁ18° in the patterns with x < 1.0 are from mineral oil, which was
used to mount these samples.
corresponding powder XRD pattern; however, the other reflections
were observed to narrow considerably (not shown).
The XRD patterns collected in this study were also used to cal-
culate the lattice parameters of the Sr2Fe2ꢀxMoxO6 materials. The a
lattice parameter was calculated assuming a cubic unit cell. The
lattice constants are plotted as a function of Mo content in
Fig. 4a, and show a sudden increase when x P 0.65. For all values
of x P 0.65, the unit cell is observed to only increase slightly. This
shift is also clearly observed in the powder XRD patterns (cf.
Fig. 4b). The sudden and relatively large change in the lattice con-
stant between Sr2Fe1.50Mo0.50O6 and Sr2Fe1.65Mo0.35O6 is indicative
of a large change in the crystal lattice. In other structural studies of
Sr2Fe2ꢀxMoxO6, this change has been shown to coincide with a
transformation from a cubic to a tetragonal crystal structure. These
studies performed Reitveld refinements on powder XRD data, and
found the quality of the fit determined using a tetragonal model
was higher than that determined using a cubic model when
x P 0.7 [9,10,36]. In this study, upon annealing under vacuum,
no significant changes in the lattice parameter were observed
between the as-synthesized and annealed samples of a given
composition.

326
J.R. Hayes, A.P. Grosvenor / Journal of Alloys and Compounds 537 (2012) 323–331
Fig. 3. The normalized powder XRD patterns from the as-synthesized (gray)
Sr2FeMoO6 material and after being annealed under vacuum (black) are compared.
Here, the patterns are offset along the y-axis for clarity. After annealing, the
reflections at 30.6° and 44.7°, which are associated with impurity phases, were no
longer observed. The (101) reflection (marked by an asterisk), is shown in the inset.
The intensity of this reflection is greatly increased upon annealing in vacuum,
indicating that a more ordered structure is present.
3.2. Fe K-edge XANES of the As-synthesized materials
Fe K-edge XANES spectra, which are presented in Fig. 5,
were collected to determine the oxidation state of Fe in the
Sr2Fe2ꢀxMoxO6 materials and to investigate the oxygen stoichiom-
etry in these systems. Spectra from a series of standard compounds
having known Fe oxidation states and coordination environ-
ments were also collected to aid in the interpretation of the
Sr2Fe2ꢀxMoxO6 spectra, and are presented in Fig. 6. As is observed
in Fig. 5a, there are four distinct features in the spectra from
Sr2Fe2ꢀxMoxO6, which have been labeled as A, A0, B, and C. Features
A and A0 make up the pre-edge region (see Fig. 5b), and are as-
signed to forbidden, quadrapolar, 1s ? Fe 3d excitations [39]. By
comparison with the Fe K-edge spectrum from FeO (Fig. 6), feature
A is assigned to Fe2+ 1s ? Fe 3d excitations, while comparison to
the Fe K-edge spectra from SrFeO2.75 and Sr2Fe2O5 indicates
that feature A0 is attributable to Fe3+ 1s ? Fe 3d excitations. In
Sr2Fe1.75Mo0.25O6, only feature A0 is observed, implying that only
Fe3+ is observed in this compound. As the Mo content is increased
to x = 0.50, feature A0 decreases in intensity and no other features
are observed. As the coordination number (CN) of the Fe center de-
creases, and the local inversion symmetry is lost, 4p states overlap
the 3d states, which adds a dipolar character to the pre-edge exci-
tation. This increased dipolar character results in the observation
of a more intense pre-edge peak (cf. spectra from FePO4ꢂxH2O
(CNFe = 6) and SrFeO2.75 (CNFe < 6) in Fig. 6) [42,45]. (It should be
noted that the increase in dipolar character is a result of changes
in the local coordination environment only, and is not a result of
long-range distortions of the crystal structure.) Comparing the
spectra presented in Fig. 5 to those in Fig. 6 leads to the conclusion
that the decrease in intensity of feature A0 observed in Fig. 5b as the
Mo concentration is increased from x = 0.25 to 0.65 is a result of
the system becoming less O deficient (i.e., the Fe CN increases).
As the Mo content is further increased to x P 0.65, the intensity
of feature A0 increases, as does the intensity of feature A, which im-
plies that some Fe2+ is present. While feature A may contribute
some intensity to feature A0 as the two features are not fully
resolved, the fact that the intensities of both features increase
Fig. 4. (a) The lattice parameters calculated using
a cubic unit cell for the
Sr2Fe2ꢀxMoxO6 system as function of Mo content (x) are shown. Between
a
x = 0.50 and x = 0.65, a large increase is observed. The dashed line is included as a
guide to the eye and the size of the symbols is representative of the uncertainty of
the measurement. (b) The 2h angle of the (220) reflection (indexing assumes a
cubic structure) was found to shift to lower 2h values as x increases. The large
change in 2h between x = 0.50 and 0.65 is reflective of the significant variation in
the lattice shape.
simultaneously as x changes from 0.50 to 1.0 indicates that
Sr2Fe2ꢀxMoxO6 becomes more O deficient as x increases from
0.65 to 1, as an increase in the concentration of Fe2+ would de-
crease the number of unoccupied Fe 3d states. Examination of
Fig. 5b also indicates that the concentration of Fe2+ increases
significantly as x is increased when x P 0.65.
To confirm the analysis of the changes in the pre-edge region,
the main-edge can also be investigated. Features B and C comprise
the main-edge, and arise from dipolar 1s ? 4p excitations. By com-
paring the Sr2Fe2ꢀxMoxO6 spectra (Fig. 5) to the spectrum from FeO
(Fig. 6), it is clear that feature B arises from the excitation of Fe2+
metal centers. As was the case for feature A in the pre-edge region,
almost no change in intensity of feature B was observed as x was
varied from 0.25 to 0.50; however, the intensity of feature B was

J.R. Hayes, A.P. Grosvenor / Journal of Alloys and Compounds 537 (2012) 323–331
327
Fig. 5. (a) Fe K-edge spectra from the as-synthesized Sr2Fe2ꢀxMoxO6 materials are
shown. Features A and A0 make up the pre-edge region and consist of transitions to
Fe 3d states while features B and C are attributed to main-edge excitations resulting
from Fe 1s ? Fe 4p excitations. As the Mo content increases, features A and B
increase in intensity and the absorption edge energy shifts to lower energy,
indicating that the average Fe oxidation state is reduced (changes are shown by
arrows). (b) The pre-edge region of the Fe K-edge spectra is shown. As the Mo
content is increased, feature A significantly increases in intensity. The intensity of
feature A0 is linked to the O deficiency of the materials and initially decreases as x is
increased from x = 0.25 to 0.50. As x is increased to values greater than 0.50, the
intensity of feature A is observed to increase.
Fig. 6. (a) Fe K-edge spectra from a series of standards is presented along with the
Fe K-edge spectrum from the as-synthesized Sr2FeMoO6 sample. (The spectra of
FePO4 and FePO4ꢂxH2O are from reference [53].) (b) The pre-edge region of the Fe K-
edge spectra is shown. Comparison of the spectra from Sr2FeMoO6 and FeO
indicates that feature A is linked to the presence of Fe2+ while the link between the
oxygen deficiency and the intensity of feature A0 is corroborated by comparison of
the spectra from SrFeO2.75 (CN < 6) and Sr2Fe2O5 (average CN = 5).
of feature C in Fig. 5. Multiscattering resonances, which are part of
the extended X-ray absorption fine structure (EXAFS), may also
contribute to the intensity of feature C. The changes in the Fe K-
edge XANES spectra as Mo content is increased may be summa-
rized as follows: (1) for values of x 6 0.50, Fe adopts a 3+ oxidation
state; (2) when x is greater than 0.50, the average oxidation state of
Fe decreases with increasing Mo content; (3) the O deficiency of
the structure decreases as the Mo content increases when
x 6 0.50, but increases with increasing Mo content when x > 0.50;
and (4) for no composition studied here was a fully oxygen stoichi-
ometric compound synthesized.
found to vary significantly with increasing Mo content when
x P 0.65. The presence of Fe2+ in these materials with varying x
is further supported by the observed shifts in the main-edge en-
ergy of these spectra, which is defined as the inflection point of
the main-edge (determined from the derivative spectrum by find-
ing the peak maximum of the feature corresponding to the main-
edge). The decrease in absorption energy with greater substitution
of Fe for Mo (
D
E ꢃ 1.5 eV), is ascribed to the presence of lower-en-
ergy valence Fe 4p-O 2p states, caused by increased ground-state
screening of the Fe nuclear charge that results from a reduction
of Fe3+ to Fe2+ [54].
3.3. Mo K-edge XANES of the As-synthesized materials
The intensity of feature C initially increases between Sr2Fe1.75-
Mo0.25O6 and Sr2Fe1.50Mo0.50O6, but then remains relatively con-
stant as the Mo content is further increased. This increase in
intensity is likely linked to the O deficiency of the system. Given
that there must be a fixed number of Fe 4p states, as the Fe 4p
character of the pre-edge peak is reduced with increasing CN
(i.e., decreased O deficiency; cf. Feature A0 Fig. 5b), the remaining
Fe 4p states shift to higher energy and interact with the O 2p states
(feature C), resulting in the observation of the increase in intensity
Mo K-edge spectra from the Sr2Fe2ꢀxMoxO6 compounds were
collected to better understand how the Mo oxidation state changes
with substitution, and are presented in Fig. 7a. In addition, to inter-
pret changes in these Mo K-edge spectra, XANES spectra from sev-
eral reference compounds were collected, and are presented in
Fig. 7b. Two prominent features are observed in these spectra,
and are labeled as D and E.
Feature D in Fig. 7 arises from forbidden transitions to Mo 4d
states (much like in the Fe K-edge), and the intensity of this feature

328
J.R. Hayes, A.P. Grosvenor / Journal of Alloys and Compounds 537 (2012) 323–331
appears to be related to the Mo content in Sr2Fe2ꢀxMoxO6 [55–58].
When x = 0.25 in Sr2Fe2ꢀxMoxO6, this feature is muted, while for
higher values of x, the intensity of feature D increases. In all cases,
the intensity of feature D is far less than that observed in the spec-
tra from SrMoO4 and MoO3, in which Mo occupies a tetrahedral
site or a highly distorted octahedral site, respectively (Fig. 7b).
(In SrMoO4, the highly distorted octahedra result in significant
overlap between the Mo 4d and 5p states [59–61].) In the Sr2Fe2ꢀx-
MoxO6 compounds, the increase in the intensity of feature D with
increasing Mo content suggests that, overall, the Mo coordination
decreases with greater Mo content. However, this interpretation
is tentative as it is difficult to separate changes in the intensity
of feature D from shifts of the main absorption edge energy, given
the poor energy resolution of Mo K-edge spectra (ꢁ4.5 eV when
core–hole lifetime effects are considered) [58].
Feature E is attributed to the main-edge excitation, which arises
from Mo 1s ? Mo 5p excitations [58]. The main-edge energy is
sensitive to changes in oxidation state and can be used to elucidate
how the oxidation state of Mo in the Sr2Fe2ꢀxMoxO6 compounds
changes with varying stoichiometry. Fig. 7c shows the derivative
spectra from the Sr2Fe2ꢀxMoxO6 compounds along with that from
the MoO2 and SrMoO4 reference compounds. When the Mo content
is low (x = 0.25), the derivative maximum is at the same energy as
that from the SrMoO4 spectrum, indicating that Mo adopts an aver-
age oxidation state of 6+ in this material. As the Mo content in-
creases in Sr2Fe2ꢀxMoxO6 to x = 0.50, the peak maximum of the
derivative spectrum is observed to shift to lower energy, indicating
a reduction in the Mo oxidation state (Fig. 7c). When the Mo con-
tent is further increased (x = 0.65), the derivative peak again shifts
to lower energy, though it is still slightly higher than that of MoO2
(Mo4+). This observation suggests that Mo largely adopts a 5+
oxidation state when x P 0.65, though given the width of the
MoO2 derivative peak, it is possible that a minor quantity of
Mo4+ is also present. (The width of the derivative peaks also shows
that a small amount of Mo6+ is likely present when x > 0.25). Inter-
estingly, as the Mo content is increased beyond x = 0.65, no further
shifts in energy are observed.
3.4. Vacuum annealing
To understand how ordering of the Mo and Fe ions in
Sr2Fe2ꢀxMoxO6 might impact the oxidation state and electronic
environment of the metal centers, the as-synthesized compounds
were annealed in an evacuated ampule and investigated by XANES.
The Fe K- and Mo K-edge XANES spectra from these annealed
compounds are compared to the corresponding spectra from the
as-synthesized samples in Figs. 8 and 9, respectively. Large changes
in the Fe K-edge spectra were observed between the as-synthe-
sized and annealed Sr2Fe2ꢀxMoxO6 compounds when x P 0.65
(Fig. 8); however, no change was observed in the spectra from
Sr2Fe1.75Mo0.25O6 and Sr2Fe1.50Mo0.50O6. For compounds where
x P 0.65, features A, A0 and B decreased in intensity while the
intensity of feature C was found to increase. These changes are
consistent with the oxidation of Fe2+ in the Sr2Fe2ꢀxMoxO6
materials and an increase in the Fe CN. The source of oxygen is
attributed to residual O2(g) left in the evacuated ampule. In previ-
ous studies on the synthesis of SrMoO3, a compound that is related
to the Sr2Fe2ꢀxMoxO6 compounds studied here, it was shown that
oxygen partial pressures as low as 10ꢀ14 Torr could lead to the oxi-
dation of SrMoO3 [61–63]. As the ampules containing the annealed
samples were only sealed under rough vacuum, the partial pres-
sure of O2(g) present was much higher than 10-14 Torr. Oxidation
of the Fe centers leads to an increase in the Fe CN, which was ob-
served as a decrease in the intensity feature A0 owing to a decrease
in mixing of the Fe 3d and 4p states. The decrease in intensity of
feature A is also explained by the oxidation of Fe2+ to Fe3+. Further,
Fig. 7. (a) Mo K-edge spectra from the as-synthesized Sr2Fe2ꢀxMoxO6 materials. As
the Mo content (x) increases, feature D increases in intensity, indicating a decrease
in the Mo CN. Arrows are included to show how features change as x increases. (b)
The Mo K-edge spectra from a series of standards, as well as the as-synthesized
Sr2FeMoO6 sample, are plotted. (Mo adopts a 4+ oxidation state in MoO2 and a 6+
oxidation state in MoO3 and SrMoO4.) (c) The derivative plot of the Mo K-edge
spectra the Sr2Fe2ꢀxMoxO6 samples as well as the MoO2 and MoO3 standards is
shown. For clarity, the MoO2 derivative spectrum has been smoothed and the
vertical dashed lines represent the peak energies of MoO2 and MoO3.
feature B (Section 3.2) was also observed to decrease in intensity,
compared to the as-synthesized compounds because of the oxida-
tion of Fe2+ to Fe3+. The increase in intensity of Feature C was a re-
sult of the greater number of Fe 4p-O 2p states available for 1s

J.R. Hayes, A.P. Grosvenor / Journal of Alloys and Compounds 537 (2012) 323–331
329
electrons to be excited to, which is indicative of an increased Fe CN
[45]. Interestingly, the magnitude of the changes observed upon
annealing increased with greater Mo content. The reason for this
observation is simple: at higher Mo concentrations, the amount
of Fe2+ present which can be oxidized is greater.
The Mo K-edge spectra of the annealed and synthesized sam-
ples (shown in Fig. 9) reveal that little to no change upon anneal-
ing. This observation does not indicate that no changes occur at the
Mo sites. Instead, this result likely indicates that the magnitude of
the changes is not detectable within the limits of the resolution of
the Mo K-edge XANES spectra reported here.
4. Discussion
In this study, a series of Sr2Fe2ꢀxMoxO6 double pervoskites have
been synthesized and studied by XANES to better understand the
relationship between the concentration of Mo and the oxidation
states of Fe and Mo, and the level of O deficiency in these materials.
In Sr2Fe1.75Mo0.25O6, the Fe K-edge XANES spectrum (Fig. 5) indi-
cates that Fe adopts a 3+ oxidation state while the Mo K-edge
XANES spectrum (Fig. 7a) indicates that Mo adopts a 6+ oxidation
state. Given the similarity of the effective ionic radii of Fe3+ and
Mo6+ (0.645 and 0.59 Å, respectively), a near-random distribution
of Mo6+ and Fe3+ would lead to an entropy stabilized system, which
could be thought of as a solid-solution derivative of the cubic
SrFeO3ꢀd perovskite structure (here, d indicates an unknown, but
small amount of O deficiency) [7,8,64]. This suggestion is con-
firmed in this study by the lack of observation of the (101)
reflection in the corresponding powder XRD patterns (Fig. 3).
(While SrFeO3ꢀd is reported to have a mix of Fe3+ and Fe4+ when
d is small, given the reducing conditions used to synthesize the
Sr2Fe2ꢀxMoxO6 materials analyzed in this study, the formation of
Fe4+ is not likely). In the case of Sr2Fe1.75Mo0.25O6, an oxygen
Fig. 9. The Mo K-edge spectra from several of the vacuum annealed Sr2Fe2ꢀxMoxO6
spectra are presented along with the corresponding as-synthesized materials
(spectra are offset for clarity). Little to no change in the spectra was observed for all
values of x.
deficient structure is required to achieve charge balance, as the
maximum number of O atoms possible per formula unit is 5.4,
assuming only Fe3+ and Mo6+ are present. Oxygen deficiency in this
compound was confirmed in this study by the intense pre-edge
feature in the Fe K-edge spectrum (Fig. 5). As the Mo content is in-
creased to Sr2Fe1.50Mo0.50O6, the level of O deficiency is reduced,
consistent with the replacement of Fe3+ by Mo6+ (as this leads to
an overall increase of the average cationic charge). The increase
in O content also explains the small increase of the lattice param-
eter (Fig. 4a) and the decrease in the pre-edge intensity in the
corresponding Fe K-edge spectrum (Fig. 5b) as addition of large
O2ꢀ anions would expand the unit cell and increase the number
of octahedrally coordinated Fe centers.
As the Mo concentration was further increased to x = 0.65
(Sr2Fe1.35Mo0.65O6), the lattice parameter was calculated to be
significantly larger than that of Sr2Fe1.50Mo0.50O6 (Fig. 4), and a
large increase in the amount of Fe2+ and Mo5+ was observed in
these materials, as shown by the changes in the Fe K-edge
(Fig. 5; Section 3.2) and Mo K-edge spectra (Fig. 6a; Section 3.3).
Here, it is proposed that between x = 0.50 and 0.65, a critical con-
centration of Mo is reached which allows for a reduction of both
Fe and Mo without creating a destabilizing amount of O vacancies,
as the oxidation state of Mo is always greater than that of Fe. The
large shift in the lattice parameter (Fig. 4a) is attributed to a change
ꢀ
in the unit cell shape from a cubic (Fm3m) structure to a tetragonal
(I4/m) structure (Fig. 1) [9,36]. As Fe3+ is reduced to Fe2+, there is a
dramatic increase in the effective ionic radius (rFe2+ = 0.780 Å vs.
rFe3+ = 0.645 Å) while there is only a minor difference between
the effective ionic radii of Mo6+ and Mo5+ (rMo5+ = 0.61 Å vs.
rMo6+ = 0.59 Å) [64]. Thus, the tetragonal distortion of the lattice
is the result of the steric distortions necessary to incorporate a sig-
nificant amount of Fe2+, which occupies the larger octahedral sites
in this structure [9,10].
Fig. 8. The Fe K-edge spectra from several of the vacuum annealed Sr2Fe2ꢀxMoxO6
samples are presented along with the corresponding as-synthesized materials
(spectra are offset for clarity). No change in the spectra were observed upon
annealing when the Mo content is low (x 6 0.50). When x > 0.50, the intensities of
features A0 and B in the spectra from the vacuum annealed samples decrease
compared to the as-synthesized samples, while the intensity of feature C increases.
These observations indicate that the materials have been oxidized upon annealing
and that the Fe CN increases.
To determine how the ordering and O deficiency of the
Sr2Fe2ꢀxMoxO6 materials is impacted by further heating, the

330
J.R. Hayes, A.P. Grosvenor / Journal of Alloys and Compounds 537 (2012) 323–331
as-synthesized materials were annealed in a vacuum environment.
Through examination of the Fe K-edge XANES spectra, oxidation of
the Fe cations was observed to occur upon annealing the as-syn-
thesized materials under vacuum. While oxidation of the Fe site
results in a decrease of the average Fe radius and would be
expected to decrease ordering based on the discussion above, anal-
ysis of the powder XRD patterns showed that the ordering of the Fe
and Mo sites increased in Sr2FeMoO6 (Section 3.1; Fig. 4). This may
be explained by the increased temperature used to anneal these
samples (the samples were reduced at 900 °C (as synthesized)
and annealed at 1050 °C), which apparently leads to increased
ordering. Thus, these results serve to further confirm the depen-
dence of the degree of ordering in Sr2Fe2ꢀxMoxO6 materials on tem-
perature, as has been found in previous studies [11,30]. These
results also suggest that the oxidation state of Fe found in Mo-rich
Sr2Fe2ꢀxMoxO6 materials (those with x P 0.65) is highly sensitive
to the atmosphere in which the materials are synthesized, consis-
tent with studies of the related SrMoO3 perovskite phase [62,63].
ing access to the powder X-ray diffractometer located at the Uni-
versity of Alberta to perform preliminary measurements. Dr.
Robert Gordon is thanked for his help in carrying out XANES mea-
surements at 20BM (PNC/XSD-CAT, APS). The PNC/XSD facilities at
the APS, and research at these facilities, are supported by the US
Department of Energy - Basic Energy Sciences, a Major Resources
Support Grant from NSERC, the University of Washington, Simon
Fraser University and the APS. Use of the APS, an Office of Science
User Facility operated for the U.S. Department of Energy (DOE) Of-
fice of Science by Argonne National Laboratory, was supported by
the U.S. DOE under Contract No. DE-AC02-06CH11357.
References
[1] I.V. Solovyev, J. Magn. Magn. Mater. 268 (2004) 194–197.
[2] W. Prellier, V. Smolyaninova, A. Biswas, C. Galley, R.L. Greene, K. Ramesha, J.
Gopalakrishnan, J. Phys.: Condens. Matter 12 (2000) 965–973.
[3] A. Millis, J. Nature 392 (1998) 147–150.
[4] S. Yousif, O.A. Yassin, J. Alloys Compd. 506 (2010) 456–460.
[5] K. Momma, F.J. Izumi, Appl. Cryst. 41 (2008) 653–658.
[6] P.M. Woodward, R.-D. Hoffman, A.W. Sleight, J. Mater. Res. 9 (1994) 2118–
2127.
5. Conclusion
[7] F. Galasso, W. Darby, J. Phys. Chem. 66 (1962) 131–132.
[8] D. Serrate, J.M. De Teresa, M.R. Ibarra, J. Phys.: Condens. Matter 19 (2007)
023201.
A series of Sr2Fe2ꢀxMoxO6 compounds were synthesized and
studied by powder XRD and XANES to determine how the oxida-
tion state of Fe and Mo depends on the Mo concentration. The
powder X-ray diffraction patterns indicated that a large variation
in the lattice parameter of these compounds occurred between
the compounds with x 6 0.50 and x P 0.65. This is in agreement
with previous studies in which a transition from a cubic to tetrag-
onal unit cell was reported [9,10,36]. Accompanying this change in
structure is a partial reduction of the Fe oxidation state from Fe3+ to
a mixture of Fe2+ and Fe3+ and a partial reduction of the Mo6+
oxidation state to a mixture of Mo5+ and Mo6+ (minor amounts of
Mo4+ may also be present). Here, changes in the oxidation states
of Fe and Mo have been spectroscopically shown to occur with
changing Mo-content. Further, the results in this study indicate
that the change in oxidation state of Fe from Fe3+ to Fe2+/3+ is the
driving factor in the transformation of the cubic lattice to a tetrag-
onal lattice, and that the reduction of Fe is possible only after a
critical concentration of Mo has been reached. The Fe K-edge XANES
spectra also indicated that a significant amount of O vacancies
exist in all of the materials studied. Upon annealing under a vac-
uum environment, Fe2+/3+ was oxidized to Fe3+ and the oxygen
deficiency of these compounds was significantly reduced. While
O vacancies are already known to occur in these systems, using
XANES, this study has shown that the amount of O deficiency is
dependent on the amount of O2(g) present in the atmosphere in
which these materials are reacted. The results reported here show
that the composition strongly impacts the oxidation states of Fe
and Mo in Sr2Fe2ꢀxMoxO6 materials, which are linked to structural
changes within this system, and the dependence on both composi-
tion and preparation method likely accounts for the wide range of
oxidation states and magnetic properties that have been reported
previously.
[9] G.Y. Liu, G.H. Rao, X.M. Feng, H.F. Yang, Z.W. Ouyang, W.F. Liu, J.K. Liang, J.
Alloys Compd. 353 (2003) 42–47.
[10] G.H. Rao, G.Y. Liu, X.M. Feng, Q. Zhang, J.K. Liang, Sci. Technol. Adv. Mater. 6
(2005) 750–754.
[11] B. Jurca, J. Berthon, N. Dragoe, P. Berthet, J. Alloys Compd. 474 (2009) 416–423.
[12] C.J. Howard, B.J. Kennedy, P.M. Woodard, Acta Crystallogr., Sect. B 59 (2003)
463–471.
[13] A. Sharma, A. Berenov, J. Rager, W. Branford, Y. Bugoslavsky, L.F. Cohen, J.L.
MacManus-Driscoll, Appl. Phys. Lett. 83 (2003) 2384–2386.
[14] J. Rager, M. Zipperle, A. Sharma, J.L. MacManus-Driscoll, J. Am. Ceram. Soc. 87
(2004) 1330–1335.
[15] K. Kuepper, I. Balasz, H. Hesse, A. Winiarski, K.C. Prince, M. Matteucci, D. Wett,
R. Szargan, E. Burzo, M. Neumann, Phys. Status Solidi 201 (2004) 3252–3256.
[16] K.-I. Kobayashi, T. Kimura, H. Sawada, K. Terakura, Y. Tokura, Nature 395
(1998) 677–680.
[17] M. Tovar, M.T. Causa, A. Butera, J. Navarro, B. Martínez, J. Fontcuberta, M.C.G.
Passeggi, Phys. Rev. B 66 (2002) 024409.
[18] H. Jalili, N.F. Heinig, K.T. Leung, Phys. Rev B 79 (2009) 174427.
[19] A.W. Sleight, J.F. Weiher, J. Phys. Chem. Solids 33 (1972) 679–687.
[20] K.-I. Kobayashi, T. Okuda, Y. Tomioka, T. Kimura, Y. Tokura, J. Magn. Magn.
Mater. 218 (2000) 17–24.
[21] S. Ray, A. Kumar, D.D. Sarma, R. Cimino, S. Turchini, S. Zennaro, N. Zema, Phys.
Rev. Lett. 87 (2001) 097204.
[22] S. Ray, A. Kumar, S. Majumdar, E.V. Sampathkumaran, D.D. Sarma, J. Phys.:
Condens. Matter 13 (2001) 607–616.
[23] D. Niebieskikwiat, F. Prado, A. Caneiro, R.D. Sánchez, Phys. Rev. B 70 (2004)
132412.
[24] A. Dinia, J. Vénaut, S. Colis, G. Pourroy, Catal. Today 89 (2004) 297–302.
[25] X.Z. Liao, A. Sharma, M. Wei, J.L. MacManus-Driscoll, W. Branford, L.F. Cohen, Y.
Bugoslavsky, Y.T. Zhu, D.E. Peterson, Y.B. Jiang, H.F. Xu, J. Appl. Phys. 96 (2004)
7747–7749.
[26] J.-S. Kang, J.H. Kim, A. Sekiyama, S. Kasai, S. Suga, S.W. Han, K.H. Kim, T. Muro,
Y. Saitoh, C. Hwang, C.G. Olson, B.J. Park, B.W. Lee, J.H. Shim, J.H. Park, B.I. Min,
Phys. Rev. B 66 (2002) 113105.
[27] C. Meneghini, S. Ray, F. Liscio, F. Bardelli, S. Mobilio, D.D. Sarma, Phys. Rev. Lett.
103 (2009) 046403.
[28] C.A. López, M.-C. Viola, J.C. Pedregosa, J.A. Alonso, M.T. Fernández-Díaz, Eur. J.
Inorg. Chem. 2010 (2010) 4110–4120.
[29] J.L. MacManus-Driscoll, A. Sharma, Y. Bugoslavsky, W. Branford, L.F. Cohen, M.
Wei, Adv. Mater. 18 (2006) 900–904.
[30] D.D. Sarma, S. Ray, K. Tanaka, M. Kobayashi, A. Fujimori, P. Sanyal, H.R.
Krishnamurthy, C. Dasgupta, Phys. Rev. Lett. 98 (2007) 157205.
[31] M.S. Moreno, J.E. Gayone, M. Abbate, A. Caneiro, D. Niebieskikwait, R.D.
Sánchez, A. de Siervo, R. Landers, G. Zampieri, Solid State Commun. 120 (2001)
161–164.
Acknowledgements
[32] N. Menéndez, M. García-Hernández, D. Sánchez, J.D. Tornero, J.L. Martínez, J.A.
Alonso, Chem. Mater. 16 (2004) 3565–3572.
[33] V. Kanchana, G. Vaitheeswaran, M. Alouani, A. Delin, Phys. Rev. B 75 (2007)
220404.
[34] J. Herrero-Martín, J. García, G. Subías, J. Blasco, M.C. Sánchez, J. Phys.: Condens.
Matter 16 (2004) 6877–6890.
[35] M. Retuerto, F. Jiménez-Villacorta, M.J. Martínez-Lope, Y. Huttel, E. Roman,
M.T. Fernández-Díaz, J.A. Alonso, Phys. Chem. Chem. Phys. 12 (2010) 13325–
13616.
The Natural Sciences and Engineering Research Council (NSERC)
of Canada supported this work through a discovery Grant awarded
to APG. JRH would also like to thank NSERC for financial support
through the Canada Graduate Scholarship program. The University
of Saskatchewan is thanked for funding provided to JRH. The Cana-
dian Foundation for Innovation (CFI) is thanked for providing funds
to purchase the PANalytical Empyrean powder X-ray diffractome-
ter used in this work. The authors would also like to thank Prof.
A. Mar (Department of Chemistry, University of Alberta) for provid-
[36] T.-T. Fang, T.-F. Ko, J. Am. Ceram. Soc. 86 (2003) 1453–1455.
[37] L.A. Grunes, Phys. Rev. B 27 (1983) 2111–2131.
[38] G.A. Waychunas, M.J. Apted, G.E. Brown, Phys. Chem. Miner. 10 (1983) 1–9.

J.R. Hayes, A.P. Grosvenor / Journal of Alloys and Compounds 537 (2012) 323–331
331
[39] F. de Groot, G. Vankó, P. Glatzel, J. Phys.: Condens. Matter 21 (2009) 104207.
[40] T.E. Westre, P. Kennepohl, J.G. DeWitt, B. Hedman, K.O. Hodgson, E.I. Solomon,
J. Am. Chem. Soc. 119 (1997) 6297–6314.
[41] M.-A. Arrio, S. Rossano, Ch. Brouder, L. Galoisy, G. Calas, Europhys. Lett. 51
(2000) 454–460.
[42] J.-P. Reuff, L. Journel, P.-E. Petit, F. Farges, Phys. Rev. B 69 (2004) 235107.
[43] F. Farges, E. Chalmin, C. Vignaud, I. Pallot-Frossard, J. Susini, J. Bargar, G.E.
Brown Jr., M. Menu, Phys. Scr. T115 (2005) 885–887.
[44] F. Farges, S. Rossano, Y. Lefrère, M. Wilke, G.E. Brown Jr., Phys. Scr. T115 (2005)
957–959.
[53] L1. Balcells, J. Navarro, M. Bibes, A. Roig, B. Martínez, J. Fontcuberta, Appl. Phys.
Lett. 78 (2001) 781–783.
[54] J.A. Sigrist, M.W. Gaultois, A.P. Grosvenor, J. Phys. Chem. A 115 (2011) 1908–
1912.
[55] K. Yokoi, N. Matsubayashi, T. Miyanaga, I. Watanabe, K. Murata, S. Ikeda, Chem.
Lett. 16 (1987) 1453–1456.
[56] P.P. Samuel, S. Horn, A. Döring, K.G.V. Havelius, S. Reschke, S. Leimkühler, M.
Haumann, C. Schulzke, Eur. J. Inorg. Chem. 2011 (2011) 4387–4399.
[57] H. Yan, R.A. Mayanovic, A.J. Anderson, P.R. Meredith, Nucl. Instrum. Methods
Phys. Res., Sect. A 649 (2011) 207–209.
[45] M.W. Gaultois, A.P. Grosvenor, J. Phys. Chem. C 114 (2010) 19822–19829.
[46] L. Chi, A.E.C. Green, R. Hammond, C.R. Wiebe, J.E. Greedan, J. Solid State Chem.
170 (2006) 165–175.
[58] F. Farges, R. Siewert, G.E. Brown Jr., A. Guesdon, G. Morin, Can. Mineral. 44
(2006) 731–753.
[59] C. Bernuy-Lopez, M. Allix, C.A. Bridges, J.B. Claridge, M. Rosseinsky, J. Chem.
Mater. 19 (2007) 1035–1043.
[47] E.R. Jette, F. Foote, J. Chem. Phys. 1 (1933) 29–36.
[48] T.J.B. Holland, S.A.T. Redfern, Mineral. Mag. 61 (1997) 65–77.
[49] S.M. Heald, D.L. Brewe, E.A. Stern, K.H. Kim, F.C. Brown, D.T. Jiang, E.D. Crozier,
R.A. Gordon, J. Synchrotron Radiat. 6 (1999) 347–349.
[50] A. Thompson, D. Attwood, E. Gullikson, M. Howells, K.-J. Kim, J. Kirz, J.
Kortright, I. Lindau, P. Pianetta, A. Robinson, J. Scofield, J. Underwood, D.
Vaughan, G. Williams, H. Winick, X-ray Data Booklet, Lawrence Berkeley
National Laboratory, Berkeley, 2009.
[51] B. Ravel, M. Newville, J. Synchrotron Radiat. 12 (2005) 537–541.
[52] K. Kuepper, M. Raekers, C. Taubitz, H. Hesse, M. Neumann, A.T. Young, C.
Piamonteze, F. Bondino, K.C. Prince, J. Appl. Phys. 104 (2008) 036103.
[60] T. Leisegang, A.A. Levin, J. Walter, D.C. Meyer, Cryst. Res. Technol. 40 (2005)
95–105.
[61] A. Kuzmin, J. Purans, J. Phys.: Condens. Matter 12 (2000) 1959–1970.
[62] H.H. Wang, D.F. Cui, Y.L. Zhou, Z.H. Chen, F. Chen, T. Zhao, H.B. Lu, G.Z. Yang,
M.C. Xu, Y.C. Lan, X.L. Chen, H.J. Qian, F.Q. Liu, J. Cryst. Growth 226 (2001) 261–
266.
[63] I. Nagai, N. Shirakawa, S. Ikeda, R. Iwasaki, H. Nishimura, M. Kosaka, Appl. Phys.
Lett. 87 (2005) 024105.
[64] R.D. Shannon, Acta Cryst. A32 (1976) 751–767.