Synthesis of a Cu-Filled Rh17S15 Framework: Microwave Polyol Process Versus High-Temperature Route

Article abstract of DOI:10.1021/acs.inorgchem.7b01102

Metal-rich, mixed copper-rhodium sulfide Cu_3-δRh₃₄S₃₀ that represents a new Cu-filled variant of the Rh₁₇S₁₅ structure has been synthesized and structurally characterized. Copper content in the [CuRh₈] cubic cluster was found to vary notably dependent on the chosen synthetic route. Full site occupancy was achieved only in nanoscaled Cu₃Rh₃₄S₃₀ obtained by a rapid, microwave-assisted reaction of CuCl, Rh₂(CH₃CO₂)₄ and thiosemicarbazide at 300 °C in just 30 min; whereas merely Cu-deficient Cu_3-δRh₃₄S₃₀ (2.0 ≥ δ ≥ 0.9) compositions were realized via conventional high-temperature ceramic synthesis from the elements at 950 °C. Although Cu_3-δRh₃₄S₃₀ is metallic just like Rh₁₇S₁₅, the slightly enhanced metal content has a dramatic effect on the electronic properties. Whereas the Rh₁₇S₁₅ host undergoes a superconducting transition at 5.4 K, no signs of the latter were found for the Cu-derivatives at least down to 1.8 K. This finding is corroborated by the strongly reduced density of states at the Fermi level of the ternary sulfide and the disruption of long-range Rh-Rh interactions in favor of Cu-Rh interactions as revealed by quantum-chemical calculations.

Full text of DOI:10.1021/acs.inorgchem.7b01102

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Article
pubs.acs.org/IC
Synthesis of a Cu-Filled Rh17S15 Framework: Microwave Polyol
Process Versus High-Temperature Route
Maria Roslova,*,† Pavlo Golub,† Lars Opherden,‡,§ Alexander Ovchinnikov,∥ Marc Uhlarz,‡
Alexey I. Baranov,†,∥ Yurii Prots,∥ Anna Isaeva,† Mauro Coduri,⊥ Thomas Herrmannsdorfer,‡
̈
Joachim Wosnitza,‡,§ Thomas Doert,† and Michael Ruck†,∥
§
†Department of Chemistry and Food Chemistry and Institute for Solid State Physics, Technische Universitat Dresden, 01062
Dresden, Germany
̈
∥Max Planck Institute for Chemical Physics of Solids, 01187 Dresden, Germany
‡Dresden High Magnetic Field Laboratory, Helmholtz-Zentrum Dresden-Rossendorf, 01314 Dresden, Germany
⊥ESRF−The European Synchrotron, 38043 Grenoble, France
S
* Supporting Information
ABSTRACT: Metal-rich, mixed copper−rhodium sulfide Cu3−δRh34S30 that
represents a new Cu-filled variant of the Rh17S15 structure has been synthesized
and structurally characterized. Copper content in the [CuRh8] cubic cluster was
found to vary notably dependent on the chosen synthetic route. Full site occupancy
was achieved only in nanoscaled Cu3Rh34S30 obtained by a rapid, microwave-assisted
reaction of CuCl, Rh2(CH3CO2)4 and thiosemicarbazide at 300 °C in just 30 min;
whereas merely Cu-deficient Cu3−δRh34S30 (2.0 ≥ δ ≥ 0.9) compositions were
realized via conventional high-temperature ceramic synthesis from the elements at
950 °C. Although Cu3−δRh34S30 is metallic just like Rh17S15, the slightly enhanced
metal content has a dramatic effect on the electronic properties. Whereas the Rh17S15
host undergoes a superconducting transition at 5.4 K, no signs of the latter were found for the Cu-derivatives at least down to 1.8
K. This finding is corroborated by the strongly reduced density of states at the Fermi level of the ternary sulfide and the
disruption of long-range Rh−Rh interactions in favor of Cu−Rh interactions as revealed by quantum-chemical calculations.
decreases the critical temperature down to ca. 3.5 K.15 Although
the previous works assumed that substitution only takes place
within the rhodium sublattice, the Rh17S15 framework offers
enough room for additional guest atoms, for example, in the
1. INTRODUCTION
Metal-rich chalcogenides of transition metals lie in focus of a
vast research playground thanks to their various electronic
properties stemming from metal−metal interactions, including
rhodium-enclosed cubic voids that may host relatively small
cations. Coinage metals may be promising candidates,
especially in view of the reported significant amounts of copper
in the naturally grown samples of miassite mineral (Rh17S15).16
On top of that, minor solubility of copper in the rhodium
sulfide in the temperature range between 540 and 900 °C has
also been accounted for ref 17. To the best of our knowledge,
effects of copper doping on the physical properties of Rh17S15
have not been investigated yet.
The present work is filling in this gap, as it provides a
comprehensive characterization of the crystal structure and
physical properties of nanoscaled and bulk Cu3−δRh34S30 phases
that were synthesized by a microwave-assisted approach and
the conventional high-temperature route, respectively. The
samples were studied by means of powder X-ray diffraction
(PXRD) and scanning electron microscopy (SEM). Magnetic
susceptibility and transport measurements down to 1.8 K
complex magnetic ordering patterns, superconductivity, charge
density waves, etc. However, nanoscaling of these materials
generally represents a considerable synthetic challenge. Among
downscaling techniques, microwave-assisted polyol synthesis is
steadily gaining popularity thanks to such benefits as rapidness,
high product yield, and a higher degree of product crystallinity
as compared to syntheses in other solvents.1−5 Nanosized
noble-metal chalcogenides are envisioned for many applica-
tions, for instance, as catalysts with high activity,6−9 and they
may also demonstrate physical properties differing from those
of the bulk analogues due to confinement effects.10
Bulk Rh17S15 exhibits a rare case of superconductivity arising
from the presence of strongly correlated 4d electrons of
rhodium at the Fermi level and demonstrates a critical field
(Hc2) value well above the Pauli paramagnetic limit.11−13 Ever
since superconductivity in Rh17S15 was discovered, a number of
attempts to study the influence of chemical composition on the
critical parameters have been repeatedly undertaken.14,15 It was
shown that partial substitution of rhodium atoms by iron, nickel
or palladium14 hinders superconductivity at least down to 1.5
K, whereas iridium doping (up to ca. 6 atom %) considerably
Received: May 4, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.7b01102
Inorg. Chem. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Inorganic Chemistry
Article
2.4. Electron Microscopy. The SEM images were collected on a
Hitachi SU8020 microscope. An acceleration voltage of 2 kV and a
current of 10 μA were used to generate the SEM images in the
secondary electron scanning mode. The energy-dispersive X-ray
spectra (EDX) were collected using an Oxford Silicon Drift X-MaxN
detector at an acceleration voltage of 30 kV and with 100 s of
accumulation time. For the determination of the composition, ∼10
points per sample were measured. Transmission electron microscopy
(TEM) investigations were performed in an FEI Tecnai-T20
microscope operated at 200 kV. For a specimen preparation
Cu3Rh34S30 nanoparticles were dispersed in ethanol with an ultrasonic
bath and then deposited on a holey carbon grid.
2.5. Resistivity and Magnetization Measurements. The
resistivity measurements were performed using a 14 T Quantum
Design PPMS equipped with a variable-temperature insert for cooling
to 1.8 K. The resistance was measured using the standard four-probe
technique. Magnetization was measured using an MPMS rf-SQUID-
magnetometer. For both resistivity and magnetization measurements
the powder was pressed to a pellet for 1 min with the applied pressure
of 0.1 GPa.
revealed metallic behavior and quenching of the super-
conducting transition.
One of the highlights of Cu3−δRh34S30 is a filled-up cubic void
in the Rh17S15 framework and the consequent formation of the
extremely rare, intermetallic [CuRh8] cluster18,19 that has not
been observed in covalently bonded systems so far, let alone
transition-metal chalcogenides. Chemical bonding between the
incorporated Cu atoms and the Rh−S framework in
Cu3Rh34S30 was characterized by delocalization indices20,21
and domain-averaged Fermi-hole analysis.22,23
Another remarkable feature is the varying occupancy of the
copper site that appears to correlate with the chosen synthetic
route. Complete filling could only be achieved via a low-
temperature, microwave-assisted polyol process, which, more-
over, provided Cu3Rh34S30 in the shape of nanoparticles with
the average diameter of 13 nm. Earlier this advanced technique
has been successfully adapted to abundant elemental-metal
nanoparticles,4,24−27 binary alloys,28−31 and oxygen-containing
materials,32−35 but there have been only a handful of reports on
preparation of oxygen-free ternary compounds.36−40 Metal-rich
Pd3Bi2S2 and Pd3Bi2Se2 nanoparticles are among the few recent,
successful applications of the method.10 The first instance of
nanosized Cu3Rh34S30 reported herewith can be also juxtaposed
with the earlier attempts to produce sulfides in the form of
nanoparticles in the Cu−Rh−S system by another wet
chemistry approach. The latter did not result in any ternary
compounds but yielded core−shell Cu1.94S@Rh2S3 nano-
particles with the Cu1.94S nanoprisms covered by Rh2S3
instead.41
2.6. Electronic Structure Calculations. Band structure calcu-
lations were performed for Cu3Rh34S30, with the respective lattice
parameters and atomic positions obtained from the powder X-ray
refinement. Scalar relativistic all-electron full-potential local orbital
calculations employing LDA PW92 functional43 were performed by
using the FPLO-9 code.44 The Blochl-corrected linear tetrahedron
̈
method45 with a 12 × 12 × 12 k-mesh was employed after checking for
convergence with respect to the number of k-points. Scalar-relativistic
projector augmented wave (PAW) calculations46 were performed with
the ABINIT code47,48 utilizing the generalized gradient approximation
(GGA) with the Perdew−Burke−Ernzerhof (PBE) functional.49
Modified ABINIT PAW data sets50 have been used with plane-wave
energy cutoff of 18 hartree and a 2 × 2 × 2 k-point grid. Subsequent
analysis of chemical bonding was performed within the framework of
the quantum theory of atoms in molecules (QTAIM).51 Electron
sharing indices20,21 and domain-average Fermi-hole (DAFH) anal-
ysis22,52 were evaluated employing domain overlap matrices computed
from PAW calculation results according to ref 53.
2. EXPERIMENTAL SECTION
2.1. Microwave-Assisted Synthesis. Ethylene glycol (Fluka,
99%) was dried under vacuum at 100 °C for 4 h prior to use. The
starting materials CuCl (Alfa Aesar, 99.9%), Rh2(CH3CO2)4 (Alfa
Aesar, 98%+), and thiosemicarbazide (Sigma-Aldrich, 99%+) were
obtained from commercial sources and were used without purification.
Microwave synthesis was performed in a dynamic mode in a CEM
Discover System operating at 300 W and 2.45 GHz. Solution (15 mL)
containing precursors was placed in a 35 mL Pyrex tube with a silicon
cap and heated to 300 °C in 3 min and held at this temperature for 30
min. After the termination of the reaction, the tube was rapidly cooled
using a high-pressurized air flow. The obtained black precipitate was
washed several times with ethanol to remove the residual solvent.
Finally, the product was dried under vacuum at room temperature
overnight. The sample obtained by the microwave-assisted synthesis
will be referred to further in the text as Sample 1.
2.2. High-Temperature Synthesis. Cu3−δRh34S30 samples were
prepared from the elements following the conventional solid-state
route. Fine powders of Cu (ABCR, 99.9%), Rh (Merck, 99.9%), and S
(Alfa Aesar, 99.99%) taken in the stoichiometric ratio of the target
compound were placed into a silica ampule. The evacuated ampule
was cooled with liquid nitrogen during sealing to prevent evaporation
of sulfur. The sealed ampule was heated in a muffle furnace to 950 °C
with 10 °C/h heating rate and held at this temperature for 7 d. After
that, the sample was removed from the furnace and allowed to cool to
room temperature. Samples obtained by high-temperature synthesis
are labeled as Samples 2−6.
For comparison, the electronic structure of the parent compound
Rh17S15 was also calculated within the same scheme as described above
for Cu3Rh34S30 but using the reported structural data.54 Although
Rh17S15 is believed to be a correlated superconductor,11−13 the LDA
density of states for this compound was found to reproduce the shape
of the photoemission spectrum and to give an accurate estimate of the
Sommerfeld coefficient.55
3. RESULTS AND DISCUSSION
3.1. Synthesis. The route used for the synthesis of
Cu3Rh34S30 employs a polyol reduction of the Rh3+ and Cu+
precursors under microwave irradiation in the presence of
thiosemicarbazide as a sulfur source. It is worth noting that the
exact reaction pathway is still unknown; however, possible key
factors for the formation of Cu3Rh34S30 can be identified. The
polyol process favors a consecutive reduction of Rh3+ and Cu+
to metallic intermediates, while the reduction rate is to a large
extend determined by the redox potential of the corresponding
redox couples. The growth and nucleation of the metallic
particles are accompanied by a reaction with sulfur ions released
by the thiosemicarbazide feedstock. Because of stability of
complexes formed by thiosemicarbazide and both rhodium and
copper cations56−59 the polyol-mediated conversation occurs at
relatively high temperature (300 °C) that leads to formation of
relatively small particles. In the kinetically driven, non-
equilibrium process, copper might in fact be rendered in a
rhodium−sulfur framework enabling the environment for the
Cu3Rh34S30 ternary phase formation. It is interesting to note
that the microwave synthesis in the binary Rh−S system does
2.3. Powder X-ray Diffraction. PXRD measurements of the
samples were performed on a PANalytical X’Pert Pro diffractometer
(Cu Kα1 radiation; λ = 1.540 56 Å; germanium monochromator), a
flat-plate silicon sample holder was used. LaB6 was used as an internal
standard. Temperature-dependent synchrotron powder diffraction data
(λ = 0.400 01 Å) were collected at the ID22 beamline of the European
Synchrotron Radiation Facility (ESRF), Grenoble. The structure
refinement was performed with the JANA2006 software.42
B
DOI: 10.1021/acs.inorgchem.7b01102
Inorg. Chem. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Inorganic Chemistry
Article
Figure 1. Morphology of the Cu3Rh34S30 sample obtained by microwave synthesis. SEM, low-resolution and high-resolution (HRTEM) images of
the nanoparticles (upper). The dimensioning in the HRTEM image corresponds to ∼1 nm. EDX mapping (lower) shows the presence of uniformly
distributed Cu, Rh, and S.
a
Table 1. Crystallographic and EDX Data for the Cu3−δRh34S30 Compounds
sample
1
2
3
4
5
6
Cu/Rh/S starting ratio
3:34:30
1:34:30
2:34:30
3:34:30
3:34:30
0.96(1)
0.9(1)
b
δ from Rietveld
0
2.10(2)
1.32(2)
0.94(2)
1.94(1)
δ from EDX
0
1.98(6)
1.18(8)
0.90(1)
a (Å)
V (Å3)
9.9891(2)
996.72(1)
9.9370(1)
981.22(1)
9.9544(1)
986.40(1)
9.9660(1)
989.79(1)
9.9664(1)
989.95(1)
9.9407(1)
982.33(3)
calculated density
(g/cm3)
7.75
7.64
7.69
7.70
7.70
7.66
Rp, %
5.44
6.44
5.85
6.73
6.55
6.55
R1, wR2 [I > 2σ(I)], %
R1, wR2 (all data), %
GOF
1.91, 2.16
2.02, 2.17
1.13
2.52, 3.17
2.65, 3.18
1.18
2.72, 2.32
2.85, 3.33
1.19
3.10, 3.47
3.17, 3.48
1.49
4.48, 5.04
4.56, 5.05
1.13
4.48, 5.04
4.56, 5.05
1.13
largest peak/hole (e/Å3)
1.2/−1.63
0.90/−0.91
1.62/−1.29
3.82/−1.34
2.23/−4.45
2.23/−4.45
a
b
Space group Pm3m (221), Z = 1, 2θ range from 7 to 130°, Cu Kα1 radiation. Uiso for all the Rh atoms as well as Uiso for all the S atoms are
̅
constrained. Uiso(Cu1) = Uiso(Rh*) constrained.
not lead to the formation of pure Rh17S15 but to an amorphous
phase consisting of particles with varying Rh/S ratio as
determined by EDX. Several consecutive annealing steps at
500 °C are required for the formation of crystalline Rh17S15. In
contrast, Cu3Rh34S30 exhibits a high degree of crystallinity even
in the form of as-prepared nanoparticles.
3.2. Morphology. Figure 1 shows a typical morphology of
the as-prepared Cu3Rh34S30 nanoparticles (Sample 1, see Table
1). Particle-size calculations employing the Scherrer equation60
whereas the calculated lattice parameter was ∼1% larger than
that of Rh17S15 (a = 9.9093(2) Å54). We started the refinement
by taking the atomic coordinates of Rh17S15 as a model. The
refinement reached convergence with Rp = 7.63 and RB = 10.57,
but the residual electron density was high, indicating that the
model was not complete. Visualization revealed that the highest
positive maximum in the difference-Fourier map is located in
the position (0, 1/2, 1/2), in the center of a cube built by the
Rh1 atoms (Figure 2). No significant positive maxima were
found close to the Rh or S sites.
yield an average particle size of 13.5
1.1 nm. TEM study
showed that the nanoparticles tend to form spongelike
networks that impede shape and size estimation. Individual
nanoparticles have irregular shapes and often can be a subject
to twinning and intergrowth, whereas their diameters hardly
reach 10−20 nm. The high-resolution imaging of individual
particles further revealed that the nanoparticles demonstrate a
good crystallinity. The lattice spacing is ∼1 nm, which is
perfectly in line with the crystal structure of Cu3Rh34S30
determined below.
3.3. Crystal Structure Determination. Initial indexing of
the PXRD pattern collected with laboratory X-ray source for
Sample 1, prepared by the microwave-assisted synthesis (see
Table 1), suggested a primitive cubic unit cell with the lattice
parameter a = 9.9891(2) Å. The powder pattern resembled that
of Rh17S15 generated from the reported structural data,54
From the X-ray data only, it was not possible to conclude
which chemical species occupies the mentioned void. There-
fore, additional experiments were undertaken to explore the
quasi-binary sections Rh17S15−Rh, Rh17S15−S, and Rh17S15−Cu
via a high-temperature route. It was revealed that Rh or S excess
leads to formation of impurities, namely, metallic Rh and Rh3S4,
respectively, confirming that Rh17S15 is a line compound.61 At
the same time, annealing of the Rh17S15−Cu mixtures yielded
cubic phases with a varying lattice parameter dependent on the
Cu content (Samples 2−5). It is worthwhile to note that minor
solubility of Cu in Rh17S15 was reported,17 but the Cu site
preference in this structure was not determined. The crystal
structure refinement of these cubic Cu3−δRh34S30 phases
employing the crystal structure of Rh17S15 as the initial model
produced a high residual electron density peak in the center of
C
DOI: 10.1021/acs.inorgchem.7b01102
Inorg. Chem. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Inorganic Chemistry
Article
atoms located slightly above each of the square faces. The latter
atoms link the type A clusters with Rh3-centered
[Rh3S14Rh22/2] octahedra (cluster B) via common corners,
so that the cuboctahedral and octahedral clusters form a ReO3-
type arrangement. The cubic void in this arrangement is
occupied by a [Rh4S36] octahedron (cluster C). The resulting
octahedral interstitials framed by four B clusters and two C
clusters accommodate isolated Cu-centered [Cu1Rh18] cubes
(cluster D).The distance between the Cu1 and Rh1 atoms is
found to be 2.556(1) Å in Cu3Rh34S30. Similar [CuRh8] units
occur also in the CuRh2Sn Heusler compound18 as well as in
the Al0.82Cu0.11Rh alloy19 with the CsCl structure. However, the
distance between the Cu and Rh atoms in Cu3Rh34S30 is
significantly shorter than in the intermetallic compounds listed
above (2.661(1) Å and 2.573(1) Å for CuRh2Sn and
Al0.82Cu0.11Rh, correspondingly).
Figure 2. Section 0 ≤ x, y ≤ 1 in z = 0 of the difference Fourier map
after the initial Rietveld refinement for Sample 3. The maximum
corresponds to the position of the Cu atom, which is not included in
the structural model. The Rh1 atoms are shifted out of the cutoff plane
by 1.45 Å, whereas the Rh2, Rh3, and S2 atoms lie in it.
PXRD analysis was performed to monitor changes in the
crystal structure caused by the Cu incorporation. Apparently,
there is a correlation between the lattice parameter and the Cu
content in the Cu3−δRh34S30 samples, namely, the lattice
parameter increases with the increasing amount of copper
(Figure 4). However, it is important to note that only samples
the [Rh1]8 cubes, similarly to the case of Sample 1. This
position was refined as partially occupied by Cu. In the case of
Sample 1, prepared by microwave-assisted route, the 3c
Wyckoff site lying in (0, 1/2, 1/2) was found to be fully
occupied. The Cu content derived from the structure analysis is
in good agreement with the EDX results. The crystallographic
data and refinement parameters are summarized in Table 1.
The atomic coordinates, the atomic displacement parameters,
as well as the list of selected interatomic distances are given in
the Supporting Information (Tables S1 and S2). Typical XRD
patterns for the freshly produced nanoparticles as well as for the
Cu3−δRh34S30 samples, obtained by high-temperature synthesis,
are given in Figures S1 and S2.
It is interesting to note that the presence of an unaccounted
electron density maximum with (0, 1/2, 1/2) fractional
coordinates was reported in the crystallographic report on the
Ni9.54Pd7.46S15 compound with mixed Ni/Pd sites.62 Our
findings allow us to suggest that Ni atoms may also occupy
the cubic void in the structure similarly to Cu in Cu3−δRh34S30.
Thus, Cu3−δRh34S30 has the filled Rh17S15 crystal structure.
The complex framework can be visualized as consisting of four
types of polyhedra (Figure 3). Empty [S212Rh26/2] cages
(cluster A) represent cuboctahedra built of S2 atoms with Rh2
Figure 4. Dependence of the lattice parameter (at room temperature)
from the Cu content (derived from PXRD data) in the series of
Cu3−δRh34S30 samples. Sample 1 was obtained by a microwave-assisted
polyol process at 300 °C, Samples 2 to 4 were obtained by high-
temperature synthesis, and Samples 5 and 6 were obtained in the same
batch at 1100 °C. The Cu content of Sample 6 was estimated by fitting
the lattice parameter values (given in Table 1).
with δ ≥ 0.9 were accessible by high-temperature synthesis.
Copper excess in the starting composition results in emergence
of the Cu2−xS (PDF No. 29−578) impurity (synthesis
temperature 950 °C, Sample 3). At the synthesis temperature
close to the melting point (∼1100 °C), two Cu3−δRh34S30
phases with different Cu occupancies can be obtained in the
same batch, accompanied by Rh (PDF No. 5−685) and Rh3S4
(ICSD No. 410813) impurities. One of the Cu3−δRh34S30
phases corresponds to the maximal observed Cu concentration
(δ ≈ 0.9, Sample 5), and its lattice parameter a = 9.9665(1) Å
resembles that for Sample 4; the second one with a = 9.9414(1)
Å (Sample 6) corresponds to δ ≈ 2.0 (see Figure 4). Thus, we
can conclude that the maximal Cu occupancy achievable by
high-temperature synthesis at 950 °C is close to 70% (δ ≈ 0.9).
The microwave-promoted chemistry is able to overcome this
limitation. The Cu3Rh34S30 sample (Sample 1), prepared in the
polyol process, has a significantly larger cell parameter than
Figure 3. Crystal structure of Cu3−δRh34S30 in the polyhedron
representation. Rh and S atoms are shown in green and orange,
respectively. Cu atoms occupy the centers of the blue cubes.
D
DOI: 10.1021/acs.inorgchem.7b01102
Inorg. Chem. XXXX, XXX, XXX−XXX