The topotactic reduction of Sr3Fe2O 5Cl2-square planar Fe(II) in an extended oxyhalide

Article abstract of DOI:10.1021/ic101371z

The topotactic reduction of the oxychloride Sr₃Fe ₂O₅Cl₂ with LiH results in the formation of Sr₃Fe₂O₄Cl₂. Neutron powder diffraction data show that Sr₃Fe₂O₄Cl ₂ adopts a body-centered tetragonal crystal structure (I4/mmm, a = 4.008(1) A, c = 22.653(1) A at 388 K) with anion vacancies located within the SrO layer of the phase. This leads to a structure consisting of infinite sheets of corner-sharing Fe(II)O₄ square planes. Variable-temperature neutron diffraction data show that Sr₃Fe ₂O₄Cl₂ adopts G-type antiferromagnetic order below T_N ～ 378(10) K with an ordered moment of 2.81(9) μ_B per iron center at 5 K consistent with the presence of high-spin Fe(II). The observed structural and chemical selectivity of the reduction reaction is discussed. The contrast between the structure of Sr ₃Fe₂O₄Cl₂ and the isoelectronic all-oxide analogue (Sr₃Fe₂O₅) suggests that by careful selection of substrate phases, the topotactic reduction of complex transition metal oxychlorides can lead to the preparation of novel anion-deficient phases with unique transition metal-oxygen sublattices which cannot be prepared via the reduction of all-oxide substrates.

Full text of DOI:10.1021/ic101371z

Inorg. Chem. 2010, 49, 9649–9654 9649
DOI: 10.1021/ic101371z
The Topotactic Reduction of Sr Fe O Cl ;Square Planar Fe(II)
3
2 5
2
in an Extended Oxyhalide
Edward Dixon and Michael A. Hayward*
Department of Chemistry, University of Oxford, Inorganic Chemistry Laboratory, South Parks Road,
Oxford, OX1 3QR, United Kingdom
Received July 9, 2010
The topotactic reduction of the oxychloride Sr Fe O Cl with LiH results in the formation of Sr Fe O Cl . Neutron
3
2
5
2
3
2
4
2
powder diffraction data show that Sr Fe O Cl adopts a body-centered tetragonal crystal structure (I4/mmm, a =
2
3
2
4
˚ ˚
.008(1) A, c = 22.653(1) A at 388 K) with anion vacancies located within the SrO layer of the phase. This leads to
4
a structure consisting of infinite sheets of corner-sharing Fe(II)O square planes. Variable-temperature neutron
4
diffraction data show that Sr Fe O Cl adopts G-type antiferromagnetic order below T ∼ 378(10) K with an ordered
3
2
4
2
N
moment of 2.81(9) μ per iron center at 5 K consistent with the presence of high-spin Fe(II). The observed structural
B
and chemical selectivity of the reduction reaction is discussed. The contrast between the structure of Sr Fe O Cl and
the isoelectronic all-oxide analogue (Sr Fe O ) suggests that by careful selection of substrate phases, the topotactic
3
2
4
2
3
2 5
reduction of complex transition metal oxychlorides can lead to the preparation of novel anion-deficient phases with
unique transition metal-oxygen sublattices which cannot be prepared via the reduction of all-oxide substrates.
Introduction
new phases with novel transition metal oxidation states and
1
-3
unusual local coordinations.
In order to reap the full
Complex transition metal oxides and related mixed anion
phases are of great importance from an academic and tech-
nological perspective due to the diverse array of physical
properties they exhibit. These properties typically arise from
the cooperative behavior of electrons located in partially
filled d states. By modifying the valence electron count (metal
oxidation state), local metal coordination (crystal field effects),
and long-range cation-anion-cation connectivity (exchange
pathways), chemists can tune and manipulate the behavior
of these electrons and thus the physical behavior of these
materials. Conventionally, such physical property tuning has
been achieved via selective cation substitutions, which readily
modify transition metal oxidation states and valence electron
counts. However there is growing interest in tuning the physi-
cal behavior of solids containing transition metals by manip-
ulating their anion lattices because this not only changes the
electron count (through reduction or oxidation) but can also
modify the local transition metal coordination sphere and the
long-range cation-anion-cation connectivity, potentially
leading to more dramatic changes in physical behavior.
Recently, we have demonstrated that binary metal hydrides
can be used as powerful solid state reducing agents for the
low-temperature deintercalation of oxide ions (topotactic
reduction) from complex transition metal oxides to yield
benefit of this synthetic approach and achieve a situation
where the directed synthesis of target phases is possible, it is
necessary to control which of the various anions in a
substrate phase are removed and how the resulting anion
vacancies order.
One strategy to exert such control over topotactic reduc-
tion reactions utilizes the strong interaction between the
anion and cation lattices in complex oxides to direct the pro-
gress and products of anion deintercalation reactions. For
example, the reduction of the isostructural and isoelectronic
A-cation ordered phases YBaCo O and LaBaCo O with
2
5
2
5
4
NaH yields YBaCo O and LaBaCo O4.25, respectively.
2
4.5
2
These two reduced phases have different anion contents and
anion vacancy ordering patterns, despite starting from iso-
structural substrate phases, demonstrating the directing in-
fluence of the A cations in this case. Building on this observa-
tion, we have sought to investigate the directing influences
of heteroanions on the progress of topotactic reductions
in mixed anion solids. Here, we describe the first topotactic
reduction of a complex oxyhalide, Sr Fe O Cl , an n = 2
3
2
5
2
Ruddlesden-Popper phase with a structure analogous to
Sr Fe O (Figure 1). Reaction of Sr Fe O Cl with LiH
3
2
7
3
2
5
2
yields Sr Fe O Cl , a phase which contains sheets of
3
2
4
2
*
To whom correspondence should be addressed. Tel.: þ44 1865 272623.
Fax: þ44 1865 272690. E-mail: michael.hayward@chem.ox.ac.uk.
1) Hayward, M. A.; Green, M. A.; Rosseinsky, M. J.; Sloan, J. J. Am.
(2) Adkin, J. J.; Hayward, M. A. J. Solid State Chem. 2006, 179, 70–76.
(3) Hayward, M. A. Chem. Mater. 2005, 17, 670–675.
(4) Seddon, J.; Suard, E.; Hayward, M. A. J. Am. Chem. Soc. 2010, 132,
2802–2810.
(
Chem. Soc. 1999, 121, 8843–8854.
r 2010 American Chemical Society
Published on Web 09/14/2010
pubs.acs.org/IC

9
650 Inorganic Chemistry, Vol. 49, No. 20, 2010
Dixon and Hayward
to remove any lithium-containing phases (LiH and Li
being dried under vacuum.
2
O) before
Characterization. X-ray powder diffraction data were col-
lected using a PANalytical X’Pert diffractometer incorporating
an X’celerator position-sensitive detector (monochromatic Cu
KR1 radiation). Data were collected from air-sensitive samples
under an inert atmosphere using a homemade gas-tight sample
holder. Electron diffraction patterns were collected from sam-
ples supported on lacy carbon grids (deposited from suspension
in chloroform) using a JEOL 2000FX microscope operating at
200 kV. Neutron powder diffraction data were collected in the
temperature range 5 < T/K< 300 from samples contained in
vanadium cans, sealed under argon with indium washers, at a
˚
wavelength of λ = 1.59 A using the D2b diffractometer (ILL
neutron source, Grenoble, France). Additional data sets were
collected in the temperature range 300<T/K<388 using the
POLARIS diffractometer (ISIS neutron source, U.K.) from
samples contained in vanadium cans sealed under argon using
a copper gasket. Rietveld profile refinement was performed
1
0
using the GSAS suite of programs. Average iron oxidation
states in all phases were determined by iodiometric titration.
Approximately 50 mg of material was placed in a three-necked
flask under an argon atmosphere. The solid was then dissolved
in 1.1 M HCl containing an excess of KI, and the liberated I was
2
2 2 3
titrated with previously standardized Na S O under an argon
Figure 1. The crystal structures of the n=2 Ruddlesden-Popper phases
Sr Fe Cl and Sr Fe
3
O
2 5
2
3
2 7
O .
square-planar Fe(II) centers, providing a structural contrast
to the recently reported phases SrFeO and Sr Fe O , which
2
3
2
5
5,6
contain similar iron coordination sites.
purge to prevent oxidation by the air. Titrations were repeated
three times or until a consistent result was obtained.
Experimental Section
Preparation of Sr
3
Fe
2
O
5
Cl
2
. Samples of Sr
3
Fe
2
O
5
Cl
2
were
prepared by a direct-combination route that has previously been
Results
7
described by Weller et al. Suitable quantities of SrO (prepared
Reactivity of Sr Fe O Cl . X-ray powder diffraction
3 2 4 2
data collected from the products of the reaction between
Sr Fe O Cl and LiH reveal that at temperatures below
by the decomposition of SrCO
SrCl
(dried at 180 °C under vacuum), and Fe
9.99%) were thoroughly mixed with an agate pestle and mortar
in an argon-filled glovebox (O and H O levels<1 ppm). The
3
at 1100 °C under vacuum),
O
3
(Alfa Aesar,
3
2
5
2
2
2
9
350 °C no reaction occurs. Reactions performed at tem-
peratures above 400 °C result in the decomposition of the
ternary oxychloride phase and the formation of Sr FeO Cl,
2
2
mixture was then heated in an evacuated silica ampule at 850 °C
for 2 periods of 24 h with one intermediate grinding. X-ray
powder diffraction data collected from this material were con-
sistent with a single phase with lattice parameters a = 3.947(1) A
˚
and c = 23.786(1) A in good agreement with previously reported
2
3
elemental iron, and LiFeO . In the temperature range
2
3
50<T/°C<400, reactions resulted in the formation of a
body-centered tetragonal phase consistent with the topo-
tactic reduction of Sr Fe Cl
˚
7
3
2
O
5
.
2
˚
˚
values (a = 3.946(1) A, c =23.786(1) A).
Reduction of Sr Fe Cl
performed using LiH as a solid state reducing agent. Small
Fe Cl were ground together in a 1:4
Structural Characterization. X-ray powder diffraction
data collected from the washed product of the reaction
between Sr Fe O Cl and LiH at 350 °C could be readily
3
O
2 5
2 3 2 5 2
. The reduction of Sr Fe O Cl was
8
samples (∼300 mg) of Sr
3
2
O
5
2
3
2
5
2
molar ratio with LiH in an argon-filled glovebox. These mix-
tures were then sealed under vacuum in Pyrex ampules and
heated at temperatures between 210 and 400 °C to monitor the
temperature dependence of the reduction reaction. Due to the
hazards associated with the production of hydrogen gas when
indexed on the basis of a body-centered tetragonal unit
˚
cell (a=4.009(1) A, c=22.638(1) A) consistent with the
˚
topotactic reduction of Sr Fe O Cl to a phase of com-
2
position Sr Fe O Cl . In contrast to the X-ray powder
3
3
2
5
2
5-x
2
diffraction data, neutron powder diffraction data col-
lected at room temperature from Sr Fe O Cl exhib-
ited a series of large d-spacing diffraction features which
are inconsistent with the simple body-centered tetragonal
cell described above (Figure 2). These additional diffrac-
tion features could be indexed using a unit cell related to
the X-ray cell by an a = 2a, b = 2b, c =c geometric
expansion. In order to confirm the size and symmetry
of the crystallographic unit cell, electron diffraction data
were collected. As can be seen in Figure 3, these data
could be readily indexed using the body-centered tetra-
gonal cell determined from the X-ray powder diffraction
data, suggesting the additional features observed in the
neutron diffraction data are due to magnetic order.
This was further confirmed by the observation that the
8
using LiH as a reducing agent, large-scale samples suitable
for characterization by neutron powder diffraction were pre-
3
2
5-x
2
pared by means of a spring-loaded venting apparatus described
was mixed in a
9
previously. Approximately 4 g of Sr
3
Fe
2
O
5
Cl
2
1
2
:4 molar ratio with LiH and heated at 350 °C for 4 periods of
4 h with intermediate grinding. The sample was then heated for
a further 48 h in an evacuated Pyrex ampule. After reaction, the
sample was washed under nitrogen with 4 ꢀ 100 mL of methanol
√
√
0
0
0
(
5) Tsujimoto, Y.; Tassel, C.; Hayashi, N.; Watanabe, T.; Kageyama, H.;
Yoshimura, K.; Takano, M.; Ceretti, M.; Ritter, C.; Paulus, W. Nature 2007,
50, 1062–1065.
6) Hiroshi, K.; Takashi, W.; Yoshihiro, T.; Atsushi, K.; Yuji, S.;
Kazuyoshi, K.; Kazuyoshi, Y.; Naoaki, H.; Shigetoshi, M.; Mikio, T.;
Monica, C.; Werner, P.; Clemens, R.; Gilles, A. Angew. Chem. 2008, 47, 5740–
4
(
5
745.
(
7) Dann, S. E.; Weller, M. T.; Currie, D. B. J. Solid State Chem. 1992, 97,
1
79–185.
(
8) Adkin, J. J.; Hayward, M. A. Inorg. Chem. 2008, 47, 10959–10964.
(10) Larson, A. C.; Von Dreele, R. B. General Structure Analysis System,
Los Alamos National Laboratory Report LAUR 86-748, Los Alamos National
Laboratory: Los Alamos, NM, 2000.
(
9) O’Malley, M.; Lockett, M. A.; Hayward, M. A. J. Solid State Chem.
2
007, 180, 2851–2858.

Article
Inorganic Chemistry, Vol. 49, No. 20, 2010 9651
3 2 4 2
Table 1. Refined Structural Parameters of Sr Fe O Cl at 388 K
Wyckoff
position
3
˚
Uiso (A )
atom
x
y
z
Sr(1)
Sr(2)
Fe(1)
O(2)
Cl(1)
2a
4e
4e
8g
4e
0
0
0
0.5
0
0
0
0
0
0
0
0.0036(5)
0.0038(4)
0.0028(2)
0.0058(2)
0.0087(2)
0.153(1)
0.425(1)
0.078(1)
0.294(1)
˚
˚
Sr
3
2
Fe O Cl
cell volume = 364.01(1) A ; weight fraction = 91.9(1)%
4
2
: space group, I4/mmm; a = 4.008(1) A; c = 22.653(1) A;
˚
3
˚
Fe: space group, Im3m; a = 2.866(1) A; cell volume =
3
3.554(3) A ; weight fraction = 8.1(1)%
χ = 1.858, wR = 2.14%, R = 3.77%
p p
˚
2
2
Table 2. Selected Bond Lengths and Bond Valence Parameters from the Refined
Structure of Sr Fe Cl
3
2
O
4
2
˚
bond length (A)
cation
Fe(1)
anion
BVS contribution
total BVS
2.002
O(2)
Cl(1)
O(2)
O(2)
Cl(1)
4 ꢀ 2.006(1)
1 ꢀ 2.977(2)
8 ꢀ 2.674(1)
4 ꢀ 2.632(2)
4 ꢀ 3.072(1)
4 ꢀ 0.479
1 ꢀ 0.079
8 ꢀ 0.223
4 ꢀ 0.249
4 ꢀ 0.219
1 ꢀ 0.158
Sr(1)
Sr(2)
1.780
2.031
1
ꢀ 3.192(2)
inspection of the data revealed a set of weak diffraction
reflections consistent with the presence of metallic iron.
A structural description of this phase was added to the
refinement model, which readily converged to give a good
statistical fit to the data (χ =1.858). Full details of the refined
structural parameters are given in Table 1 with selected
Figure 2. Observed, calculated, and difference plots from the structural
refinement of Sr Fe O Cl at 388 K (top) and structural and magnetic
3 2 4 2
refinement at 298 K (bottom) against neutron powder diffraction data.
The major magnetic reflections are marked with arrows.
2
bond lengths from the refined structure of Sr Fe O Cl
4
3
2
2
shown in Table 2. Plots of the observed and calculated
data are shown in the Supporting Information. In order
to confirm the refined composition, iodiometric titrations
were performed, the results of which are consistent with an
average iron oxidation state of þ2 for the bulk phase. Due
to the highly air-sensitive nature of Sr Fe O Cl , it was
3
2
4
2
not possible to confirm the stoichiometry of the phase
using reoxidative thermogravimetric analysis. Indeed,
bulk samples of this material are pyrophoric, and extreme
care should be taken when handling this material.
Magnetic Characterization. Neutron powder diffrac-
tion data collected at 5 K show additional features rela-
tive to the 388 K data set which were assigned to magnetic
order, as described above. A series of magnetic ordering
models consistent with the crystallographic symmetry
were compared to the 5 K data, and the best fit was
obtained by a simple G-type antiferromagnetic model
with moments aligned parallel to the z axis (Figure 4a).
The refinement converged readily to yield an ordered
Figure 3. Electron diffraction data collected from the [001] and [010]
3 2 4 2
zone axes of Sr Fe O Cl consistent with a tetragonal unit cell: a =
˚ ˚
4.008 A, c = 22.653 A.
intensities of these additional neutron diffraction features
diminish with increasing temperature such that, at 388 K,
their intensity was zero (Figure 2), consistent with a loss
of magnetic order at this temperature.
To avoid the complications of simultaneous structu-
ral and magnetic refinements, the structural refinement of
Sr Fe O Cl was performed against neutron powder
moment of 2.81(9) μ per iron center at 5 K. Complete
B
3
2
5-x
2
details of the magnetic refinement and plots of observed
and calculated data are shown in the Supporting Infor-
mation. Analogous models were refined against neutron
diffraction data collected in the temperature range 5 <
T/K<388. Figure 4b shows a plot of the refined ordered
moment as a function of the temperature, indicating a
magnetic ordering temperature of TN ∼ 378(10) K for
Sr Fe O Cl . Due to the presence of metallic iron in the
diffraction data collected at 388 K. A structural model
based on that of Sr Fe O Cl was refined against this
3
2
5
2
data set. During the refinement, all atomic positional and
thermal parameters were allowed to vary, and to account
for the nonstoichiometry of the reduced phase, the frac-
tional occupancies of the anion sites were also allowed to
vary. The occupancy of the O(1) “bridging axial” anion
site (Figure 1) rapidly declined to zero within error with
an associated improvement in the fit to the data. In the
final stages of the refinement, the fractional occupancies
of all of the atomic sites were allowed to vary, but all
other atomic positions remained fully occupied within
error, indicating a composition of Sr Fe O Cl . Close
3
2
4
2
sample, it was not possible to confirm the magnetic order-
ing temperature from susceptibility measurements.
Discussion
The topotactic reduction of Sr Fe O Cl with LiH results
3
2
5
2
3
2
4
2
in the deintercalation of oxide ions from the SrO layer of the

9
652 Inorganic Chemistry, Vol. 49, No. 20, 2010
Dixon and Hayward
Figure 4. (a) The G-type antiferromagnetic model refined for Sr
Cl . (b) A plot of the refined magnetic moment per iron center as a
function of the temperature for Sr Fe Cl , indicating a magnetic
ordering temperature of T
∼ 378 K.
3 2
Fe -
O
4
2
3
O
2 4
2
N
material and the formation of Sr Fe O Cl , as shown in
3
1
2
4
2
1
Figure 5. Bond valence sums (BVS) calculated for the iron
centers in Sr Fe O Cl (Table 2) are consistent with the pre-
3
2
4
2
sence of Fe(II). In addition, they reveal that the long Fe-Cl
˚
bond (2.977(2)A) contributes only 4.2% to the BVS total of
the iron center, so the iron coordination geometry is effec-
tively square-planar. There are relatively few examples in the
literature of square-planar Fe(II) centers in extended solids
for comparison. Those that are reported have been prepared
by similar topotactic reduction syntheses as Sr Fe O Cl .
3
2
4
2
SrFeO for example has been prepared by the topotactic
2
5
Figure 5. Reduction of Sr Fe O Cl and Sr Fe O . The reduction of
3
2
5
2
3
2 7
reduction of SrFeO3 with CaH . Comparison of the
-x
2
Sr
sites to yield Sr
the removal of ions from bridging axial sites to form Sr Fe O . A sub-
Fe O Cl
3 2 5 2
proceeds via the removal of oxide ions from bridging axial
structural parameters of Sr Fe O Cl and SrFeO (Figure 6)
3
2
4
2
2
3
Fe Cl . The reduction of Sr Fe proceeds initially by
2
O
4
2
3
2 7
O
reveals a remarkable similarity in the in-plane Fe-O bond
3
2 6
5
˚
˚
lengths of 2.006(1) A and 1.996(1) A, respectively.
Structural Selectivity. The reduction of Sr Fe O Cl
sequent reduction step which involves the deintercalation of oxide ions
from equatorial sites and a rearrangement of the anion lattice leads to the
3
2
5
2
13
3 2 5
formation of the spin-ladder phase Sr Fe O .
with LiH removes oxide ions exclusively from the axial
AO layers of the host structure. This selectivity is in direct
contrast to that shown by a number of other AO(ABO3)n
Ruddlesden-Popper oxides such as LaSrCoO and Sr -
4
3
Mn O . During the topotactic reduction of these phases,
2
7
anions are selectively deintercalated from the “equatorial”
BO2 layers of their structures.
1
2,13
These contrasting
-x
selectivities can be rationalized by considering the different
coordination requirements of the different cations in each
system. Following the logic described previously, and
9
taking Sr Fe O Cl as an example, it is clear that in order
3
2
5
2
to best satisfy the competing coordination requirements
of both strontium and iron cations on reduction, oxide
ions will be removed from sites which form the strongest
bonds to the iron and the weakest bonds to the strontium
sites. We can evaluate this idea using bond valence sums
1
1
(
BVS) as has been demonstrated previously for the
9
Sr Mn O phases. BVS calculations can give a crude
15-x
7
4
approximation of the cation-anion bond strengths in a
solid and can thereby determine which anion is most
likely to be deintercalated upon reduction. Contributions
made by the iron and strontium cations to the BVS of the
different anion sites in Sr Fe O Cl are shown in Table 3.
Figure 6. The structures of the topotactically reduced phases Sr
Cl and SrFeO , showing the different separations between the FeO
layers in each phase.
3 2
Fe -
O
4
2
2
2
oxide ions will be removed from the O(1) “bridging axial”
anion site on reduction in agreement with observation.
Analogous calculations for LaSrCoO and Sr Mn O
7
3
2
5
2
The largest value of Fe BVS - Sr BVS clearly predicts that
4
3
2
shown in the Supporting Information correctly predict the
opposite selectivity, which is observed experimentally.
(
11) Brese, N. E.; O’Keeffee, M. Acta Crystallogr., Sect. B: Struct. Sci.
991, B47, 192–197.
12) Hayward, M. A.; Rosseinsky, M. J. Chem. Mater. 2000, 12, 2182–
195.
12,13
1
The topotactic reduction of the all-oxide analogue
of Sr Fe O Cl , Sr Fe O , provides another instructive
(
2
3
2
5
2
3
2
7

Article
Inorganic Chemistry, Vol. 49, No. 20, 2010 9653
Table 3. Bond Valence Sums Calculated for the Anion Sites of Sr
3
Fe
2
O
5
Cl
2
More generally, the removal of chloride ions from a
complex oxyhalide would presumably occur via reaction 1,
leading to the formation of LiCl, compared to reaction 2
for the removal of oxide ions, which leads to the formation
O(1)
O(2)
Cl(1)
0.086
Fe BVS
Sr(1) BVS
Sr(2) BVS
Fe BVS - Sr BVS Total
1.474
0.649
1.068
0.235
0.636
0.197
of Li O as a byproduct.
2
1.155
-1.069
0.825
Sr3Fe2O5Cl2 þ x LiH f Sr3Fe2O5Cl2- x
3
x
þ x LiCl þ
3 H
ð1Þ
comparison. As shown in Figure 5, the reduction of this
phase occurs via a two-step process. Initially, oxide ions
are deintercalated from the “bridging axial” anion sites in
the host structure (in agreement with BVS predictions,
as shown in the Supporting Information), leading to the
formation of Sr Fe O , a structural analogue of Sr Fe -
3
2
2
Sr3Fe2O5Cl2 þ 2x LiH f Sr3Fe2O5- xCl2
3
þ x Li2O þ x H2
Given that the conversion of LiH to Li
ð2Þ
O or LiCl
3
3
3
2
6
3
2
2
O Cl . However, Sr Fe O can be further reduced to
6
provides the thermodynamic driving force for the topo-
tactic reduction reactions, the greater lattice energy of
4
2
3
2
a phase of composition Sr Fe O . BVS calculations,
3
2
5
-
1
Li O compared to LiCl (-2799 kJ mol vs -834 kJ
detailed in the Supporting Information, predict this
reduction should occur via the removal of oxide ions
from “equatorial” anion sites. However, the reduction is
observed to progress with the removal of some equato-
rial oxide ions and a rearrangement of the anion lattice
to yield a ladder-type structure incorporating pseudo-
2
-
1 14
)
mol
is clearly a deciding factor for the course of the
reaction. This further suggests that in general we would
expect to preferentially deintercalate oxide ions, rather
than chloride ions, from the majority of complex oxyha-
lide hosts.
6
Magnetic Behavior. Neutron powder diffraction data
show that Sr Fe O Cl adopts a G-type antiferromag-
square-planar Fe(II) centers, as shown in Figure 5. This
3
2
4
2
anion lattice rearrangement suggests that the square-
planar Fe(II) coordination formed is energetically prefer-
able to the “tetrahedral” coordination that would result
from the simple topotactic deintercalation of oxide ions
from the equatorial layers of Sr Fe O . While it is clear
netic arrangement below TN ∼ 378 K, exhibiting an
ordered moment of 2.81(9) μ per iron center at 5 K.
B
The magnitude of the ordered moment is consistent with
a d high-spin electronic configuration expected for Fe(II)
6
3
2
6
(
S=2, μordered_exp.=4 μ ) with a reduction in the observed
B
that Sr Fe O and Sr Fe O Cl adopt strongly related
2
3
2
6
3
2
4
moment due to covalency within the iron-anion contacts
and/or slight disorder within the magnetic system. As a
comparison, the isoelectronic and structurally similar
structures, there is no evidence for a related reduction/
rearrangement in the oxyhalide phase. This is unsurpris-
ing because the corresponding reduction would require
the formation of an iron phase with an average oxidation
state of below Fe þ2. To the best of our knowledge, there
are no known examples of a topotactic reduction result-
ing in the formation of iron with such an oxidation state.
In addition, an analogous rearrangement of “Sr Fe O Cl ”
infinite-layer phase SrFeO (Figure 6) has an ordering
2
temperature of TN ∼ 473 K and an ordered moment of
5
3
.6 μ per iron center at 10 K.
B
Neutron powder diffraction data also show that Sr3-
Fe O Cl retains tetragonal symmetry on cooling to 5 K.
2
4
2
3
2
3
2
This indicates that in common with SrFeO , Sr Fe O Cl
2
2
3
2
4
to a ladder-type structure is clearly less feasible than that
of Sr Fe O , as this would leave the strontium cations
does not have a degenerate electronic ground state, as
3
2
5
such a state would be susceptible to a Jahn-Teller-like
within the SrCl rock salt layers (Sr(2)) significantly
under bonded” in large O Cl coordination sites.
Chemical Selectivity. Complex transition metal oxyha-
1
5
distortion. This observation and the similarity of the
iron coordination environments in the two phases make it
reasonable to assume that the iron centers in Sr Fe O Cl
2
“
2
5
lides provide an opportunity to evaluate the ability of the
binary metal hydrides to selectively deintercalate different
anion types. As described above, it is clear that during the
reduction of Sr Fe O Cl there is complete selectivity for
3
2
4
will adopt the same local orbital configuration as those in
SrFeO . Previous reports on the electronic structure of
SrFeO indicate that the antiferromagnetic S = 2 ground
2
2
3
2
5
2
state contains Fe(II) centers with an electronic configura-
2
the removal of oxide ions over chloride ions. This
selectivity can be rationalized on the basis of two factors.
In the specific case of Sr Fe O Cl , it can be seen from
2
2
1
tion of: (d ) (d , d ) (d ) (d ) .
1 5,15
2
z2
xz yz
xy
x -y
Considering the
1
6
Goodenough-Kanamori rules, it can be seen that the
transition metal d orbital expected to have the strongest
3
2
5
2
Figure 1 that each Sr(2) cation resides in a nine-coordinate
coordination site of stoichiometry O Cl . BVS calcula-
tions performed for the Sr Fe O Cl structure (Table 3)
contribution to the superexchange interactions is the
2
4
5
1
d
(d
2
orbital. Thus, the 180 °Fe (d
2
2
) - O (2p) - Fe
superexchange interaction accounts for the
3
2
5
2
x
-y
x
-y
1
indicate that the chloride ions contribute only 4.2% of the
total valence sum of the iron centers, but are “bonded”
much more strongly to the Sr(2) cation. Removal of chlo-
ride ions from the structure of Sr Fe O Cl as part of the
x
2
-y
2
)
strong antiferromagnetic coupling between iron centers
within each FeO layer of Sr Fe Cl
Coupling between FeO layers is more complex. In
the case of adjacent FeO layers within SrFe blocks
O
.
2
2
3
2
4
3
2
5
2
2
reduction process would therefore have little effect on the
iron centers but leave the coordination sphere of the Sr(2)
cations exceptionally anion-deficient, contrary to the
ideas of structural selectivity described above.
2
O
2 4
(14) Handbook of Chemistry and Physics, 89th ed.; CRC Press: Boca Raton,
FL, 2008.
15) Xiang, H. J.; Wei, S.-H.; Whangbo, M. H. Phys. Rev. Lett. 2008, 100,
67207.
(16) Goodenough, J. B. Magnetism and the Chemical Bond; Wiley: New York,
1963.
(
1
(
13) Gillie, L. J.; Wright, A. J.; Hadermann, J.; Van Tendeloo, G.; Greaves,
C. J. Solid State Chem. 2003, 175, 188–196.

9
654 Inorganic Chemistry, Vol. 49, No. 20, 2010
Dixon and Hayward
2
2
(
tions result in G-type antiferromagnetic ordering. How-
Figure 6), direct Fe (dz
2
) -Fe (dz2) exchange interac-
magnetic ordering temperature when comparing LaBaMnO3.5
to LaSrMnO3.5 associated with the decrease in interlayer
separation on the substitution of barium for strontium.
1
7
ever, in order for the FeO layers separated by SrCl rock
2
salt layers to couple (Figure 6), a through-space mecha-
nism must be employed if full three-dimensional anti-
ferromagnetic order is to occur. The lower ordering
temperature and smaller ordered moment observed for
Sr Fe O Cl , compared to SrFeO , suggests that this
Conclusion
In conclusion, we have demonstrated that binary metal
hydrides can be used to selectively deintercalate oxide ions
from complex transition metal oxychlorides. In addition, it
can be seen that the topotactic reduction reactions of oxy-
chlorides and isostructural complex oxides proceed to differ-
ent extents to give isoelectronic phases with significantly
different anion lattices (e.g., the two Fe(II) phases Sr Fe -
3
2
4
2
2
through-space coupling is a limiting factor. The insertion
of SrCl layers into the structure of Sr Fe O Cl to form
3
2
4
2
isolated double layer blocks and a more two-dimensional
structure (Figure 6) leads to a reduction in the strength
of the magnetic interactions between these double layer
blocks (compared to the regularly spaced FeO layers
2
within the structure of SrFeO ) and an associated low-
2
3
2
O Cl and Sr Fe O (Figure 5)). This suggests that by careful
4
2
3
2
5
selection of substrate phases, the topotactic reduction of
complex transition metal oxychlorides can lead to the pre-
paration of novel anion-deficient phases with unique transi-
tion metal-oxygen sublattices which cannot be prepared via
the reduction of all oxide substrates.
ering of both the magnetic ordering temperature and
ordered moment.
The influence of dimensionality can also be seen in
the magnetic ordering behavior of another isoelectronic
phase, Sr Fe O . As shown in Figure 5, Sr Fe O also
3
2
5
3
2
5
Acknowledgment. We thank R. Smith and E. Suard
for assistance in collecting the neutron diffraction data.
Experiments at the ISIS pulsed neutron facility were sup-
ported by a beam time allocation from the Science and
Technology Facilities Council. E.D. thanks the EPSRC
for a studentship.
contains Fe(II) centers in pseudo-square-planar coordi-
nation. However, in contrast to SrFeO and Sr Fe O Cl ,
Sr Fe O adopts a spin-ladder-type structure in which the
2
3
2
4
2
3
2
5
dominant magnetic exchange pathways can be considered
one-dimensional. Thus, it is no surprise that while Sr3-
Fe O adopts a similar magnetic structure to Sr Fe O Cl
2
2
5
3
2
4
6
it has a lower ordering temperature (“below 298 K”).
Supporting Information Available: A complete description of
the ordered magnetic model of Sr Fe O Cl . Observed, calcu-
Similar correlations between dimensionality and magne-
tic ordering temperatures are common in Ruddlesden-
Popper systems, as exemplified by the increase in 3D
3
2
4
2
lated, and difference plots from the structural refinement at
388 K, and the structural and magnetic refinement at 5 K of
Sr Fe O Cl . Bond valence sum predictions for LaSrCoO ,
3
2
4
2
4
(
17) Kitchen, H. J.; Saratovsky, I.; Hayward, M. A. Dalton Trans. 2010,
Sr Mn O , Sr Fe O , and Sr Fe O . This material is available
3 2 7 3 2 7 3 2 6
3
9, 6098–6105.
free of charge via the Internet at http://pubs.acs.org.