Synthesis, crystal structure, and electronic properties of the tetragonal (REIREII)3SbO3 Phases (RE I = La, Ce; REII = Dy, Ho)

Article abstract of DOI:10.1021/ic302292w

In our efforts to tune the charge transport properties of the recently discovered RE₃SbO₃ phases (RE is a rare earth), we have prepared mixed (RE^IRE^II)₃SbO₃ phases (RE^I = La, Ce; RE^II = Dy, Ho) via high-temperature reactions at 1550 C or greater. In contrast to monoclinic RE₃SbO ₃, the new phases adopt the P4₂/mnm symmetry but have a structural framework similar to that of RE₃SbO₃. The formation of the tetragonal (RE^IRE^II)₃SbO ₃ phases is driven by the ordering of the large and small RE atoms on different atomic sites. The La_1.5Dy_1.5SbO₃, La_1.5Ho_1.5SbO₃, and Ce_1.5Ho _1.5SbO₃ samples were subjected to elemental microprobe analysis to verify their compositions and to electrical resistivity measurements to evaluate their thermoelectric potential. The electrical resistivity data indicate the presence of a band gap, which is supported by electronic structure calculations.

Full text of DOI:10.1021/ic302292w

Article
pubs.acs.org/IC
Synthesis, Crystal Structure, and Electronic Properties of the
Tetragonal (REIREII)3SbO3 Phases (REI = La, Ce; REII = Dy, Ho)
Scott Forbes,† Peng Wang,† Jinlei Yao,†,§ Taras Kolodiazhnyi,‡ and Yurij Mozharivskyj*,†
†Department of Chemistry and Chemical Biology, McMaster University, 1280 Main Street West, Hamilton, Ontario, L8S 4M1,
Canada
‡Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan
§School of Mathematics and Physics, Suzhou University of Science and Technology, Suzhou 215009, China
S
* Supporting Information
ABSTRACT: In our efforts to tune the charge transport properties of
the recently discovered RE3SbO3 phases (RE is a rare earth), we have
prepared mixed (REIREII)3SbO3 phases (REI = La, Ce; REII = Dy, Ho)
via high-temperature reactions at 1550 °C or greater. In contrast to
monoclinic RE3SbO3, the new phases adopt the P42/mnm symmetry
but have a structural framework similar to that of RE3SbO3. The
formation of the tetragonal (REIREII)3SbO3 phases is driven by the
ordering of the large and small RE atoms on different atomic sites. The
La1.5Dy1.5SbO3, La1.5Ho1.5SbO3, and Ce1.5Ho1.5SbO3 samples were
subjected to elemental microprobe analysis to verify their compositions and to electrical resistivity measurements to evaluate
their thermoelectric potential. The electrical resistivity data indicate the presence of a band gap, which is supported by electronic
structure calculations.
In pursuit of novel PGEC phases, our group has focused on
the synthesis of complex rare-earth (RE) antimonide suboxides,
such as RE3SbO3, RE8Sb3−δO8, and RE2SbO2.12,13 In the case of
RE3SbO3, we have combined the electrically conductive,
semimetallic RESb and thermally insulating RE2O3 binaries
into a single structure, with the goal of conserving the desirable
properties of each. The new RE3SbO3 phases feature high
thermal stability and low thermal conductivity but possess high
electrical resistivities, which make them unsuitable for thermo-
electric applications. Although the desired properties could not
be obtained, we have succeeded in the direct combination of
RESb and RE2O3 and the opening of a band gap within the
RESb framework. As such, there is a potential for the RE3SbO3
phases if their electrical resistivities can be reduced to a more
suitable level, either by elemental substitution or further
structural modifications.
Originally, we explored substitution of Ce for the trivalent
RE in RE3SbO3. However, it was established that Ce is also
trivalent in the prepared (CeRE)3SbO3 phases. These new,
mixed phases were discovered to have tetragonal symmetry,
while the parent RE3SbO3 structures were monoclinic. The
subsequent experiments revealed that this structural change is
possible only through the use of two RE atoms with a
sufficiently large difference in atomic radii, which results in the
preferential ordering of large and small rare-earths on different
atomic sites. In this paper, we discuss the synthetic conditions,
characterization, and electronic structures of the new,
INTRODUCTION
■
The ever-increasing global demand for natural resources
continues to fuel the research on and development of new
and efficient ways for energy use. One potential approach is the
use of thermoelectric materials for energy recovery from the
waste heat, released by internal combustion engines, e.g., in cars
and trucks.1,2 For such applications, thermoelectrics yield many
advantages over other forms of energy recovery and production
such as compactness, reliability, a lack of moving parts, and a
long operational life. These properties have already allowed
thermoelectrics to be used in some unique applications, such as
power generation on spacecrafts3 or wristwatches.4
The current drawback of thermoelectric materials is their low
efficiency, which reaches 5−15% of the Carnot efficiency in the
best materials known.5 As such, thermoelectrics have only seen
use in small, niche applications in which compactness and
reliability outweigh efficiency. A good thermoelectric material
must have a high Seebeck coefficient and electrical conductivity
and a low thermal conductivity. Unfortunately, the Seebeck
coefficient and electrical conductivity of a material compromise
each other and usually cannot both be optimized simulta-
neously. Likewise, designing a material with good electrical
properties and a low thermal conductivity presents a great
challenge due to the thermal conductivity provided by charge
carriers. Different strategies have been employed to optimize
thermoelectric properties. One such method is to prepare
phases with electrical properties of a crystalline solid but with
thermal conductivities of a glass.6 Such a material, known as a
phonon-glass electron-crystal (PGEC), has been realized in few
systems, e.g., in Zn4Sb3
Received: October 18, 2012
Published: January 8, 2013
7−9
and PbTe-based phases.10,11
© 2013 American Chemical Society
1025
dx.doi.org/10.1021/ic302292w | Inorg. Chem. 2013, 52, 1025−1031

Inorganic Chemistry
Article
Table 1. Phase Analysis of the (REIREII)3SbO3 Samples Prepared at 1550 °C
system
REI2.5REII0.5SbO3
REI2REIISbO3
REI1.5REII1.5SbO3
REIREII2SbO3
REI0.5REII2.5SbO3
Ce−Gd−Sb−O
Ce−Ho−Sb−O
Ce−Er−Sb−O
La−Dy−Sb−O
La−Ho−Sb−O
-
-
-
-
-
-
Monoclinic RE3SbO3
-
-
Tetragonal RE3SbO3
Monoclinic RE3SbO3 and RE2SbO2
RE2SbO2
-
Tetragonal RE3SbO3 and RE2SbO2
-
Monoclinic RE3SbO3 Monoclinic RE3SbO3 Tetragonal RE3SbO3
Monoclinic RE3SbO3 Monoclinic RE3SbO3 Tetragonal RE3SbO3
Tetragonal RE3SbO3
RE2SbO2
RE2SbO2
RE2SbO2
Table 2. Crystallographic and Refinement Data for the (REIREII)3SbO3 Crystals
La1.5Dy1.5SbO3
LaDy2SbO3
La1.5Ho1.5SbO3
Ce1.5Ho1.5SbO3
Ce1.34(9)Ho1.66(9)SbO3
refined composition
space group
a (Å)
La1.47(7)Dy1.53(7)SbO3
P42/mnm
La0.96(9)Dy2.04(9)SbO3
P42/mnm
La1.32(4)Ho1.68(4)SbO3
P42/mnm
P42/mnm
11.947(1)
11.864(2)
11.920(2)
11.880(2)
3.8906(8)
549.1(2)
c (Å)
volume (Å3)
3.9263(8)
3.8920(8)
3.9057(8)
560.4(1)
547.8(3)
555.0(2)
Z
4
4
4
4
index ranges
−19 ≤ h ≤ 15
−18 ≤ k ≤ 18
−6 ≤ l ≤ 6
6.82−68.96°
6684
−15 ≤ h ≤ 18
−18 ≤ k ≤ 18
−6 ≤ l ≤ 6
6.68−69.01°
4159
−18 ≤ h ≤ 15
−18 ≤ k ≤ 18
−6 ≤ l ≤ 5
6.84−68.70°
5522
−18 ≤ h ≤ 18
−18 ≤ k ≤ 14
−6 ≤ l ≤ 6
6.86−68.96°
5304
2θ range
total reflections
goodness-of-fit on F2
R indices
1.187
1.169
1.091
1.029
R1 = 0.0378
wR2 = 0.0760
R1 (all data) = 0.0532
0.0026(2)
R1 = 0.0557
wR2 = 0.0759
R1 (all data) = 0.0883
0.0012(1)
R1 = 0.0380
wR2 = 0.0445
R1 (all data) = 0.0625
0.00175(9)
R1 = 0.0519
wR2 = 0.0617
R1 (all data) = 0.1019
0.00073(6)
extinction coefficient
Table 3. Atomic Parameters for the (REIREII)3SbO3 Crystals
atom
site
occupancy
x
y
z
Ueq
La1.47(7)Dy1.53(7)SbO3
0.11537(4)
La/Dy1
La/Dy2
Sb1
4f
8i
4f
4g
8i
0.93/0.07(3)
0.11537(4)
0.65651(5)
0.37828(5)
0.6896(5)
0.3383(8)
0
0
0
0
0
0.0119(2)
0.0135(2)
0.0124(3)
0.013(2)
0.028(2)
0.27/0.73(3)
0.12314(4)
1
1
1
0.37828(5)
O1
0.3104(5)
O2
0.0660(7)
La0.96(9)Dy2.04(9)SbO3
0.11601(7)
La/Dy1
La/Dy2
Sb1
4f
8i
4f
4g
8i
0.75/0.25(4)
0.11601(7)
0.65631(6)
0.37824(8)
0.6890(8)
0.339(1)
0
0
0
0
0
0.0135(2)
0.0116(2)
0.0123(4)
0.016(3)
0.025(3)
0.11/0.89(4)
0.12298(7)
1
1
1
0.37824(8)
O1
0.3110(8)
O2
0.0660(9)
La1.32(4)Ho1.68(4)SbO3
0.11539(5)
La/Ho1
La/Ho2
Sb1
4f
8i
4f
4g
8i
0.92/0.08(2)
0.11539(5)
0.65604(4)
0.37756(5)
0.6900(5)
0.3394(8)
0
0
0
0
0
0.0119(2)
0.0139(1)
0.0119(2)
0.012(2)
0.029(2)
0.20/0.80(2)
0.12319(4)
1
1
1
0.37756(5)
O1
0.3100(5)
O2
0.0660(6)
Ce1.34(9)Ho1.66(9)SbO3
0.11523(9)
La/Ce1
La/Ce2
Sb1
4f
8i
4f
4g
8i
0.88/0.12(4)
0.11523(9)
0.65599(6)
0.37815(9)
0.6878(8)
0.339(1)
0
0
0
0
0
0.0130(3)
0.0130(2)
0.0121(4)
0.013(3)
0.032(3)
0.23/0.77(4)
0.12372(8)
1
1
1
0.37815(9)
O1
0.3122(8)
O2
0.067(1)
tetragonal (REIREII)3SbO3 phases (REI = La, Ce; REII = Dy,
Ho).
̂
wt. %, Rhone-Poulenc). To avoid the volatility of elemental antimony
and its reactivity with tantalum at elevated temperatures, the
corresponding RESb binaries were first prepared. The RESb binaries
were prepared by mixing RE filings with ground antimony powder in
equimolar amounts within a glovebox filled with argon. These mixtures
were then pressed into pellets of 1−2 g and sealed below 10−5 Torr in
silica tubes. Samples were annealed in box furnaces at a temperature of
600 °C for 12 h and then at 850 °C for 48 h to complete the reaction.
EXPERIMENTAL SECTION
■
Synthesis. Samples were made using high-purity RE metals (99.99
wt. %, SmartElements) (RE = La, Ce, Dy, Ho), antimony metal
(99.999 wt. %, CERAC Inc.), and RE2O3 and CeO2 powder (99.999
1026
dx.doi.org/10.1021/ic302292w | Inorg. Chem. 2013, 52, 1025−1031

Inorganic Chemistry
Article
a
Table 4. Compositions of the Tetragonal (REIREII)3SbO3 Phases Based on the EPMA and the X-ray Single-Crystal Analyses
loading Composition
EPMA
X-ray
La1.5Dy1.5SbO3
La1.5Ho1.5SbO3
Ce1.5Ho1.5SbO3
La1.45(7)Dy1.55(7)Sb1.00(8)O6.42(5)
La1.54(6)Ho1.52(6)Sb1.00(8)O5.98(5)
Ce1.54(6)Ho1.47(6)Sb1.00(7)O5.13(5)
La1.47(7)Dy1.53(7)SbO3
La1.32(4)Ho1.68(4)SbO3
Ce1.34(9)Ho1.66(9)SbO3
a
The EPMA results were normalized to one Sb atom.
Upon quenching in air, black pellets of RESb were obtained. X-ray
powder diffraction was performed on all samples to confirm their
purity.
Coordination Polyhedra Analysis. Coordination polyhedra of
the new tetragonal (REIREII)3SbO3 structures, the previously studied
monoclinic Ho3SbO3 phase,12 and the monoclinic Dy3SbO3 phase
prepared for this work were analyzed with the Dido95 program.18 For
this purpose, we have calculated the volumes of the RE sites, which are
represented by the volumes of the Wigner−Seitz polyhedra. Atomic
positions determined from the X-ray single-crystal solutions were used
for calculations.
Microprobe Measurements. The quantitative elemental analysis
of the selected samples was performed employing the electron probe
microanalysis (EPMA) through a wavelength dispersive (WDS) X-ray
spectroscopy (model JXA-8500F, JEOL, Tokyo, Japan). LaB6, CeB6,
Al2O3, Ho5Sb3, and DySb were used as standards to determine the
concentrations of La, Ce, Ho, Sb, Dy, and O in the title compounds.
The EPMA was performed on polished surface samples by averaging
the data taken from 5 to 10 locations of selected grains. Only the
REI1.5REII1.5SbO3 samples were analyzed for measurement as these
samples yielded the best purities.
Electrical Resistivity Measurements. Samples of La1.5Dy1.5SbO3,
La1.5Ho1.5SbO3, and Ce1.5Ho1.5SbO3 were ground and pressed into
pellets with a tungsten carbide press die in a glovebox under argon
atmosphere. Only the REI1.5REII1.5SbO3 set of samples was chosen for
measurement as the highest purities were obtained for this
stoichiometry. Samples were then sealed in evacuated silica tubes
and annealed at 1000 °C for 24 h to ensure rigidity. X-ray powder
diffraction revealed no decomposition or generation of impurities.
The electrical resistivity was measured in the 2−400 K range on
rectangular-shaped samples with a four-probe technique using a press
contact assembly produced by Wimbush Science & Technology and
on a PPMS instrument (Quantum Design, USA). During the
measurements, the samples were heated with a speed of 1 K/min to
allow thermal equilibration with a cryostat.
Samples of REI1.5REII1.5SbO3 were prepared by mixing the RESb
binaries and RE2O3 for each of the respective REs present in a 1:1:1:1
molar ratio. The uneven REI:REII ratio (REIREII)3SbO3 samples were
prepared by mixing the respective RESb and RE2O3 binaries to
produce the desired stoichiometry. The (CeREII)3SbO3 samples were
prepared by mixing filed RE metal, the RESb binaries, and CeO2 in the
desired molar ratio. The respective mixtures were pressed into pellets
of 1 g and placed inside tantalum ampules which were then sealed by
arc-melting under an argon atmosphere. High-temperature reaction
conditions were accomplished through the use of an induction furnace.
The tantalum ampules containing samples were placed inside a
molybdenum susceptor which was heated in the radio frequency
furnace under dynamic vacuum below 10−5 Torr. An optical pyrometer
was used to monitor the temperature of the reaction. All samples were
heated at 1550 °C for 10 h to ensure a complete reaction. Samples
were allowed to cool to room temperature inside the furnace under
vacuum over a period of 1 h. The obtained products were black,
molten, crystalline, and stable in air. According to X-ray powder
diffraction analysis, samples exposed to air for several weeks showed
no signs of decomposition. A list of all samples prepared is presented
in Table 1.
X-ray Single Crystal Diffraction. Single crystals picked up from
the samples were analyzed on a STOE IPDSII diffractometer using Mo
Kα radiation in the whole reciprocal sphere. A numerical absorption
correction was based on the crystal shape originally determined by
optical face indexing but was later optimized against equivalent
reflections using the STOE X-Shape software.14 Crystal structures
were determined and solved using the SHELX software.15 All
structures studied adopt the P42/mnm space group. The lattice
parameters of structures were observed to expand with larger RE
atoms. During refinement, the RE atoms were permitted to mix on the
RE1 and RE2 sites, yielding mixed occupancy on both sites for all
structures. While the occupancies of Dy and Ho on the RE1 site in
La1.47(7)Dy1.53(7)SbO3 and Ce1.34(9)Ho1.66(9)SbO3 are within three
standard deviations from zero, the Hamilton test16 indicated that
their presence on this site can be accepted with a higher than 0.995
confidence level. A summary of the refinement results is presented in
Tables 2 and 3. Further information on the crystal structures of all
compounds presented, including the monoclinic Dy3SbO3 structure
collected for this work, can be obtained from the Fachinformations-
zentrum Karlsruhe, 76344 Eggenstein-Leopoldshafen, Germany (fax
(49) 7247-808-666; e-mail crysdata@fiz.karlsruhe.de), by quoting the
CSD depository numbers 425138 for Ce1.5Ho1.5SbO3, 425139 for
La1.5Dy1.5SbO3, 425140 for La1.5Ho1.5SbO3, 425141 for Dy3SbO3, and
425150 for LaDy2SbO3 samples.
X-ray Powder Diffraction. All samples studied were analyzed by
X-ray powder diffraction using a PANalytical X’Pert Pro diffractometer
with the Cu Kα1 radiation and an X’Celerator detector. This was done
to determine reaction progression, sample purity, and lattice constants.
Samples were ground in 20−50 mg amounts in a mortar and pestle
until a fine powder was obtained. The obtained powder was
distributed evenly on disks manufactured from single crystals of
silicon which were coated with a thin film of Vaseline. Diffraction data
were collected in the 20−70° 2θ range for all samples. The Rietvald
refinement method (Rietica program17) was employed to determine
sample purity and lattice constants. The structural parameters obtained
from the single-crystal refinements were used as starting models.
Electronic Band Structure Calculations. The band structures of
the tetragonal fully ordered REIREII SbO3 phases (LaDy2SbO3,
2
LaHo2SbO3, and CeHo2SbO3) were calculated using the tight-binding,
linear-muffin tin orbital method19 with the atomic sphere approx-
imation (TB-LMTO-ASA) as implemented in the Stuttgart program.20
The experimental lattice and atomic parameters obtained from the X-
ray single-crystal refinements of the LaDy2SbO3, La1.5Ho1.5SbO3, and
Ce1.5Ho1.5SbO3 samples were used for calculations. Since the atomic
mixing cannot be treated during calculations, the RE1 and RE2 sites in
each structure were assumed to be fully occupied by the larger and
smaller RE, respectively. All 4f electrons were considered as core
electrons. Exchange and correlation were treated by the local density
approximation (LDA).21 A scalar relativistic approximation22 was
employed to account for all relativistic effects except spin−orbit
coupling. Overlapping Wigner−Seitz cells were constructed with radii
determined by requiring the overlapping potential to be the best
approximation to the full potential, according to the atomic sphere
approximation (ASA). Automatic sphere generation23 was performed
to construct empty spheres to be included in the unit cell to satisfy the
overlap criteria of the TB-LMTO-ASA model.
RESULTS AND DISCUSSION
■
Composition and Formation of (REIREII)3SbO3. The X-
ray single-crystal refinements of the La1.5Dy1.5SbO3 and
Ce1.5Ho1.5SbO3 crystals yielded compositions (Table 3) that
are within 1 and 2 standard deviations from the loading
compositions, respectively. Refinement of the La1.5Ho1.5SbO3
crystal yielded the La1.32(4)Ho1.68(4)SbO3 composition that is
1027
dx.doi.org/10.1021/ic302292w | Inorg. Chem. 2013, 52, 1025−1031

Inorganic Chemistry
Article
statistically different from the loading one. In some cases, X-ray
single-crystal results may not represent the actual composition,
particularly when an atomic mixing is present and the difference
between the atomic scattering factors is not large. To establish
true compositions (and verify refinement results), electron
probe microanalyses (EPMA) were performed on three
REI1.5REII1.5SbO3 samples (Table 4). Upon inspecting Table
4, it is readily apparent that the equimolar loading ratios of the
RE atoms are preserved across all samples within one standard
deviation. An excess of oxygen is observed, but this can be
attributed to the surface oxidation during polishing. On the
basis of the EPMA results, we can conclude that
(REIREII)3SbO3 phases maintain the same rare-earth compo-
sitions that are used in their formation.
basis of the crystal radii of the rare-earth atoms24 used, the
lowest large-to-small ratio of rare-earth radii shown to produce
the phase is 1.23, accomplished with Ce1.5Ho1.5SbO3.
Furthermore, an attempt to prepare the tetragonal
Ce1.5Gd1.5SbO3 analogue was unsuccessful. Considering the
crystal radii ratio of Ce to Gd is 1.19, it appears the minimum
ratio required may lie between these two values. The reasons
for the formation of the tetragonal structure as opposed to the
monoclinic phase for the mixed (REIREII)3SbO3 phases are
discussed below.
Structures and Preferential Site Occupancy. The
(REIREII)3SbO3 phases of interest adopt a novel crystal
structure with the P42/mnm space group. The structure is
defined by a REO framework composed of edge-sharing RE4O
tetrahedra. Three RE4O tetrahedra are fused via edge-sharing to
form RE8O3 units referred to as building blocks. Within the ab
plane, these blocks connect to each other via corners to build
2D slabs (Figure 1). In turn, the slabs stack along the c
Samples with loading compositions containing different
ratios of REI to REII were also prepared using reaction
conditions similar to the samples with the equimolar ratios. The
synthesis of LaDy2SbO3 (REI = La and REII = Dy) was
confirmed to produce the tetragonal RE3SbO3 structure, but
with a smaller unit cell when compared to La1.5Dy1.5SbO3.
Reactions of the REI2REIISbO3 and REI2.5REII0.5SbO3 phases
yielded only the monoclinic RE3SbO3 phases; these samples
were not analyzed further. In contrast, the REIREII SbO3 and
2
REI0.5REII2.5SbO3 samples, excluding LaDy2SbO3, were shown
to produce the RE2SbO2 phase. The rationalization for forming
each phase at different REI to REII ratios will be discussed later.
The tetragonal (REIREII)3SbO3 phases can be obtained only
at reaction temperatures of 1550 °C and greater; at
temperatures less than 1500 °C, either the corresponding
RE8Sb3−δO8 or RE2SbO2 phase is formed depending on the REI
and REII used and their ratios. When only one type of RE is
used, the resulting phase is always either the monoclinic
RE3SbO3 phase or the monoclinic RE8Sb3‑δO8 phase at
temperatures of at least 1550 °C. The formation conditions
for (REIREII)3SbO3 are summarized in Scheme 1. It is worth
Scheme 1. Procedures of Formation of the Phases Studied in
the RE−Sb−O System
Figure 1. Tetragonal (REIREII)3SbO3 structure and its building block
type composed of RE4O tetrahedra.
direction via edge-sharing to form the 3D REO framework.
Additionally, square and rectangular channels along the c
direction are created by this arrangement. The rectangular
voids are occupied by two Sb atoms, and the square voids are
left unoccupied. The Sb−Sb separations are too large to
support dimer formation, as the shortest distance between
these atoms is ∼4 Å.25 An interesting aspect of the
(REIREII)3SbO3 phases is the position and displacement
parameters for the O2 atoms located in the tetrahedral voids
on each side of a RE8O3 building block. The O2 atoms are
displaced from the center of RE tetrahedra toward the center of
a terminal triangular face. The large displacement parameters of
the O2 atoms are also directed toward the centers of these
terminal triangular faces. Such a feature is likely due to anion−
anion repulsion between O1 and O2 atoms in the same
building block, which pushes the outside O2 atoms further
away from their ideal position at the center of each RE4O unit,
and as a result yields high displacement parameters.
mentioning that prolonged heat treatments below 1550 °C will
irreversibly convert the tetragonal (REIREII)3SbO3 phase into
the corresponding monoclinic RE3SbO3 and RE8Sb3−δO8
phases, with mixing on all RE sites. The tetragonal phase can
be prepared for samples containing two REs of significantly
different sizes, which limits the possible combinations. On the
It is useful to view these phases in a simplified form as
RE3SbO3, which has the same composition as the previously
studied monoclinic RE3SbO3 phases.12 Both structures include
REO frameworks which define the overall structure and empty
channels occupied by Sb atoms. Aside from their differences in
1028
dx.doi.org/10.1021/ic302292w | Inorg. Chem. 2013, 52, 1025−1031

Inorganic Chemistry
Article
Figure 2. Comparison of the tetragonal (REIREII)3SbO3 (left), monoclinic RE3SbO3 (middle), and monoclinic RE8Sb3−δO8 (right) phases. Different
building block types are highlighted with different colors.
symmetry, the two RE3SbO3 phases can be easily distinguished
by their building block types and arrangements (Figure 2). The
previously studied monoclinic RE3SbO3 phase contains two
distinct building blocks, A and B, which are four and two
tetrahedral RE4O units long, respectively. In contrast, the newly
discovered tetragonal (REIREII)3SbO3 phase contains only one
type of building block, which is three RE4O tetrahedra long in
all directions, accounting for its higher symmetry. Both
structures also have large displacement parameters for the
oxygen atoms located at the ends of the building blocks. This
unusual feature appears to be preserved among the rare-earth
antimonide oxide series, as the RE8Sb3−δO8 phase (Figure 2)
also has terminal O atoms with large displacement parameters.
The (REIREII)3SbO3 structure contains two RE sites which
both have mixed occupancy. The RE1 site (4f) located in the
center of each building block is preferentially (>75%) occupied
by larger RE atoms (Table 3). Conversely, the RE2 position
(8i), which makes the outside corners of the RE8O3 building
blocks, strongly (>73%) prefers the smaller RE atoms. The site
preference was analyzed based on the polyhedra volume
(Figure 3) and electronegativity of the REI and REII atoms. In
Table 5. Coordination Polyhedra Volumes of the Largest
and Smallest Sites in the Tetragonal (REIREII)3SbO3 and
Monoclinic RE3SbO3 Phases
smallest RE site largest RE site
relative size
difference
a
a
phase
volume
volume
Tetragonal Phases
La1.47(7)Dy1.53(7)SbO3
La1.32(4)Ho1.68(4)SbO3
Ce1.34(9)Ho1.66(9)SbO3
La0.96(9)Dy2.04(9)SbO3
17.02(3) Å3
16.76(3) Å3
16.73(3) Å3
16.62(3) Å3
21.20(4) Å3
21.15(4) Å3
20.67(4) Å3
20.72(4) Å3
24.6(3)%
26.2(3)%
23.5(3)%
24.7(3)%
Monoclinic Phases
Dy3SbO3
Ho3SbO3
15.98(3) Å3
15.79(3) Å3
18.44(4) Å3
18.28(4) Å3
15.4(3)%
15.8(3)%
a
In tetragonal (REIREII)3SbO3, the largest RE site is RE1, and the
smallest is RE2.
phases. Table 5 gives the site volumes for the monoclinic
Dy3SbO3 and Ho3SbO3 phases in addition to those for the
mixed phases. It is evident that in the monoclinic RE3SbO3
phase the RE sites are quite similar in volume to one another, as
the maximum relative volume difference of the RE sites in the
monoclinic RE3SbO3 structures is only about 15%, yet this
difference is about 10% higher in each of the tetragonal
(REIREII)3SbO3 structures. Thus, it is easy for any RE atom to
occupy the RE sites evenly in the monoclinic RE3SbO3 phase.
However, in the tetragonal (REIREII)3SbO3 phase, the two RE
sites are significantly different in size, and if only one RE atom
is used, it is not possible for simultaneous occupation of both
sites due to their large volume difference. As a result, the
structure cannot be formed. Through the usage of two different
REs with significantly different sizes, a preference of one RE to
occupy one site over another is introduced, and this supports
the novel tetragonal REO framework. However, a full
separation of the REI and REII atoms on the two sites is not
achieved.
Figure 3. Coordination polyhedra of the RE1 site (left) and the RE2
site (right) in the tetragonal (REIREII)3SbO3 phases.
(REIREII)3SbO3, RE1 atoms occupy distorted square antiprisms
with one face consisting of oxygen atoms and another face
consisting of Sb atoms, while RE2 atoms sit in distorted
octahedra made of four oxygen atoms and two antimony atoms.
The site volumes, calculated with the Dido95 program,18 are
given in Table 5. The RE1 site volume is almost 25% larger
than the RE2 site volume, and thus this site should be
preferentially occupied by larger RE atoms. Also, the electro-
negativity difference favors such an RE distribution; the larger
RE atoms are more electropositive and as such they can
support anion-richer environments (RE1 site) better than the
smaller and more electronegative RE atoms.
The tetragonal structure can be preserved with the
REIREII SbO3 stoichiometry, for at least LaDy2SbO3, as the
2
larger RE1 site can easily accommodate more of the small RE
with some additional occupation on the RE2 site. The net
result is no change in the coordination polyhedra volume
between the La1.5Dy1.5SbO3 and LaDy2SbO3 structures
(24.6(3)% vs 24.7(3)%). It is not known why LaDy2SbO3
forms the tetragonal (REIREII)3SbO3 phase while the
LaHo2SbO3 and CeHo2SbO3 samples do not. However,
based on the occupations of the RE sites in the LaDy2SbO3
The presence of the RE1 and RE2 sites with largely different
volumes within the tetragonal (REIREII)3SbO3 structures
appears to be the reason for the existence of these mixed
crystal solution, it is likely that the REIREII SbO3 stoichiometry
2
is near the limit of compositions that can form the tetragonal
(REIREII)3SbO3 phase, and perhaps a small change in site
1029
dx.doi.org/10.1021/ic302292w | Inorg. Chem. 2013, 52, 1025−1031

Inorganic Chemistry
Article
occupations for the LaHo2SbO3 and CeHo2SbO3 samples
results in the destruction of the desired phase.
The REI2REIISbO3 and REI2.5REII0.5SbO3 phases studied were
shown to not preserve the tetragonal (REIREII)3SbO3 structure.
This is because the larger RE1 site is already nearly fully
occupied by the large RE (∼90% in each structure), and as a
result, the smaller RE2 site must be occupied instead. This in
turn destroys the size difference between the two sites and leads
to destabilization of the phase. Finally, the REIREII SbO3 and
2
REI0.5REII2.5SbO3 samples (excluding LaDy2SbO3) were shown
to form the RE2SbO2 phase. This may occur since the high REII
content cannot maintain the large size difference in the RE sites
for the tetragonal structure, and as pointed out in previous
literature,12 the monoclinic RE3SbO3 phases require higher
temperatures to form for the late rare earths. Since the
Ce0.5Ho2.5SbO3 and La0.5Ho2.5SbO3 samples mostly consist of
holmium, they can be viewed as Ho3SbO3, and thus
temperatures of at least 1600 °C will be required to form
monoclinic RE3SbO3. Since neither the monoclinic nor the
tetragonal RE3SbO3 phase can form, the RE2SbO2 phase
becomes the preferred outcome for the REI0.5REII2.5SbO3
samples.
Electrical Resistivity and Electronic Structure. Assum-
ing that La, Dy, and Ho atoms are in +3 oxidation states and
that Sb and O atoms act as isolated anions, the electronic
formulas of (REIREII )3SbO3 can be written as
(REI(3+)REII(3+))3(Sb3−)(O2−)3. It is readily apparent that the
La1.5Dy1.5SbO3 and La1.5Ho1.5SbO3 structures are charge
balanced and may be semiconducting. However, the
Ce1.5Ho1.5SbO3 structure is somewhat ambiguous; CeO2 was
used as a reactant which contains cerium atoms in the +4
oxidation state. While it is true that Ho metal is oxidized in the
reaction with CeO2 to make Ce1.5Ho1.5SbO3, cerium in the
mixed valence state could still be present in the structure, giving
the sample metallic behavior. This ambiguity was later resolved
through electrical conductivity measurements.
The Ce1.5Ho1.5SbO3, La1.5Dy1.5SbO3, and La1.5Ho1.5SbO3
samples show an Arrhenius-type exponential decrease in
electrical resistivity with increasing temperature, which is
indicative of semiconducting behavior and in support of the
charge-balanced formulas. This also suggests that
Ce1.5Ho1.5SbO3 contains only Ce3+ and no Ce4+, as this
would lead to extra carrier electrons and metallic-type behavior
in the sample. The activation energies were calculated based on
Arrhenius-type behavior for a nondegenerate semiconductor
with both electrons and holes as charge carriers. The room-
temperature electrical conductivities and activation energies are
presented in Table 6. The electrical resistivities for all samples
measured in this work are shown in Figure 4.
Figure 4. Electrical resistivities of the Ce1.5Ho1.5SbO3, La1.5Dy1.5SbO3,
and La1.5Ho1.5SbO3 samples.
behavior may be due to (1) a highly conductive impurity, which
dominates the electrical transport properties at low temper-
ature, (2) minor deficiencies, which would change conductivity
from semiconductor-like to metal-like, or (3) a semiconductor-
4−
to-metal transition, which may occur as the result of the Sb2
dimer formation, accompanied by the release of conduction
electrons. To date, we have not resolved the origin of this
unexpected behavior; we leave this work for future inves-
tigation.
Upon inspection, it is also observed that all samples
measured have electrical resistivities of ∼0.1−1 Ω cm or
greater at all temperatures measured, which are roughly 2
orders of magnitude greater than the resistivities of benchmark
thermoelectric materials.26 Additionally, these samples gen-
erally have electrical resistivities roughly equal to the
monoclinic RE3SbO3 and RE8Sb3−δO8 phases. Taking these
facts into consideration, these new phases do not pose any
foreseeable use for thermoelectric applications in their current
states.
Semiconducting properties of the (REIREII)3SbO3 phases
were verified through the electronic structure calculations on
the fully ordered REIREII SbO3 structures. The calculated
2
densities of states, DOS, of LaDy2SbO3, LaHo2SbO3, and
CeHo2SbO3 are very similar. The electronic structure of
LaDy2SbO3 is presented in Figure 5, and those of LaHo2SbO3
and CeHo2SbO3 can be found in the Supporting Information.
An important feature of the DOS is the presence of a small
band gap of about 0.3 eV, which separates the conduction band
from the valence band. In agreement with the
(RE(3+)IRE(3+)II)3(Sb3−)(O2−)3 formula, the conduction band
is primarily composed of the RE d states, and the valence band
is dominated by the Sb p states, as the O p states are much
lower in energy.
The La-containing samples displayed fairly similar electrical
resistivities above 50 K, which is expected considering the
similar chemistry between Dy and Ho. However, below 50 K,
the resistivity of La1.5Dy1.5SbO3 was found to increase. Such
CONCLUSIONS
■
Table 6. Room-Temperature Electrical Resistivities and
Activation Energies of the La1.5Dy1.5SbO3, La1.5Ho1.5SbO3,
and Ce1.5Ho1.5SbO3 Samples
The mixed tetragonal (REIREII)3SbO3 compounds are small-
band-gap semiconductors similar to the previously studied
monoclinic RE3SbO3 and RE8Sb3−δO8 phases. On the basis of
the electronic structure calculations, combining semimetallic
RESb with corresponding RE2O3 opens a band gap between the
valence band dominated by the Sb states and the conduction
band composed of the RE states. The semiconducting
properties of the Ce1.5Ho1.5SbO3 phase support the +3
sample
electrical resistivity (Ω cm)
activation energy (eV)
La1.5Dy1.5SbO3
La1.5Ho1.5SbO3
Ce1.5Ho1.5SbO3
1.05
1.05
0.223
0.11
0.093
0.044
1030
dx.doi.org/10.1021/ic302292w | Inorg. Chem. 2013, 52, 1025−1031

Inorganic Chemistry
Article
(2) LaGrandeur, J.; Crane, D.; Hung, S.; Mazar, B.; Eder, A. Int. Conf.
Thermoelectr. 2006, 25th (Pt. 1), 343−348.
(3) Yang, J.; Caillat, T. MRS Bull. 2006, 31 (3), 224−229.
(4) Watanabe, S. Manufacture of thermoelectric generator.
JP06338636A, 1994.
(5) Rowe, D. M. CRC Handbook of Thermoelectrics: Macro to Nano;
CRC: Boca Raton, 2005.
(6) Slack, G. A. In CRC Handbook of Thermoelectrics; Rowe, D. M.,
Ed.; CRC Press: Boca Raton, FL, 1995.
(7) Caillat, T.; Fleurial, J. P.; Borshchevsky, A. J. Phys. Chem. Solids
1997, 58 (7), 1119−1125.
(8) Mozharivskyj, Y.; Pecharsky, A. O.; Bud’ko, S.; Miller, G. J. Chem.
Mater. 2004, 16 (8), 1580−1589.
(9) Snyder, G. J.; Christensen, M.; Nishibori, E.; Caillat, T.; Iversen,
B. B. Nat. Mater. 2004, 3 (7), 458−463.
(10) Beyer, H.; Nurnus, J.; Bottner, H.; Lambrecht, A.; Roch, T.;
Bauer, G. Appl. Phys. Lett. 2002, 80 (7), 1216−1218.
(11) Harman, T. C.; Spears, D. L.; Manfra, M. J. J. Electron. Mater.
1996, 25 (7), 1121−1127.
(12) Wang, P.; Forbes, S.; Kolodiazhnyi, T.; Kosuda, K.;
Mozharivskyj, Y. J. Am. Chem. Soc. 2010, 132 (25), 8795−8803.
(13) Wang, P. L.; Kolodiazhnyi, T.; Yao, J.; Mozharivskyj, Y. J. Am.
Chem. Soc. 2012, 134 (3), 1426−1429.
Figure 5. Total and partial densities of state for LaDy2SbO3.
(14) STOE X-Shape; STOE & Cie GmbH: Darmstadt, Germany,
2004.
oxidation state for Ce, as a mixed valence state would yield a
metallic behavior. The REI1.5REII1.5SbO3 samples display room-
temperature electrical resistivities in the 0.1−1 Ω cm range,
similar to the RE3SbO3 and RE8Sb3−δO8 phases, which is too
large for thermoelectric applications.
(15) Sheldrick, G. M. SHELXL; University of Gottingen: Germany,
1997.
(16) Hamilton, W. C. Acta Crystallogr. 1965, 18 (3), 502−10.
(17) Hunter, B. A.; Howard, C. J. Rietica Australian Nuclear Science
and Technology Organization; Lucas Heights Research Laboratories:
Menai, Australia, 2000.
(18) Koch, E.; Fischer, W. Z. Kristallogr. 1996, 211, 251−253.
(19) Andersen, O. K.; Pawlowska, Z.; Jepsen, O. Phys. Rev. B:
Condens. Matter Mater. Phys. 1986, 34, 5253−69.
(20) Jepsen, O.; Burkhardt, A.; Andersen, O. K. The TB-LMTO-ASA
The tetragonal (REIREII)3SbO3 phases are obtained at
temperatures above 1550 °C and only when two RE atoms
with highly different atomic radii are used. While the REI and
REII atoms are distributed on both RE sites, there is a
pronounced site preference; larger atoms tend to occupy the
larger site (4f) and smaller atoms occupy the smaller site (8i).
Compared to the analogous monoclinic RE3SbO3 structures,
the RE site volumes differ significantly in the tetragonal
(REIREII)3SbO3 structures. Such difference supports incorpo-
ration of two different RE atoms into the structure and,
ultimately, its stability.
Program, version 4.7; Max-Planck-Institut fur Festkorperforschung:
̈
̈
Stuttgart, Germany, 1999.
(21) Andersen, O. K.; Jepsen, O. Phys. Rev. Lett. 1984, 53, 2571−
2574.
(22) Andersen, O. K.; Jepsen, O.; Glotzel, D. In Highlights of
Condensed Matter Theory; Bassani, F., Fumi, F., Tosi, M. P., Eds.;
Elsevier: Amsterdam, 1985.
(23) Jepsen, O.; Andersen, O. K. Phys. Rev. B.: Condens. Matter 1995,
97, 35−47.
ASSOCIATED CONTENT
* Supporting Information
■
S
(24) Shannon, R. D. Acta Crystallogr., Sect. A 1976, A32 (5), 751−67.
(25) Papoian, G. A.; Hoffmann, R. Angew. Chem., Int. Ed. 2000, 39
(14), 2409−2448.
Magnetic measurements of the Ce1.5Ho1.5SbO3 sample are
presented, as well as the calculated densities of states for the
REI1.5REII1.5SbO3 phases adjusted to REIREII SbO3 stoichiome-
(26) Sootsman, J. R.; Chung, D. Y.; Kanatzidis, M. G. Angew. Chem.,
Int. Ed. 2009, 48 (46), 8616−8639.
2
tries via full occupation of the RE sites based on the site
preferences derived from the X-ray single-crystal solutions. The
crystallographic data for the Dy3SbO3 phase as well as the other
phases are presented. This material is available free of charge via
the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: mozhar@mcmaster.ca.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
This work was supported by a Discovery Grant from the
Natural Sciences and Engineering Research Council of Canada.
■
REFERENCES
■
(1) Bell, L. E. Science (Washington, DC, U. S.) 2008, 321 (5895),
1457−1461.
1031
dx.doi.org/10.1021/ic302292w | Inorg. Chem. 2013, 52, 1025−1031