[Fe24Se26]: A Host-Guest Compound with Unique Fe-Se Topology

Article abstract of DOI:10.1002/chem.201504840

The novel host-guest compound [Cs₆Cl][Fe₂₄Se₂₆] (I4/mmm; a=11.0991(9), c=22.143(2) ?) was obtained by reacting Cs₂Se, CsCl, Fe, and Se in closed ampoules. This is the first member of a family of compounds with unique Fe-Se topology, which consists of edge-sharing, extended fused cubane [Fe₈Se₆Se_8/3] blocks that host a guest complex ion, [Cs₆Cl]⁵⁺. Thus Fe is tetrahedrally coordinated and divalent with strong exchange couplings, which results in an ordered antiferromagnetic state below T_N=221 K. At low temperatures, a distribution of hyperfine fields in the M?ssbauer spectra suggests a structural distortion or a complex spin structure. With its strong Fe-Se covalency, the compound is close to electronic itinerancy and is, therefore, prone to exhibit tunable properties.

Full text of DOI:10.1002/chem.201504840

DOI: 10.1002/chem.201504840
Full Paper
&
Magnetic Properties
[Cs6Cl][Fe24Se26]: A Host–Guest Compound with Unique Fe–Se
Topology
Martin Valldor,*[a] Bodo Bçhme,[a] Yurii Prots,[a] Horst Borrmann,[a] Peter Adler,[a]
Walter Schnelle,[a] Yves Watier,[b] Chang Yang Kuo,[a] Tun-Wen Pi,[c] Zhiwei Hu,[a]
Claudia Felser,[a] and Liu Hao Tjeng[a]
Abstract: The novel host–guest compound [Cs6Cl][Fe24Se26]
(I4/mmm; a=11.0991(9), c=22.143(2) Š) was obtained by re-
acting Cs2Se, CsCl, Fe, and Se in closed ampoules. This is the
first member of a family of compounds with unique Fe–Se
topology, which consists of edge-sharing, extended fused
cubane [Fe8Se6Se8/3] blocks that host a guest complex ion,
[Cs6Cl]5+. Thus Fe is tetrahedrally coordinated and divalent
with strong exchange couplings, which results in an ordered
antiferromagnetic state below TN =221 K. At low tempera-
tures, a distribution of hyperfine fields in the Mçssbauer
spectra suggests a structural distortion or a complex spin
structure. With its strong Fe–Se covalency, the compound is
close to electronic itinerancy and is, therefore, prone to ex-
hibit tunable properties.
[9,10]
Introduction
later revised to K6ÀxFe24ÀyS26(S,Cl)1Àz
.
Bartonite is structurally
related to djerfisherite [K6Cl][Li][Fe24S26][11] and pentlandite
[12]
The chemistry of heavier chalcogenides has made a giant leap
forward since the discovery of superconductivity in Fe1+xSe.[1]
This binary belongs to a group of emerging compounds with
tetrahedral coordination of Fe2+ by chalcogenides with consid-
erable electronic itinerancy, which leads to interesting physical
phenomena: Spin- or charge-density waves observed in
Fe1+xSe[2,3] are discussed as being closely related to the super-
conductivity that appears at relatively lower temperatures (Tc).
Recently, it was discovered that Tc of FeSe can be increased
from 8 to 46 K by interleaving monolayers with molecular[4] or
ionic[5] spacers, to form homogeneous interfaces between FeSe
and insulators. The latter also falls into the concept of host–
guest compounds, in which the layers of FeSe host the polariz-
able intercalated guest ions and molecules, as was discussed
for the mineral tochilinite, which is based on Fe1ÀxS.[6] Obvious-
ly, new topologies of Fe–Se-based materials have to be further
explored, in search of novel interfaces and lattices that are
prone to exhibit exotic electronic states.
(Ni,Fe)9S8 through a common fused-cubane building block,
[Fe8S6]. The close relation between this fused-cubane cluster
and the active site in the enzyme nitrogenase, which catalyzes
the essential nitrogen fixation reaction in nature,[13,14] was al-
ready realized by Pohl and Opitz.[15] Strong electron correla-
tions dominate in these iron sulfide-based systems, for exam-
ple, the minerals are insulators as indicated by their yellow-
brown color.[9,10] To enhance the chance of electronic itinerancy
in such fused-cubane-based structures, S2À can be replaced
with Se2À, which is anticipated by the chemical analogy be-
tween the two chalcogenides. Herein, we report the first suc-
cessful synthesis of Se-substituted bartonite, [Cs6Cl][Fe24Se26],
a host–guest compound with a novel Fe–Se framework, to-
gether with investigations of its fundamental physical proper-
ties.
Results
A few decades ago, the mineral bartonite[7] was first discov-
ered and found to have variable composition,[8] which was
Synthesis and chemical composition
The obtained polycrystalline sample and the small single crys-
tals were black and elemental analyses (EDX) on ten crystallites
from the powder sample gave an average composition of
Cs6.3(4)Cl1.07(8)Fe25(1)Se25.1(6). Within analytical errors, this result
agrees well with the nominal composition and the results of
the crystal structure determination.
[a] Dr. M. Valldor, Dr. B. Bçhme, Dr. Yu. Prots, Dr. H. Borrmann, Dr. P. Adler,
Dr. W. Schnelle, Dr. C. Y. Kuo, Dr. Z. Hu, Prof. Dr. C. Felser,
Prof. Dr. L. H. Tjeng
Max Planck Institute for Chemical Physics of Solids
01187 Dresden (Germany)
E-mail: martin.valldor@cpfs.mpg.de
[b] Dr. Y. Watier
ESRF, 71 Avenue des Martyrs, 38000 Grenoble (France)
Structural analysis
[c] Dr. T.-W. Pi
The crystal structure of the title compound (Figure 1) consists
of a host lattice built from octamers of fluorite-like edge-shar-
ing FeSe4 tetrahedra (Figure 2a–c), which form two crystallo-
National Synchrotron Radiation Research Centre, Hsinchu 30076 (Taiwan)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201504840.
Chem. Eur. J. 2016, 22, 4626 – 4631
4626
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
on the basis of the nested polyhedra units approach[17] and re-
garded as a hierarchical derivate of the TiAl3-type,[16] in which
the [Cs6Cl] units occupy the Ti and the 3[Fe8Se6Se8/3] net occu-
1
pies the Al positions. In turn, the TiAl3-type constitutes a super-
structure of fcc Cu[18] or a stuffed In-type crystal structure.[19]
From the expected charges of Cs, Cl, and Se and by assum-
ing an ideal composition, a charge deficit on Fe is apparent.
This can be balanced by 98% occupancy of the Fe sites in the
host lattice; a similar situation is discussed for the closely relat-
ed djerfisherite compound.[11] Alternatively, if all Fe sites are
fully occupied, one delocalized electron could be distributed
per three octamers. Or this charge deficit could also be ac-
counted for if Cl was completely replaced with O that enters
as an impurity in the solid-state reaction from the Fe metal.
However, the expected CsÀO separation is about 2.86 Š, as
taken from Cs2O[20] and Cs11[Fe5S8]2O,[21] in which O is octahe-
drally coordinated by Cs. Here the corresponding separation is
3.26 Š on average (Figure 2d). This is almost identical to the
CsÀCl separation in the host–guest compound [Cs6Cl][RE21Ch34]
(RE=Dy, Ho, Ch=S, Se, Te) of 3.27 Š,[22] in which the charges
add up properly with only RE+3. Moreover, Cl was identified in
fitting relative amounts, as estimated from EDX data. Thus,
a minor iron deficiency in the title compound is the most
probable explanation, but cannot be unambiguously resolved
with the diffraction data at hand. As estimated by a synchro-
tron powder X-ray diffraction measurement (Figure S2 in the
Supporting Information), the powder sample is relatively pure.
Also, diffraction data obtained at 100 K suggest that the body-
centered tetragonal symmetry persists even at relatively low
temperatures.
Figure 1. Left: The complete [Cs6Cl][Fe24Se26] crystal structure, in which the
[Cs6Cl]5+ octahedra and the 13[Fe8Se6Se8/3] units are highlighted. For better
visualization, the latter (octamers) are shown in two colors to separate Fe1
(orange) and Fe2 (dark green). Right: The crystal structure of TiAl3 is
shown for comparison.
[16]
graphically distinct extended fused cubane [Fe8Se6Se8/3] units.
Each octamer of one type, Fe1(2), is connected to eight(four)
octamers of the other type (Fe2(1)). Layers of each octamer are
found perpendicular to the unique crystallographic axis. In
a simplified description, half of the possible Fe1 octamers are
replaced with [Cs6Cl] octahedra (Figure 2d) in a checkerboard
fashion, which is reversed in the neighboring layers to manifest
the body centering (Figure 1). The structure can be described
X-ray spectroscopy
The Fe-L2,3 XAS spectrum of [Cs6Cl][Fe24Se26] lies at the same
energy as that of Fe2+ reference Fe0.04Mg0.96O,[23] and is shifted
by more than 1 eV to a lower energy with respect to Fe3+ ref-
erence Fe2O3, which demonstrates the Fe2+ valence state in
[Cs6Cl][Fe24Se26] (Figure 3). The similar multiplet spectral fea-
tures of [Cs6Cl][Fe24Se26] and Fe0.04Mg0.96O further indicates the
same Fe2+ valence state, however, the features of the title
compound are much broader. We ascribe this to the very
strong Fe–Se covalence, not so much due to a large transfer
integral but due to the relatively small Fe3d–Se4p charge-
transfer energy, which makes these states practically degener-
ate. We note that the Fe–Se transfer integral is expected to be
rather similar to that of Fe–O or Fe–S as a consequence of the
larger Se atomic radius being compensated by the larger
FeÀSe separation.
Magnetism
The temperature-dependent magnetic susceptibility (c) reveals
that the title compound does not behave as a normal para-
magnet even at 750 K, which indicates relatively strong spin–
spin interactions (Figure 4). In the temperature range of 240 to
200 K, an anomaly in c is observed (Figure 4, upper right
inset). In combination with data presented later, this suggests
Figure 2. Crystal structure building blocks of [Cs6Cl][Fe24Se26]. a)–c) The parts
of an [Fe8Se6Se8/3] extended fused cubane and d) the Se coordination
around the [Cs6Cl]5+ octahedra. Chosen interatomic distances [Š] are shown
and the dashed lines are a guide for the eye between equivalent atoms.
Chem. Eur. J. 2016, 22, 4626 – 4631
www.chemeurj.org
4627
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
quettes that are superimposed instead of being antiferromag-
netically alternated along the ladders in BaFe2Se3.
57Fe Mçssbauer spectroscopy
The Mçssbauer spectra of [Cs6Cl][Fe24Se26] (Figure 5 and Fig-
ure S3 in the Supporting Information) reveal a single slightly
Figure 3. Fe-L2,3 X-ray absorption spectroscopy (XAS) data of [Cs6Cl][Fe24Se26]
(middle), Fe0.04Mg0.96O (top, as Fe2+ reference), and Fe2O3 (bottom, as Fe3+
reference). The dashed vertical lines represent the main intensity of each
spectrum and are added as a guide for the eye.
Figure 5. Mçssbauer spectra of [Cs6Cl][Fe24Se26] above and below TN. The ex-
perimental data are plotted as dots, the calculated spectra are represented
by black solid lines, and the colored lines correspond to the subspectra.
broadened quadrupole doublet for T>TN, which indicates
rapid spin–spin fluctuations (k>108 sÀ1). The two iron sites
(Fe1, Fe2) are very similar and not resolved. The isomer shifts
(IS; 0.55 mmsÀ1 at RT) are smaller than those for compounds
3
with electronically localized Fe2+ (e3t2 ), such as oxysulfides
and oxyselenides (ISꢀ0.68 mmsÀ1),[26] but larger than for elec-
tronically more itinerant compounds, such as semiconducting
BaFe2S3 (IS=0.41 mmsÀ1)[27] or the metallic superconductor
b-FeSe (IS=0.44 mmsÀ1).[2] This is in line with covalent Fe–Se
bonding as suggested by the XAS results. Also, the quadrupole
splitting (QS=0.58 mmsÀ1) is smaller than expected for a local-
3
ized e3t2 configuration. The surprisingly complex magnetic hy-
perfine pattern at 5 K (T<TN) was simulated by a superposition
of six sextets with equal IS (see Section S3 in the Supporting
Information for details and fit results). The sextets differ in
their magnetic hyperfine fields (Bhf) and quadrupole coupling
parameters (QS). The latter reflect the magnitude of the electric
field gradient (efg) and the orientation between the efg and
Bhf. The complex pattern may indicate a structural phase transi-
tion that involves tilts of FeSe4 tetrahedra, as suggested for
BaFe2Se3,[28] and/or a complex spin structure, such as that
found in the spiral antiferromagnet Sr2FeO4.[29,30] Note that, be-
cause no obvious crystallographic changes were observed at
100 K (Figure S2 in the Supporting Information), the complex
spin structure is more probable. Thus the magnetic ground
state of the title compound might contain spin plaquettes, as
suggested for BaFe2Se3,[25] but with an additional ordering
complexity.
Figure 4. Magnetic susceptibility of [Cs6Cl][Fe24Se26]. The temperature-de-
pendent static-field magnetic susceptibility is shown. Upper inset: Magnifica-
tion of data close to the TN. Lower inset: Field-dependent magnetization
data.
antiferromagnetic ordering with a NØel temperature (TN) of
221 K. The magnetization curve at 300 K indicates about 1%
unreacted a-Fe,[24] which agrees with the Rietveld refinement.
For comparison, the magnetism of the bartonite sulfide miner-
al was only briefly mentioned as “weakly magnetic”,[12] but no
data have been presented so far. Although the magnetic lattice
of BaFe2Se3 is very different (quasi-1D spin-ladders), the mag-
netic behavior is similar to that of the title compound, but
with a TN =256 K.[25] Thus, the magnetic Fe4 plaquettes that are
[25]
suggested in BaFe2Se3 might be relevant in the title com-
pound; the octamers could then be a combination of two pla-
Chem. Eur. J. 2016, 22, 4626 – 4631
www.chemeurj.org
4628
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Resistivity measurements
The covalent bonding character of the Fe–Se host lattice
agrees with the observed electric conductivity. At high temper-
atures, thermally activated conductivity behavior is observed
with small activation energy (Figure 6). Below 221 K, the resis-
Figure 7. Specific heat capacity of [Cs6Cl][Fe24Se26]. Inset: The integrated
area of the Cp/T(T) plot that was used to obtain the magnetic ordering en-
tropy.
easy access to novel compounds with novel Fe–Se lattices, as
presented herein.
The building block [Fe8Se6Se8/3] is a chemical extension
of the known fused cubanes that can be found as complex
ions [Fe8S6I8]4À, in which eight iodine atoms sit at the terminat-
ing corner sites.[15] Fe8S6 is further found in, for example,
djerfisherite,[11] pentlandite,[12] [Fe8S6Cl4(PCy3)4]·3THF,[31] (PCy3 =
Figure 6. Electrical resistivity as a function of inverse temperature at two dif-
ferent magnetic fields. The NØel temperature is indicated by a vertical
broken line and an Arrhenius estimation of the activation energy (dashed
line) has been fitted to the high-temperature data at 0 T.
tricyclohexylphosphine,
THF=tetrahydrofurane),
and
tivity (1) increases as expected for a lattice with antiferromag-
netic spin ordering. Due to strong exchange interactions in the
title compound, a magnetic field has a relatively weak influ-
ence on the conducting electrons, but it increases 1 above TN,
whereas it decreases 1 below TN. Thus, a considerable magne-
to-resistance is observed. The estimated thermal activation
energy gap of 7.6 meV represents a minimum, which is possi-
bly influenced by impurity states. The fundamental thermal ac-
tivation energy might be larger.
(PPN)2[Fe8S6(NO)8]
(PPN=bis(triphenylphospine)iminium),[32]
which emphasizes its commonness. All cubane-like compounds
are of great interest with respect to catalytic activity in the nat-
ural enzyme nitrogenase.[13–15] The title compound proves that
Se-based fused cubane is obtainable, which opens up pros-
pects to alter the catalytic activity of Fe cubane and related
cluster centers.
The strongly ionic guest [A6Cl]5+ (A=Cs, K) occurs within
strongly covalent host lattices in at least three well-character-
ized compounds: bartonite,[9,10] djerfisherite,[11] and [Cs6Cl]
[RE21Ch34] (RE=Dy, Ho, Ch=S, Se, Te).[21] By comparing barton-
ite with [Cs6Cl][RE21Ch34], it is evident that the resulting charge
of the host lattice is important for the overall periodicity of the
[A6Cl]5+ ion. The higher the charge of the chalcogenide-based
host, the shorter the distance between the guest ions, as ex-
pected.
Specific heat capacity
In the temperature range of 170 to 250 K, the specific heat
data exhibit an anomaly with a maximum close to 221 K
(Figure 7). The asymmetry and broadness of this anomaly indi-
cate a second-order phase transition, which confirms the mag-
netic ordering of the title compound. However, close to 221 K
the released spin entropy per iron spin (Figure 7, inset) is
36.7 JmolÀ1 KÀ1/24=1.53 JmolÀ1 KÀ1, which is far from the ex-
pected Rln((2J)+1)=13.4 JmolÀ1 KÀ1 for J=2. Possibly the en-
tropy is released within a larger temperature range than pre-
sumed in this simple estimation.
The crystal structure of the title compound, except for the
part replaced with the guest ion, is a cubic close-packed Se lat-
tice with Fe2+ solely in the tetrahedral voids, which emphasiz-
es its close structural relationship to the superconductor
Fe1+xSe. The main difference is the distribution of Fe2+ ions:
The title compound has Fe2+ ions in eight neighboring tetra-
hedral voids surrounded by a “shell” of unoccupied tetrahedral
voids, whereas Fe1+xSe has Fe2+ in all tetrahedral voids within
layers that are separated by layers of empty tetrahedra.[1] Nev-
ertheless, the FeÀSe separations in the title compounds (Fig-
ure 2b) are remarkably similar to those in Fe1+xSe (2.39 Š).[2]
Also, the FeÀFe separations of about 2.76 to 2.88 Š in the oc-
tamers are not far from those found in Fe1+xSe (2.67 Š),[2]
Discussion
Although the bartonite mineral was suggested to form in
aqueous solution under hydrothermal conditions,[10] the iso-
structural title compound was synthesized from binary precur-
sors and elements. The success of this route ensures relatively
Chem. Eur. J. 2016, 22, 4626 – 4631
www.chemeurj.org
4629
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
K0.8Fe2Se2 (2.77 Š),[33] and Fe pnictides that show superconduc-
tivity.
96 h, before the furnace was switched off. The well-shaped crystals
were mechanically separated from the CsCl residue and were
stable in air for several days.
Evidently, [Cs6Cl][Fe24Se26] is an antiferromagnetic insulator
and not a superconductor. Yet we can infer that the compound
is very close to being a conductor; its bandgap must be rather
small. Although the activation energy, as deduced from the re-
sistivity measurements, may not be a reliable indicator, data
from other measurements do provide strong hints: 1) the
broad XAS features suggest that the Fe3d and Se4p orbitals
are degenerate in energy, which facilitates the hopping of
charge carriers, 2) the Mçssbauer parameters refute a well-lo-
X-ray diffraction
Single-crystal data were obtained by using a Rigaku AFC7 with
a CCD camera as the detector (Saturn 724+), and a MoKa X-ray
source (l=0.71073 Š). The empirically absorption-corrected data
(multi-scan) were further treated by using JANA2006.[34] A structur-
al model was found by using charge flipping in Superflip. In the
final refinement, all atomic positions were refined as fully occupied
and with harmonic anisotropic displacement parameters. Details
can be found in the Supporting Information (Section S1).
3
calized e3t2 electronic configuration, and 3) the high NØel tem-
perature together with the absence of a Curie–Weiss behavior
up to 750 K imply relatively strong exchange or super-ex-
change interactions and, therefore, rather low virtual excitation
energies. Therefore, we expect that pressure, a decrease in the
FeÀFe separation, or doping can easily bring the title com-
pound into itinerancy with suppressed magnetism and per-
haps even superconductivity.
Powder X-ray data were obtained at the high-resolution powder
diffraction beamline ID22 (ESRF, Grenoble) by using a constant
wavelength of l=0.400737(6) Š and a 0.3 mm glass capillary as
the sample holder, Debye–Sherrer mode, and a nine-crystal multia-
nalyzer as the detector. The Rietveld refinement was completed by
using JANA2006.[34]
Elemental analysis
Conclusion
A scanning electron microscope (SEM XL30) equipped with an
energy-dispersive X-ray spectrometer (EDX) from Philips, working
at 25 kV, was used for elemental analysis. The sample was briefly
exposed to air during the transport into the microscope, but there
was no obvious sample degradation.
The host–guest compound [Cs6Cl][Fe24Se26] can be obtained
through solid-state reactions in closed vessels. The Fe–Se lat-
tice is built from edge-sharing fused-cubane entities
[Fe8Se6Se8/3], in which the divalent, tetrahedrally coordinated
Fe is found as octamers. An antiferromagnetic state is ob-
served below TN =221 K and the complex Mçssbauer spectra
suggest a non-trivial magnetic ground state. The strong cova-
lent bonding based on a near degeneracy of the Fe3d and
Se4p orbitals together with small energies for virtual excita-
tions indicate that the title compound is close to electronic
itinerancy.
Spectroscopy
Soft X-ray absorption spectroscopy (XAS) at the Fe-L2,3 edge was
measured at the BL08B beamline of the National Synchrotron Radi-
ation Research Centre (NSRRC) in Taiwan. The Fe-L2,3 XAS spectra of
[Cs6Cl][Fe24Se26] together with Fe0.04Mg0.96O[23] and Fe2O3 as Fe2+
and Fe3+ references, respectively, were taken in the total electron
yield (TEY) mode with a photon energy resolution of 0.2 eV. Clean
sample surfaces were obtained by cutting pellets in situ just
before collecting the data in an ultrahigh vacuum chamber with
the pressure in the mid-10À10 mbar range.
Experimental Section
Mçssbauer spectra were collected by using a standard WissEl spec-
trometer operated in the constant acceleration mode and with
a
in a Plexiglas container. The Fe content was ꢀ10 mgcmÀ2. Spectra
between 4.7 and 290 K were obtained by using a Janis-SHI-850–5
closed cycle refrigerator (CCR). The isomer shifts are given relative
to a-Fe. The data were evaluated by using the program Moss-
Winn[35] and the thin absorber approximation.
Sample preparation
57Co/Rh source. A powder sample was mixed with BN and placed
The Cs2Se precursor was synthesized by reacting Cs metal (3n,
Chempur) with elemental Se (5n, Alfa). Cs metal was placed in a Ta
tube with one end open. The corresponding stoichiometric
amount of Se was placed in a smaller Ta tube that fit into the first
one. Care was taken to not let the tube fall over. Subsequently, the
outer Ta tube was weld-sealed. The sample was heated for 12 h at
2108C, then heated at a rate of 508hÀ1 to 8008C, and finally cooled
to RT at the same rate. Cs2Se was bright grey on grinding.
Physical properties
Stoichiometric amounts of powdered Cs2Se, CsCl (3n, TRC), Fe (3n,
Alfa), and Se (2.5:1:24:23.5) were mixed in an agate mortar in a con-
trolled atmosphere (O2 and H2O <0.1 ppm, MBraun Labmaster
glove box). Pellets of the mixture were reacted in corundum cruci-
bles inside evacuated (<10À7 bar) silica tubes. The sample was an-
nealed for 60 h at 5008C with slow heating and cooling. Without
exposure to air, the sample was reground, repelletized, and heated
again in the same way to afford the compound powder investigat-
ed herein.
Magnetization data were obtained by using a MPMS-XL instrument
(Quantum Design). In the range of 2–350 K, a polycarbonate capsu-
le was used as sample holder, and between 300 and 750 K, a high-
purity silica tube was used. Temperature-dependent data were ob-
tained at a constant field of 2 T and the high-temperature data
were shifted to account for the temperature-independent signal
from the furnace.
A standard four-point method was used, in which the application
of ac or dc current delivered temperature-dependent resistivity
data by using a physical property measurement system (PPMS,
Quantum Design). Gold contacts were fastened to a polycrystalline
piece with silver-filled epoxy inside a glovebox. The setup was
Small single crystals were grown in CsCl flux inside an evacuated
silica ampoule. A mixture with a Fe/Se/SeO2/CsCl molar ratio of
4:1:1:3 was heated to 7508C within 6 h. Subsequently the mixture
was held at this temperature for 96 h, then cooled to 5008C over
Chem. Eur. J. 2016, 22, 4626 – 4631
www.chemeurj.org
4630
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[9] H. T. Evans Jr., J. R. Clark, Am. Mineral. 1981, 66, 376–384.
quickly transferred to the cryostat with exposure to air for only
a few seconds.
[10] V. N. Yakovenchuk, Y. A. Pakhomovsky, Yu. P. Men’shikov, G. Yu. Ivanyuk,
S. V. Krivovichev, P. C. Burns, Can. Mineral. 2003, 41, 503–511.
[11] B. S. Tani, Am. Mineral. 1977, 62, 819–823.
[12] M. Lindqvist, D. Lindqvist, A. Westgren, Sv. Kem. Tids. 1936, 48, 156–
160.
[13] J. Kim, D. C. Rees, Science 1992, 257, 1677–1682.
[14] J. Kim, D. C. Rees, Nature 1992, 360, 553–560.
[15] S. Pohl, U. Opitz, Angew. Chem. Int. Ed. Engl. 1993, 32, 863–864; Angew.
Chem. 1993, 105, 950–952.
To extract the specific heat, the nonadiabatic thermal relaxation
method of the HC option setup of the PPMS (Quantum Design)
was used. A well-sintered sample piece was attached to the holder
by using Apiezon N grease. Due to the air sensitivity of the sample,
the mounting was performed in a glovebox and the sample was
protected by a thin layer of grease during the short exposure to
air between box and cryostat.
[16] G. Brauer, Z. Anorg. Allg. Chem. 1939, 242, 1–22.
[17] B. Chabot, K. Cenzual, E. ParthØ, Acta Crystallogr. Sect. A 1981, 37, 6–11.
[18] W. L. Bragg, Philos. Mag. 1914, 28, 255–360.
Acknowledgements
[19] A. W. Hull, W. P. Davey, Phys. Rev. 1921, 17, 266–267.
[20] A. Helms, W. Klemm, Z. Anorg. Allg. Chem. 1939, 242, 33–40.
[21] M. Schwarz, C. Rçhr, Z. Anorg. Allg. Chem. 2014, 640, 2792–2800.
[22] H. Lin, L. H. Li, L. Chen, Inorg. Chem. 2012, 51, 4588–4596.
[23] T. Haupricht, R. Sutarto, M. W. Haverkort, H. Ott, A. Tanaka, H. H. Hsieh,
H. J. Lin, C. T. Chen, Z. Hu, L. H. Tjeng, Phys. Rev. B 2010, 82, 035120.
[24] C. Koz, M. Schmidt, H. Borrmann, U. Burkhardt, S. Rçbler, W. Carrillo-
Cabrera, W. Schnelle, U. Schwarz, Yu. Grin, Z. Anorg. Allg. Chem. 2014,
640, 1600–1609.
We thank Rüdiger Kniep, Daniel Khomskii, and Deepa Kasina-
than for their invaluable comments. Ralf Koban is acknowl-
edged for assistance with the thermodynamic measurements.
We also thank Petra Scheppan for the electron microscopy
studies (EDX).
[25] J. M. Caron, J. R. Neilson, D. C. Miller, A. Llobet, T. M. McQueen, Phys.
Rev. B 2011, 84, 180409.
[26] M. Valldor, P. Adler, Yu. Prots, U. Burkhardt, L. H. Tjeng, Eur. J. Inorg.
Chem. 2014, 36, 6150–6155.
Keywords: inorganic synthesis · iron · magnetic properties ·
Mçssbauer spectroscopy · X-ray diffraction
[27] W. M. Reiff, I. E. Grey, A. Fan, Z. Eliezer, H. Steinfink, J. Solid State Chem.
1975, 13, 32–40.
[28] K. Kome˛dera, A. K. Jasek, A. Błachowski, K. Ruebenbauer, M. Piskorz, J.
[1] F. C. Hsu, J. Y. Luo, K. W. Yeh, T. K. Chen, T. W. Huang, P. M. Wu, Y. C. Lee,
Y. L. Huang, Y. Y. Chu, D. C. Yan, M. K. Wu, Proc. Natl. Acad. Sci. USA
2008, 105, 14262–14264.
[2] T. M. McQueen, A. J. Williams, P. W. Stephens, J. Tao, Y. Zhu, V. Ksenofon-
tov, F. Casper, C. Felser, R. J. Cava, Phys. Rev. Lett. 2009, 103, 057002.
[3] S. Tan, Y. Zhang, M. Xia, Z. Ye, F. Chen, X. Xie, R. Peng, D. Xu, Q. Fan, H.
Xu, J. Jiang, T. Zhang, X. Lai, T. Xiang, J. Hu, B. Xie, D. Feng, Nat. Mater.
2013, 12, 634–640.
[4] M. Burrard-Lucas, D. G. Free, S. J. Sedlmaier, J. D. Wright, S. J. Cassidy, Y.
Hara, A. J. Corkett, T. Lancester, P. J. Baker, S. J. Blundell, S. J. Clarke, Nat.
Mater. 2013, 12, 15–19.
[5] T. P. Ying, X. L. Chen, G. Wang, S. F. Jin, T. T. Zhou, X. F. Lai, H. Zhang,
W. Y. Wang, Sci. Rep. 2012, 2, 426.
˙
´
Zukrowski, A. Krzton-Maziopa, E. Pomjakushina, K. Conder, Solid State
Commun. 2015, 207, 5–8.
[29] S. E. Dann, M. T. Weller, D. B. Currie, M. F. Thomas, A. D. Al-Rawwas, J.
Mater. Chem. 1993, 3, 1231–1237.
[30] P. Adler, J. Solid State Chem. 1994, 108, 275–283.
[31] C. Goh, B. M. Segal, J. Huang, J. R. Long, R. H. Holm, J. Am. Chem. Soc.
1996, 118, 11844–11853.
[32] H. Kalyvas, D. Coucouvanis, Inorg. Chem. 2006, 45, 8462–8464.
[33] J. Guo, S. Jin, G. Wang, S. Wang, K. Zhu, T. Zhou, M. He, X. Chen, Phys.
Rev. B 2010, 82, 180520.
ˇ
ˇ ˇ
[34] V. Petrícek, M. DuSek, L. Palatinus, JANA2006, Institute of Physics, Acad-
emy of Sciences of the Czech Republic, Praha (Czech Republic), 2013.
[35] http://mosswinn.com.
[6] Y. Peng, G. Xi, C. Zhong, L. Wang, J. Lu, X. Sun, L. Zhu, Q. Han, L. Chen,
L. Shi, M. Sun, Q. Li, M. Yu, M. Yin, Geochim. Cosmochim. Acta 2009, 73,
4862–4878.
[7] G. K. Czamanske, M. A. Lanphere, R. C. Erd, M. C. Blake Jr., Earth Planeta-
ry Sci. Lett. 1978, 40, 107–110.
Received: December 2, 2015
Published online on February 16, 2016
[8] G. K. Czamanske, R. C. Erd, B. F. Leonard, J. R. Clark, Am. Mineral. 1981,
66, 369–375.
Chem. Eur. J. 2016, 22, 4626 – 4631
www.chemeurj.org
4631
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim