Room-temperature synthesis, hydrothermal recrystallization, and properties of metastable stoichiometric FeSe

Article abstract of DOI:10.1021/ic300798p

Room-temperature precipitation from aqueous solutions yields the hitherto unknown metastable stoichiometric iron selenide (ms-FeSe) with tetragonal anti-PbO type structure. Samples with improved crystallinity are obtained by diffusion-controlled precipitation or hydrothermal recrystallization. The relations of ms-FeSe to superconducting η-FeSe_1-x and other neighbor phases of the iron-selenium system are established by high-temperature X-ray diffraction, DSC/TG/MS (differential scanning calorimetry/ thermogravimetry/mass spectroscopy), ⁵⁷Fe Moessbauer spectroscopy, magnetization measurements, and transmission electron microscopy. Above 300 °C, ms-FeSe decomposes irreversibly to η-FeSe_1-x and Fe₇Se₈. The structural parameters of ms-FeSe (P4/nmm, a = 377.90(1) pm, c = 551.11(3) pm, Z = 2), obtained by Rietveld refinement, differ significantly from literature data for η-FeSe_1-x. The Moessbauer spectrum rules out interstitial iron atoms or additional phases. Magnetization data suggest canted antiferromagnetism below T_N = 50 K. Stoichiometric non-superconducting ms-FeSe can be regarded as the true parent compound for the 11 iron-chalcogenide superconductors and may serve as starting point for new chemical modifications.

Full text of DOI:10.1021/ic300798p

Article
pubs.acs.org/IC
Room-Temperature Synthesis, Hydrothermal Recrystallization, and
Properties of Metastable Stoichiometric FeSe
F. Nitsche,† T. Goltz,‡ H.-H. Klauss,‡ A. Isaeva,† U. Muller,† W. Schnelle,§ P. Simon,§ Th. Doert,*,†
̈
and M. Ruck†,§
†Department of Chemistry and Food Chemistry, Technische Universitat Dresden, D-01062 Dresden, Germany
̈
‡Institute of Solid State Physics, Technische Universitat Dresden, D-01062 Dresden, Germany
̈
§Max Planck Institute for Chemical Physics of Solids, D-01187 Dresden, Germany
S
* Supporting Information
ABSTRACT: Room-temperature precipitation from aqueous solutions yields the hitherto
unknown metastable stoichiometric iron selenide (ms-FeSe) with tetragonal anti-PbO type
structure. Samples with improved crystallinity are obtained by diffusion-controlled precipitation
or hydrothermal recrystallization. The relations of ms-FeSe to superconducting β-FeSe1−x and
other neighbor phases of the iron−selenium system are established by high-temperature X-ray
diffraction, DSC/TG/MS (differential scanning calorimetry/thermogravimetry/mass spectros-
copy), 57Fe Mossbauer spectroscopy, magnetization measurements, and transmission electron
̈
microscopy. Above 300 °C, ms-FeSe decomposes irreversibly to β-FeSe1−x and Fe7Se8. The
structural parameters of ms-FeSe (P4/nmm, a = 377.90(1) pm, c = 551.11(3) pm, Z = 2),
obtained by Rietveld refinement, differ significantly from literature data for β-FeSe1−x. The
Mossbauer spectrum rules out interstitial iron atoms or additional phases. Magnetization data
̈
suggest canted antiferromagnetism below TN = 50 K. Stoichiometric non-superconducting ms-FeSe can be regarded as the true
“parent” compound for the “11” iron-chalcogenide superconductors and may serve as starting point for new chemical
modifications.
and a considerable homogeneity range, we refer to them only as
Fe3Se4 and Fe7Se8 for shortness and clarity.
INTRODUCTION
■
Compounds containing square iron nets received new attention
from physicists and chemists alike after the discovery of
superconductivity up to 55 K in doped rare-earth iron
pnictides.1,2 Besides the search for new structures with potential
for high-temperature superconductivity, a closer inspection of
known compounds revealed superconducting properties. β-
FeSe1−x with anti-PbO type structure3 is one example.
The structural background of the nonstoichiometry of β-
FeSe1−x is still under debate: Either excess iron atoms occupy
interstitial positions between the layers in analogy to Fe2As,7 or
some of the selenium positions are void.8,9 Whereas doping is
needed for the iron pnictide phases (e.g., BaFe2As2 so-called
“Ba-122” and LaFeAsO “La-1111”)10 to become superconduct-
ing, the intrinsic doping of β-FeSe1−x (“11”) by deviation from
the precise 1:1 composition induces superconductivity.8,9,11 In
contrast to the “122” and “1111” phases, substitution of iron by
other transition metals leads to suppression or only small
enhancement of superconductivity in iron selenide.12−14
Crystals of tetragonal β-FeSe1−x grown at temperatures above
the peritectoid point are often pseudomorphic showing the
hexagonal habitus of δ-FeSe.15−19 Consequently, growth of
single crystals has to be performed within the stability range of
β-FeSe1−x, as shown by Oyler et al.,20 by the reaction of
elemental precursors in high-boiling organic solvents.
Considerable research on the iron−selenide system has been
performed, and the complex phase diagram is well established.4
Tetragonal β-FeSe1−x was first reported by Hagg and
̈
Kindstrom5 in 1933. It is found to undergo a peritectoid
̈
decomposition at 457 °C into α-Fe and δ-FeSe (NiAs type)
and exhibits a narrow homogeneity range around 49 at. % Se
(see Figure 4).4,6 Hexagonal δ-FeSe is formed in an eutectoid
reaction of δ-FeSe1−x and Fe3Se4 (defect NiAs type, see Table
1) at 350 °C. Although Fe3Se4- and Fe7Se8-type phases are
nonstoichiometric compounds with vacancy-ordered structures
For the sulfide FeS, a synthesis in aqueous media at room
temperature has been reported.21,22 Furthermore, theoretical
studies of the solubility of iron chalcogenides in water have
been performed.23 The synthesis of FeSe in water is feasible
only in the absence of oxygen, since the product11 and the
Table 1. Overview of the Phases Discussed in the Text
designation
structure type
crystal system
β-FeSe1−x
ms-FeSe
δ-FeSe
anti-PbO
anti-PbO
NiAs
tetragonal
tetragonal
hexagonal
Fe3Se4
defect-NiAs
defect-NiAs
orthorhombic
monoclinic
Received: April 18, 2012
Published: June 20, 2012
Fe7Se8 (α, β)
© 2012 American Chemical Society
7370
dx.doi.org/10.1021/ic300798p | Inorg. Chem. 2012, 51, 7370−7376

Inorganic Chemistry
Article
reactants are easily oxidized. Following these ideas, we
conducted a diffusion-controlled precipitation from aqueous
solutions and obtained metastable FeSe (ms-FeSe) with anti-
PbO type structure and stoichiometric 1:1 composition. It
represents the undoped non-superconducting “parent” com-
pound for the “11” system and provides a starting point for
future chemical modifications, to induce superconductivity in
analogy to the “122” and “1111” systems.
We report on the synthesis and hydrothermal recrystalliza-
tion of ms-FeSe. The relation of ms-FeSe to the neighbor
phases of the Fe−Se system is investigated and discussed.
Finally, the crystal structure data, the Mossbauer spectrum, and
̈
the magnetic and thermal properties are given.
EXPERIMENTAL SECTION
■
All starting materials and products are oxygen-sensitive and were
handled and stored in an argon-filled glovebox (M. Braun; p(O2)/p0 ≤
1 ppm, p(H2O)/p0 ≤ 1 ppm, argon purified with molecular sieve and
copper catalyst).
Starting Materials. Li2Se was prepared by reacting lithium (>99%,
Acros Organics) and selenium (99.999%, Chempur, treated with H2 at
150 °C) under Schlenk conditions in freshly dried tetrahydrofuran
(THF, 99.7%, VWR) using naphthalene (99%, Acros Organics) as
catalyst.24 The obtained creme-colored solid was washed three times
with THF and dried under vacuum.
FeI2 was prepared by the reaction of the elements (Fe: 99.5%, Alfa
Aeser; I2: >99.8%, Merck, sublimated twice over BaO) in a silica
ampule (15 cm length, 14 mm diameter) using 10 at. % excess of iron.
The iron-containing side of the ampule was placed in a horizontal tube
furnace at 550 °C, while the other side of the ampule resided outside
the furnace to limit the pressure of gaseous iodine. FeI2 exhibits a
noticeable vapor pressure at 550 °C25 and was separated from the
excess iron by sublimation.
Figure 1. H-tube for diffusion-controlled reaction of Li2Se (a) with
FeI2 (b) to ms-FeSe (d) with silica wool (c) inserted to prevent
reactants from uncontrolled mixing during filling with distilled water.
Photograph taken after 10 days of reaction.
or silica capillaries (Hilgenberg) for room- and high-temperature
measurements, respectively. The temperature in the heating chamber
for PXRD was calibrated by the melting points of Sn (232 °C, 99.99%,
Fluka) and Zn (420 °C, 99.999%, Strem Chemicals). In the high-
temperature experiment the temperature of measurement was accessed
with a rate of 10 K min−1 and then kept constant while two powder
diffraction patterns were taken subsequently with a measurement time
of three hours each.
Both starting materials are easily oxidized by molecular oxygen, yet
they are stable and soluble in oxygen-free water. Thus, the water used
for reactions was distilled under streaming argon to remove molecular
oxygen (denoted as distilled water in the following text). Aqueous
solutions of Li2Se and FeI2 are deep red and light green, respectively.
Combining both liquids immediately yields a black precipitate which is
amorphous according to X-ray diffraction.
Diffusion-Controlled Precipitation. An H-shaped tube was used
for the controlled precipitation of ms-FeSe (Figure 1). Under argon,
Li2Se and FeI2 powders were filled separately in the legs of the tube.
To hinder uncontrolled dissolution and mixing of the reactants, silica
wool was placed on top of the powders. Distilled water was added as
diffusion medium and the tube was plugged. After 40 days at room
temperature, a black precipitate had formed, which was then removed
from the tube and washed three times with distilled water and dried
subsequently under vacuum.
Hydrothermal Recrystallization. FeI2 and Li2Se were dissolved
separately in distilled water in Schlenk tubes, and the solutions were
combined in a silica tube. The resulting aqueous solution with the
immediately formed black precipitate was frozen by immersing the
container in liquid nitrogen before the open end was fused. The
approximately half-filled ampule was inserted into a Teflon-lined steel
autoclave (DAB-2, BERGHOF), which was then also half filled with
water to generate the necessary counter-pressure for the ampule. The
steel autoclave was stored in a heat-cabinet at 200 °C. After three days,
the ampule was opened under argon in a glovebag (ABCR). The
colorless solution was decanted, and the black polycrystalline material
was washed three times with distilled water and dried subsequently
under vacuum.
Characterization. Powder X-ray diffraction (PXRD) patterns at
room temperature were measured with an X’Pert PRO diffractometer
(PANalytical, Cu Kα1, Ge monochromator, PIXcel detector), while
high-temperature measurements (HT-PXRD) were performed on a
Stadi P diffractometer (STOE, Cu Kα1, Ge monochromator, linear
position sensitive detector). The samples were enclosed in borosilicate
Lattice parameters and phase fractions were obtained by Rietveld
refinement using GSAS,26 EXPGUI27, and Jana2006,28 applying
models for axial divergence and anisotropic line broadening.
Background contributions were fitted by Legendre polynomials.
Absorption corrections for cylindrical samples corresponding to a
packing of 60 vol. % in the 0.2 mm capillaries were applied. δ-FeSe,
Fe3Se4, α-Fe7Se8, and β-Fe7Se8 were treated within NiAs-type models,
since reflection splitting of the defect-NiAs type structures could not
be observed due to the broad reflections. Nevertheless, differences in
lattice parameters allow to distinguish the monoclinic phases (Fe7Se8)
from δ-FeSe during cooling from high temperature.
The Fe/Se ratio of the sample was determined with an ICP-OES
Vista RL (Varian) by previously dissolving the samples in a mixture of
hydrochloric and nitric acid. The reported ratio is the average of three
measurements from one sample.
Transmission electron microscopy (TEM) images at the middle
resolution regime, as well as electron diffraction patterns, were
recorded using a FEI Tecnai 10 microscope (LaB6-source, 100 kV,
TEM camera TVIPS F224HD 2k × 2k). The images were analyzed
using Digital Micrograph29 and ImageJ.30
Differential scanning calorimetry measurements (DSC) combined
with thermogravimetry (TG) and mass spectroscopy (MS) were
performed with a STA 409 CD Skimmer (NETZSCH) operated
inside a glovebox with high-purity argon.31 Alumina-lined Pt-crucibles
with lid were used and double runs with heating rates of 5 and 10 K
min−1 were conducted. Baseline contributions were fitted with
quadratic interpolation and subtracted.
57Fe Mossbauer spectra were measured in transmission geometry at
̈
room temperature using a 57Co/Rh γ-radiation source (emission line
half width at half-maximum Γ = 0.135(2) mm s−1) and a Kr/CO2
proportional counter as detector. The spectrometer was calibrated to
α-Fe at room temperature.
Full-potential linearized augmented plane-wave method (FP-
LAPW) and basis sets implemented within the Elk code32 were
utilized for quantum chemical calculations. The generalized gradient
7371
dx.doi.org/10.1021/ic300798p | Inorg. Chem. 2012, 51, 7370−7376

Inorganic Chemistry
Article
approximation (GGA) was employed for the exchange-correlation
functional as proposed by Perdew and Wang.33 A number of 90
irreducible k-points was considered. The electric field gradient (EFG)
was computed from the converged self-consistent field (SCF) energy
taking spin−orbit coupling into account. The constant magnetic field
of 1.7 T was applied in the z-direction in order to break spin
symmetry. The basis set was expanded with a number of FP-LAPW
basis functions up to RMT Kmax = 12 (where RMT is the average radius
of the muffin-tin spheres and Kmax is the maximum value of the wave
vector K = k + G) and the maximum length of G for expanding the
interstitial density and potential is equal to 22. The atomic orbitals up
to an angular momentum l = 12 were used to expand the wave
functions inside the muffin-tin spheres.
Magnetic measurement were performed with a SQUID-magneto-
meter (Quantum Design MPMS-XL7) in the temperature range
between 1.8 and 400 K and external fields of 2 mT to 7 T. Samples
were prepared in an argon-filled glovebox, transferred under argon,
and measured in helium atmosphere. The susceptibility of the silica
sample holder was determined prior to the measurement and
subtracted.
Magnetite, identified as an impurity phase with 0.4(1) wt %
during Rietveld refinement, is believed to bias only the
magnetic characterization.
ms-FeSe, as well as β-FeSe1−x, crystallizes in the anti-PbO
type structure. However, there are subtle yet significant
differences (see Table 2): The c-axis of ms-FeSe is shorter
36
,
a
Table 2. Structure Data of ms-FeSe and β-FeSe1−x
param.
a (pm)
ms-FeSe
β-FeSe1−x
377.90(1)
551.11(3)
1.45835(9)
0.2655(3)
239.0(1)
377.376(2)
552.482(5)
1.46401(2)
0.2652(1)
238.89(3)
104.34(1)
372.18(5)
c (pm)
c/a
zSe
Fe−Se (pm)
Fe−Se−Fe (deg)
Se···SevdW (pm)
104.48(5)
371.7(2)
a
both: space group P4/nmm, no. 129, 1st setting.
5,8,11,17,34−38
RESULTS AND DISCUSSION
and the a-axis is longer than for β-FeSe1−x
,
■
Diffusion-controlled room-temperature precipitation from
aqueous solutions yielded tetragonal ms-FeSe according to
the following equation:
resulting in a significantly smaller c/a ratio. In contrast, the
distance of iron and selenium is equal. Consequently, the
structural difference of ms-FeSe relative to β-FeSe1−x can be
interpreted as an in-plane expansion of the iron−selenium slab
accompanied by a contraction in stacking direction. The
“tetrahedral” angle (Fe−Se−Fe in Table 2) increases
correspondingly.
Plate-shaped crystals of ms-FeSe with broad particle-size
distribution around 100 nm are identified by TEM (Figure 3).
Electron diffraction patterns of isolated particles could be
indexed with the tetragonal structure model of FeSe (Table 2).
The major faces are {0 0 1} and the side faces belong to the
form {1 1 0}. The preferred growth directions of ms-FeSe, thus,
are ⟨1 0 0⟩. The observed anisotropy of the crystals supports
the model used for Rietveld refinement.
While dissolving the samples for elemental analysis, some
residual material remained. However, the determined ratio of
iron and selenium gave 49.8(2) at. % Se, which is in excellent
agreement with the 1:1 composition expected for a
precipitation following eq 1. To further probe the composition
of ms-FeSe, the relation to the phases in the literature-known
phase diagram (Figure 4) was assessed by HT-PXRD and
thermal analysis.
H2O
FeI2(s) + Li2Se(s) ⎯⎯⎯→ ms‐FeSe(s) + 2Li(+aq) + 2I(−aq)
(1)
No additional phases were detected by powder X-ray
diffraction. Preliminary HT-XRD and TG/MS investigations
showed decomposition of the phase above 303(5) °C to Fe3Se4
and β-FeSe1−x. MS showed only loss of residual solvent water
but no selenium containing species.
Based on these results, a faster synthesis route was
developed. The black precipitate obtained by fast mixing of
solutions of Li2Se and FeI2 appears to be amorphous in PXRD.
Hydrothermal treatment at 200 °C increases the crystallinity
and yields ms-FeSe within three days instead of 40 days of the
diffusion-controlled reaction. All further characterization was
done with this sample.
Rietveld refinement of PXRD data of ms-FeSe (Figure 2)
succeeded with a model for anisotropic particle size. Applying
the Scherrer equation to the profile parameters leads to a
particle size of 16 nm in c direction and 27 nm in the directions
of the basal plane (further data in the Supporting Information).
DSC experiments with ms-FeSe gave very similar results for
different heating rates (5 and 10 K min−1). However, the
second heating of the sample differs markedly from the first
(Figure 5), indicating that the sample is out of thermodynamic
equilibrium before the first run. The second heating run can be
interpreted based on the phase diagram given in Figure 4. The
first endothermic signal corresponds to the eutectoid reaction
of β-FeSe1−x, Fe3Se4, and δ-FeSe at 332(2) °C. The
symmetrical signal of the eutectoid reaction is invariant to
the heating rate and superimposed with an asymmetrical signal,
which corresponds to the solvus line of δ-FeSe ending at
441(1)°C and 50 at. % Se (extrapolation to zero heating rate)
in agreement with the data reported by Katsuyama et al.6
Consequently, the first heating run can be interpreted as an
exothermal decomposition of ms-FeSe with 1:1 composition
starting above 300 °C with the superimposed eutectoid reaction
approaching the equilibrium evolution above 430 °C.
The HT-PXRD shows that the decomposition of ms-FeSe is
starting at a kinetical temperature of 303(5) °C. Whereas the
first HT-PXRD pattern at this temperature indicates only ms-
Figure 2. Rietveld refinement of ms-FeSe with 0.4(1) wt % magnetite
impurity. Experimental data plotted as dots, fitted pattern as solid line.
Lower panel: difference plot.
7372
dx.doi.org/10.1021/ic300798p | Inorg. Chem. 2012, 51, 7370−7376

Inorganic Chemistry
Article
Figure 5. DSC measurement of ms-FeSe with a heating rate of 5 K
min−1. The black curve indicates the first heating run and the gray one
the second. Dotted line marks zero.
Figure 3. TEM images of ms-FeSe. Upper panel: overview, showing
the particle-size distribution. (a) Plate-shaped crystallites oriented
along the basal plane. (b) Crystallites with basal plane perpendicular to
viewing direction. Middle panel: view along [0 0 1]. Lower panel: view
along [1 1 0].
̅
FeSe (point 1 in Figure 6), the second pattern, taken at the
same temperature three hours later, shows a mixture of β-
FeSe1−x and a NiAs type phase (point 2). Upon further heating,
Figure 6. Evolution of phases and their lattice parameters as derived
from Rietveld refinements. Red markers indicate the first heating run,
blue markers the subsequent cooling, and orange markers the second
heating run. Hexagons and circles indicate NiAs structure related
phases. Diamonds indicate β-FeSe1−x, and squares ms-FeSe. Points 1
and 2 are subsequent observations in the course of heating; points 3
and 4 are during cooling. Error bars represent the confidence interval
of 2σ and are within markers in case of phase fractions. Lines are
guides for the eye.
Figure 4. Phase diagram of iron and selenium in analogy to refs 4 and
6. Red points correspond to the results of Rietveld refinements of HT-
PXRD data with the resulting solvus and eutectoid lines in green.
Though ms-FeSe is not a thermodynamically stable phase, it is
included in the phase diagram in order to visualize the relation of the
compound to the neighbor phases.
7373
dx.doi.org/10.1021/ic300798p | Inorg. Chem. 2012, 51, 7370−7376

Inorganic Chemistry
Article
the content of the NiAs type phase increases following the
solvus line in the phase diagram (Figure 4). HT-PXRD at
432(5) °C reveals that the sample has transformed completely
into δ-FeSe, which corresponds well to the temperature of the
solvus line of δ-FeSe at 50 at. % Se.
For temperatures above the decomposition of metastable ms-
FeSe, the system is in equilibrium. The phase fractions
determined by HT-PXRD can thus be transferred into the
phase diagram (red markers in Figure 4) by applying the lever
rule with the prerequisite of 49.2 at. % Se as the solvus line for
β-FeSe1−x. The so-obtained phase relations (green lines in
Figure 4) follow tentatively the literature-known phase diagram
with a slightly lower temperature for the eutectoid reaction and
a different progression of the solvus line.
In summary, HT-PXRD and DSC measurements support the
1:1 composition obtained for ms-FeSe from elemental analysis.
Furthermore, ms-FeSe and β-FeSe1−x can be distinguished by
their unit cell volumes and c/a ratios (Table 2).
Upon cooling from 460 °C, δ-FeSe was supercooled down to
240(5) °C. The almost linear decrease of volume indicates
constant composition (Figure 6). At 240 °C, the first HT-
PXRD pattern shows δ-FeSe only (point 3 in Figure 6), while
the subsequent HT-PXRD pattern indicates the beginning
decomposition into β-FeSe1−x and Fe7Se8 (points 4). The
formation of tetragonal β-FeSe1−x below 300 °C and its
increasing phase fraction is in agreement with previous studies,4
and suggests that the nonstoichiometric β-FeSe1−x is a
thermodynamically stable phase at room temperature. These
results refute the phase diagramm given by McQueen et al.8
with tetragonal β-FeSe1−x as a high-temperature modification of
hexagonal δ-FeSe above 300 °C.
While reheating the obtained phase mixture of β-FeSe1−x, δ-
FeSe, and Fe7Se8 (orange markers in Figure 6) the volume and
c/a ratio of β-FeSe1−x coincide with the data of the first heating,
indicating that the high-temperature behavior is reproducible.
The pronounced decrease of the c/a ratio above 300 °C
indicates a structural change of β-FeSe1−x beyond thermal
expansion. To emphasize the effect, the linear thermal
expansion coefficients αa ≈ αc ≈ 3 × 10−5 K were determined
by linear regression based on the lattice parameters obtained for
cooling and second heating in the range from room
temperature to 250 °C. These coefficients are in good
agreement with literature data for Fe1.06Te and FeTe0.5Se0.5.39
After subtraction of the linear thermal expansion according to
Figure 7. Temperature dependence of the lattice parameters of ms-
FeSe and β-FeSe1−x. The data in the two lower panels are corrected for
thermal expansion. Red symbols indicate the first heating, blue
symbols the cooling, and orange symbols the second heating run.
Diamonds indicate β-FeSe1−x, squares ms-FeSe. Error bars represent
2σ. Lines are guides for the eye.
a′(ϑ) = a(ϑ) − α (ϑ − 20°C)
a
(and correspondingly for c), the divergence of the a- and c-axes
above 300 °C is discernible (Figure 7), emphasizing a structural
change of β-FeSe1−x beyond thermal expansion. This effect can
either be attributed to a structural reorganization (e.g., from
selenium vacancies to iron interstitials) or a change of
composition. It could, in any case, explain the influence of
thermal treatment on the superconducting phase transition for
various samples.9,11,40
Figure 8. Mossbauer spectrum of ms-FeSe at room temperature.
Experimental data plotted as dots, fitted doublet as solid line. Lower
panel: difference plot.
̈
Since the crystal structures of ms-FeSe and β-FeSe1−x differ
in detail, small differences of the physical properties can be
the isomer shift in ms-FeSe is slightly smaller. This reflects an
increase of the local electron density at the iron atoms as it
would be expected for an itinerant system with a smaller c-axis
(Table 2).
The asymmetry of the doublet pattern for ms-FeSe shows
increased absorption at higher velocities. In contrast,
expected. The room-temperature Mossbauer spectrum (Figure
̈
8) consists of an asymmetric doublet related to ms-FeSe. No
sign of the Fe3O4 impurity was detected. From the spectrum, an
isomer shift of δ = 0.43(1) mm s−1 and a quadrupole splitting
of ΔEQ = 0.28(1) mm s−1 is obtained for ms-FeSe (β-FeSe1−x
:
δ = 0.45−0.47 mm s−1, ΔEQ = 0.26−0.29 mm s−1).7,8,38,41
Mossbauer spectra of β-FeSe and Fe1+xTe show increased
̈
1−x
While the quadrupole splitting is comparable for both phases,
absorption at lower velocities, which has been attributed to
7374
dx.doi.org/10.1021/ic300798p | Inorg. Chem. 2012, 51, 7370−7376

Inorganic Chemistry
Article
either occupation of interstitial sites by iron atoms or preferred
orientation of the polycrystalline material (texture effect).42−44
Following this argumentation, we can rule out interstitial iron
atoms for ms-FeSe for two reasons: First, the asymmetry is
opposite to experimental findings in all nonstochiometric
samples. Second, assuming interstitial sites would lead to more
than 2% off-stochiometry which is not consistent with the
results from HT-PXRD, DSC, and elemental analysis.
CONCLUSION
■
Precipitation from aqueous solutions of FeI2 and Li2Se yielded
the yet unknown metastable phase ms-FeSe which has
stoichiometric composition and anti-PbO type structure. It
possibly exhibits magnetic ordering below 50 K and is
nonsuperconducting down to 1.8 K in an external field of 2
mT. Above 300 °C, ms-FeSe decomposes irreversibly into the
thermodynamically stable phases Fe7Se8 and β-FeSe1−x
.
The asymmetry can, however, be explained by preferred
orientation. The plate-shaped particles (Figure 3) tend to
orientate preferably with their basal plane parallel to the flatbed
sample holder. In an ideal powder without preferred
orientation, the vector of the principal axis vzz of the electrical
field gradient of the individual particle and the wavevector of
the incident γ-radiation enclose in average the “magic angle” of
54.7° resulting in two symmetric absorption lines with equal
intensity. A preferred orientation of the particles will lead to a
deviation of this average angle (θγ, Q) from the “magic angle”,
and thus to a change in the relative line intensities. With vzz > 0,
ms-FeSe reveals a significantly shorter c and longer a-axis
compared to β-FeSe1−x, while the site parameter z for selenium
is unchanged. As a consequence of the different local
environment of iron relative to β-FeSe1−x, the isomer shift in
the Mossbauer spectrum is smaller. The difference of the line
̈
intensities of the doublet is attributed to texture effects due to
preferred orientation of the crystallites.
There is evidence that β-FeSe1−x undergoes a structural
change above 300 °C, which may be due to a change of
composition or to a reorganization form interstitial iron atoms
to selenium defects. This may explain the dependence of
superconductivity on thermal treatment as pointed out by
McQueen et al.8 In contrast to recent reports,8 β-FeSe1−x
proves to be a thermodynamically stable phase at room-
temperature.
Metastable stoichiometric ms-FeSe is the undoped “parent”
compound for the tetragonal iron−selenium (“11”) phases and
thus opens up new ways for future chemical modifications in
analogy to the “122” and “1111” systems. Substitution of iron
with cobalt in the structure may lead to enhanced super-
conductivity and new properties, providing additional means to
study the “simplest” system of iron-based superconductors.
obtained from quantum chemical calculations, one finds θγ,Q
=
50(2)° showing the presence of a preferred orientation
consistent with plate-shaped particles.
Preliminary magnetic measurements of ms-FeSe in various
external fields (Figure 9) indicate a ferro-like signal around 50
ASSOCIATED CONTENT
* Supporting Information
■
S
Detailed data of the Rietveld refinement of ms-FeSe (PDF) and
corresponding crystallographic information file (CIF). This
material is available free of charge via the Internet at http://
pubs.acs.org/.
Figure 9. Temperature dependence of the magnetization of ms-FeSe
in various external fields. Inset: Zero-field cooled (zfc) and field cooled
(fc) magnetization measurement after thermal treatment of ms-FeSe
up to 500 °C.
AUTHOR INFORMATION
Corresponding Author
*Email: thomas.doert@chemie.tu-dresden.de.
■
Notes
The authors declare no competing financial interest.
K, which saturates above 3.5 T. The very low moment of 0.03
μB per formula unit indicates a weak ordering which could be
due to a canted antiferromagnetic state. It has to be noted that
this may not necessarily reflect the bulk properties of ms-FeSe
due to the small particle size. The ferrimagnetic magnetite
impurity leads to a field dependency of the signal in the whole
temperature range up to 400 K, hampering the interpretation of
the measurement.
It is, however, evident that ms-FeSe is not superconducting
down to 1.8 K in an external field of 2 mT. After thermal
treatment up to 500 °C ms-FeSe has decomposed into the
thermodynamically stable phases and magnetization indicates
superconductivity below 10 K originating from β-FeSe1−x
(Figure 9, inset). The increase of χ (with increasing
temperature) from large positive values at 1.8 K after zero-
field cooling (zfc) is due to spin freezing. It might also be
explained by the additional phases δ-FeSe and Fe7Se8 (see
phase fractions in Figure 6), which contribute to the overall
susceptibility. Similar observations have also been made for
other superconducting β-FeSe1−x samples.9,15,37,45,46
ACKNOWLEDGMENTS
■
The authors like to thank Prof. Dr. Peer Schmidt and Mr.
Philipp Materne for discussions, Dr. Stefan Hoffmann and
Susann Scharsach for thermal analysis, Dr. Gudrun Auffermann
for elemental analysis, and Ms. Gudrun Kadner for
experimental help. Part of this work was financially supported
by the German Research Foundation (DFG) within the priority
program SPP1458, project KL-1086/10-1.
REFERENCES
■
(1) Kamihara, Y.; Watanabe, T.; Hirano, M.; Hosono, H. J. Am.
Chem. Soc. 2008, 130, 3296−3297.
(2) Ren, Z.-A.; Che, G.-C.; Dong, X.-L.; Yang, J.; Lu, W.; Yi, W.;
Shen, X.-L.; Li, Z.-C.; Sun, L.-L.; Zhou, F.; Zhao, Z.-X. EPL (Europhys.
Lett.) 2008, 83, 17002.
(3) Hsu, F.; Luo, J.; Yeh, K.; Chen, T.; Huang, T.; Wu, P. M.; Lee, Y.;
Huang, Y.; Chu, Y.; Yan, D.; Wu, M. Proc. Natl. Acad. Sci. U.S.A. 2008,
105, 14262−14264.
(4) Okamoto, H. J. Phase Equilib. 1991, 12, 383−389.
7375
dx.doi.org/10.1021/ic300798p | Inorg. Chem. 2012, 51, 7370−7376

Inorganic Chemistry
Article
(5) Hagg, G.; Kindstrom, A. Z. Phys. Chem. 1933, B22, 453−464.
(38) Gom
́
ez, R. W.; Marquina, V.; Per
́
ez-Mazariego, J. L.; Escamilla,
̈
̈
(6) Katsuyama, S.; Ueda, Y.; Kosuge, K. Mater. Res. Bull. 1990, 25,
913−922.
R.; Escudero, R.; Quintana, M.; Hernandez-Gomez, J. J.; Ridaura, R.;
́
́
Marquina, M. L. J. Supercond. Novel Magn. 2010, 23, 551−557.
(39) Bud’ko, S. L.; Canfield, P. C.; Sefat, A. S.; Sales, B. C.; McGuire,
M. A.; Mandrus, D. Phys. Rev. B 2009, 80, 134523.
(40) McQueen, T. M.; Williams, A. J.; Stephens, P. W.; Tao, J.; Zhu,
Y.; Ksenofontov, V.; Casper, F.; Felser, C.; Cava, R. J. Phys. Rev. Lett.
2009, 103, 057002.
(41) Hamdeh, H. H.; El-Tabey, M. M.; Asmatulu, R.; Ho, J. C.;
Huang, T. W.; Yeh, K. W.; Wu, M. K. EPL (Europhys. Lett.) 2010, 89,
67009.
(42) Ward, J. B.; McCann, V. H. J. Phys. C: Solid State Phys. 1979, 12,
873−879.
(43) Mizuguchi, Y.; Furubayashi, T.; Deguchi, K.; Tsuda, S.;
Yamaguchi, T.; Takano, Y. Phys. C 2010, 470 (Supplement1),
S338−S339.
(44) Szyman
Gawryluk, D. J.; Berkowski, M.; Puzniak, R.; Wisn
Sci. Technol. 2011, 24, 105010.
(45) Błachowski, A.; Ruebenbauer, K.; Zukrowski, J.; Przewoznik, J.;
Wojciechowski, K.; Stadnik, Z. J. Alloys Compd. 2010, 494, 1−4.
(46) Fedorchenko, A. V.; Grechnev, G. E.; Desnenko, V. A.; Panfilov,
A. S.; Gnatchenko, S. L.; Tsurkan, V. V.; Deisenhofer, J.; Krug von
Nidda, H.; Loidl, A.; Chareev, D. A.; Volkova, O. S.; Vasiliev, A. N.
Low Temp. Phys. 2011, 37, 83−89.
(7) Tsuji, T.; Howe, A. T.; Greenwood, N. N. J. Solid State Chem.
1976, 17, 157−163.
(8) McQueen, T. M.; Huang, Q.; Ksenofontov, V.; Felser, C.; Xu, Q.;
Zandbergen, H.; Hor, Y. S.; Allred, J.; Williams, A. J.; Qu, D.;
Checkelsky, J.; Ong, N. P.; Cava, R. J. Phys. Rev. B 2009, 79, 014522.
(9) Williams, A. J.; McQueen, T. M.; Cava, R. J. Solid State Commun.
2009, 149, 1507−1509.
(10) Paglione, J.; Greene, R. L. Nat. Phys. 2010, 6, 645−658.
(11) Pomjakushina, E.; Conder, K.; Pomjakushin, V.; Bendele, M.;
Khasanov, R. Phys. Rev. B 2009, 80, 024517.
(12) Wu, M. K.; et al. Phys. C 2009, 469, 340−349.
(13) Thomas, E. L.; Wong-Ng, W.; Phelan, D.; Millican, J. N. J. Appl.
Phys. 2009, 105, 073906.
́ ́
ski, K.; Olszewski, W.; Dobrzyn
(14) Mizuguchi, Y.; Tomioka, F.; Tsuda, S.; Yamaguchi, T.; Takano,
Y. J. Phys. Soc. Jpn. 2009, 78, 074712.
́
́
(15) Zhang, S. B.; Sun, Y. P.; Zhu, X. D.; Zhu, X. B.; Wang, B. S.; Li,
G.; Lei, H. C.; Luo, X.; Yang, Z. R.; Song, W. H.; Dai, J. M. Supercond.
Sci. Technol. 2009, 22, 015020.
(16) Patel, U.; Hua, J.; Yu, S. H.; Avci, S.; Xiao, Z. L.; Claus, H.;
Schlueter, J.; Vlasko-Vlasov, V. V.; Welp, U.; Kwok, W. K. Appl. Phys.
Lett. 2009, 94, 082508.
(17) Mok, B. H.; Rao, S. M.; Ling, M. C.; Wang, K. J.; Ke, C. T.; Wu,
P. M.; Chen, C. L.; Hsu, F. C.; Huang, T. W.; Luo, J. Y.; Yan, D. C.;
Ye, K. W.; Wu, T. B.; Chang, A. M.; Wu, M. K. Cryst. Growth Des.
2009, 9, 3260−3264.
́
́
(18) Takeya, H.; Kasahara, S.; Hirata, K.; Mochiku, T.; Sato, A.;
Takano, Y. Phys. C 2010, 470 (Supplement1), S497−S498.
(19) Hara, Y.; Takase, K.; Yamasaki, A.; Sato, H.; Miyakawa, N.;
Umeyama, N.; Ikeda, S. Phys. C 2010, 470 (Supplement 1), S313−
S314.
(20) Oyler, K. D.; Ke, X.; Sines, I. T.; Schiffer, P.; Schaak, R. E. Chem.
Mater. 2009, 21, 3655−3661.
(21) Berner, R. A. J. Geol. 1964, 72, 293−306.
(22) Lennie, A. R.; Redfern, S. A. T.; Schofield, P. F.; Vaughan, D. J.
Mineral. Mag. 1995, 59, 677−683.
(23) Makarov, G. V.; Voevoda, N. Y.; Batrakov, V. V.; Kudasheva, T.
V. Zh. Fiz. Khim. 1993, 67, 1573−1576.
(24) Thompson, D. P.; Boudjouk, P. J. Org. Chem. 1988, 53, 2109−
2112.
(25) Knacke, O.; Kubaschewski, O. In Thermochemical Properties of
Inorganic Substances, 2nd ed.; Hesselmann, K., Ed.; Springer-Verlag:
Germany, 1991; Vol. 1−2.
(26) Larson, A. C.; Von Dreele, R. B. Los Alamos National Labratory
Report; Los Alamos National Laboratory: Los Alamos, NM, 2000; pp
86−748.
(27) Toby, B. H. J. Appl. Crystallogr. 2001, 34, 210−213.
(28) Petricek, V.; Dusek, M.; Palatinus, L. Jana2006; Institute of
Physics: Praha, Czech Republic, 2006.
(29) Digital Micrograph; Gatan Inc.: Pleasanton, CA, 2003.
(30) Abramoff, M. D.; Magelhaes, P. J.; Ram, S. J. Biophotonics Int.
2004, 11, 36−42.
(31) Hoffmann, S.; Schmidt, M.; Scharsach, S.; Kniep, R.
Thermochim. Acta 2012, 527, 204−210.
(32) The Elk FP-LAPW Code, 2009−2012. http://elk.sourceforge.
net/ (accessed: October 10, 2011).
(33) Perdew, J. P.; Wang, Y. Phys. Rev. B 1992, 45, 13244−13249.
(34) Grønvold, F.; Hertzberg, S.; Weeks, O. B.; Schwieter, U.;
Paasivirta, J. Acta Chem. Scand. 1968, 22, 1219−1240.
(35) Schuster, W.; Mikler, H.; Komarek, K. L. Monatsh. Chem. 1979,
110, 1153−1170.
(36) Margadonna, S.; Takabayashi, Y.; McDonald, M. T.;
Kasperkiewicz, K.; Mizuguchi, Y.; Takano, Y.; Fitch, A. N.; Suard,
E.; Prassides, K. Chem. Commun. 2008, 5607−5609.
(37) Janaki, J.; Geetha Kumary, T.; Mani, A.; Kalavathi, S.; Reddy, G.;
Narasimha Rao, G.; Bharathi, A. J. Alloys Compd. 2009, 486, 37−41.
7376
dx.doi.org/10.1021/ic300798p | Inorg. Chem. 2012, 51, 7370−7376