Cationic clathrate I Si46-xPxTey (6.6(1) ≤ y≤ 7.5(1), x ≤ 2y): Crystal structure,homogeneity range, and physical properties

Article abstract of DOI:10.1021/ic8023887

A new cationic clathrate I Si_46-xP_xTe_y (6.6(1) ≤ y ≤ 7.5(1), x ≤ 2y at 1375 K) was synthesized from the elements and characterized by X-ray powder diffraction, thermal analysis, scanning electron microscopy, wavelength dis

Full text of DOI:10.1021/ic8023887

Inorg. Chem. 2009, 48, 3720-3730
Cationic Clathrate I Si P Te (6.6(1) e y e 7.5(1), x e 2y): Crystal
46-x x
y
Structure, Homogeneity Range, and Physical Properties
†,‡,§
J. V. Zaikina,
and A. V. Shevelkov*
‡,§
‡
‡
‡
‡
K. A. Kovnir, U. Burkhardt, W. Schnelle, F. Haarmann, U. Schwarz, Yu. Grin,
‡
,†
Chemistry Department, LomonosoV Moscow State UniVersity, 119991 Moscow, Russia, and
Max-Planck-Institut f u¨ r Chemische Physik fester Stoffe, 01187 Dresden, Germany
Received December 15, 2008
x y
A new cationic clathrate I Si46-xP Te (6.6(1) e y e 7.5(1), x e 2y at 1375 K) was synthesized from the elements
and characterized by X-ray powder diffraction, thermal analysis, scanning electron microscopy, wavelength dispersive
31
X-ray spectroscopy (WDXS), neutron powder diffraction, and P NMR spectroscopy. The thermal behaviors of the
magnetic susceptibility and resistivity were investigated as well. Si46-x Te reveals a wide homogeneity range due
to the presence of vacancies in the tellurium guest positions inside the smaller cage of the clathrate I structure.
The vacancy ordering in the structure of Si46-x Te
causes the change of space group from Pm 3¯ n (ideal clathrate
P
x
y
P
x
y
I) to Pm 3¯ accompanied by the redistribution of P and Si atoms over different framework positions. Neutron powder
diffraction confirmed that P atoms preferably form a cage around the vacancy-containing tellurium guest position.
31
Additionally, P NMR spin-spin relaxation experiments revealed the presence of sites with different coordination
of phosphorus atoms. Precise determination of the composition of Si46-x Te by WDXS showed slight but noticeable
P
x
y
deviation (x e 2y) of phosphorus content from the Zintl counting scheme (x ) 2y). The compound is diamagnetic
while resistivity measurements show activated behavior or that of heavily doped semiconductors. Thermal analysis
revealed high stability of the investigated clathrate: Si46-x
P Te
x y
melts incongruently at ∼1460 K in vacuum and is
stable in air against oxidation up to 1295 K.
Introduction
cations of alkali, alkaline-earth metals, or europium as guest
atoms in the large polyhedral cages. This type of clathrates
is known as anionic clathrates or simply clathrates since the
negatively charged host framework is balanced by guest
cations. The host framework of cationic clathrates or
clathrates with inversed host-guest polarity is built of Si,
Ge, and Sn and always with an addition of a Group 15
element (pnicogen), tellurium, or even iodine, with halogen
or tellurium anions serving as guests. Starting with the
Clathrates belong to the group of inclusion compounds,
in which noncovalent interactions between host and guest
substructures play a key role in the arrangement and
1
stabilization of the crystal structure. Sodium silicide Na
8
Si46
,
2
discovered over 40 years ago, was the first example of an
intermetallic compound that can adopt the crystal structure
of the type I gas hydrate, also known as the clathrate I
structure type. Subsequently, more than 100 type I clathrates
of the group 14 elements were synthesized and scrutinized.
Their crystal structure is best described as a three-
dimensional host framework based on Si, Ge, or Sn trapping
preparation of germanium clathrates Ge38
Z X (Z ) P, As,
8 8
3
4
or Sb; X ) Cl, Br, or I) by von Schnering and Menke, a
5
6
number of tin and several new germanium cationic type I
(
3) (a) Kovnir, K. A.; Shevelkov, A. V. Russ. Chem. ReV. 2004, 73, 923–
938, and references therein. (b) Jung, W.; L o¨ rincz, J.; Ramlau, R.;
Borrmann, H.; Prots, Y.; Haarmann, F.; Schnelle, W.; Burkhardt, U.;
Baitinger, M.; Grin, Y. Angew. Chem., Int. Ed. 2007, 46, 6725–6728.
(c) Condron, C. L.; Kauzlarich, S. M.; Nolas, G. S. Inorg. Chem. 2007,
46, 2556–2562. (d) B o¨ hme, B.; Guloy, A.; Tang, Z.; Schnelle, W.;
Burkhardt, U.; Baitinger, M.; Grin, Y. J. Am. Chem. Soc. 2007, 129,
5348–5349. (e) Melnychenko-Koblyuk, N.; Grytsiv, A.; Rogl, P.;
Rotter, M.; Lackner, R.; Bauer, E.; Fornasari, L.; Marabelli, F.; Giester,
G. Phys. ReV. B 2007, 76, 195124. (f) Kaltzoglou, A.; Ponou, S.;
F a¨ ssler, T. F. Eur. J. Inorg. Chem. 2008, 538–542.
*
To whom correspondence should be addressed. E-mail: shev@
inorg.chem.msu.ru. Fax: (+7-495) 939 47 88.
‡
Max-Planck-Institut f u¨ r Chemische Physik fester Stoffe.
Lomonosov Moscow State University.
Present address: Department of Chemistry and Biochemistry, Florida
†
§
State University, Tallahassee, FL, 32306.
1) Jeffrey G. A. In Inclusion Compounds; Atwood, J. L., Davies, J. E. D.,
MacNicol, D. D., Eds.; Academic Press Inc.: London, 1984.
(
(
2) Kasper, J. S.; Hagenmuller, P.; Pouchard, M.; Cros, C. Science 1965,
1
50, 1713–1714.
3720 Inorganic Chemistry, Vol. 48, No. 8, 2009
10.1021/ic8023887 CCC: $40.75  2009 American Chemical Society
Published on Web 03/12/2009

x y
Cationic Clathrate I Si46-xP Te
12b
clathrates were obtained. However, only three silicon-based
geneity range, was performed. We have applied a number
of physical and chemical characterization methods to over-
come most obstacles and to shed some light on the relation-
ship between the composition, crystal structure, and physical
7
6 6.5
cationic clathrates have been known up to now: Si40P I ,
8
9
1.5 8
Si38Te16, and Si44.5I I . The latter two compounds were
synthesized under high pressure only.
The electron balance between hosts and guests in cationic
and anionic clathrates is achieved by substitution and/or
vacancy formation in either host or guest positions, and in
many cases the clathrates composition fulfills the 8-n rule,
properties for the new cationic clathrate Si46-x
e y e 7.5(1), x e 2y).
x y
P Te (6.6(1)
Experimental Section
10
thus, the Zintl concept. The interest into the clathrate topic
has been renewed owing to the concept phonon glass –
Sample Preparation. Preparation and storage of the samples
and the starting materials were carried out in an argon-filled
11
electron crystal (PGEC) introduced by Slack. This concept
claimed that an advanced material for thermoelectric cooling
or power generation would be found among compounds in
which guest atoms trapped inside oversized cages are able
to rattle. This would lead to a suppression of the thermal
conductivity because of effective phonon scattering without
impeding the transport of charge carriers. Such a combination
of properties would ensure a high figure of merit, ZT )
glovebox (content of O
2
<1.2 ppm, H O <1 ppm). Silicon
2
(
99.9999%, ABCR), red phosphorus (99.995%, Chempur), and
tellurium (99.995%, Chempur) were used. Pieces of silicon and
tellurium were separately reground in a tungsten carbide mortar
and then were sifted (mesh size 0.63 or 0.8 µm). Phosphorus was
mechanically pulverized in an agate mortar. Powders of the elements
(0.5-2 g total weight) were mixed in respective stoichiometric ratios
and were sealed in fused silica ampules (inner diameter 9.5 mm,
outer diameter 13 mm, length approximately 60 mm) under vacuum.
Samples were synthesized by a two-step annealing. Ampules with
the stoichiometric mixtures of the starting materials were placed
in furnaces, heated to 1375 K in 24 h, and annealed for 100 h.
Afterward, the samples were cooled, ground, and annealed again
in fused silica ampules sealed under vacuum at the same temperature
for 144 h. Finally, the ampules were cooled to room temperature
in the furnace. An additional series of syntheses was performed
with the same conditions but with a lower temperature of 925 K.
For neutron powder diffraction studies, the samples (∼5 g) were
prepared by either two-step annealing at 1375 K or two-step
annealing at 1375/1075 K, but in all cases large fused silica ampules
2
S Tσ/κ (T is temperature, σ is electrical conductivity, S is
the Seebeck coefficient, and κ is thermal conductivity) of a
thermoelectric material.
Our systematic study of cationic clathrates
5a-e,6b,7
had led
to the investigation of the ternary system Si-P-Te, which
turned out to be a challenge because of various preparative
1
2a
and characterization problems.
munication about thermoelectric properties of a clathrate I
with an unspecified composition Si46-x Te was published,
Recently, a short com-
P
x
8
but no detailed investigation of its composition and crystal
structure, in particular phosphorus distribution and homo-
(
inner diameter 18.5 mm, outer diameter 19.5 mm, length ≈110
mm) were used. Ampules were placed into a furnace, heated to
1375 K in 24 h, and annealed for 100 h. The samples were reground
and annealed in fused silica ampules sealed under vacuum at the
same temperature of 1375 K for 144 h (sample 1) or at 1075 K for
(
4) (a) von Schnering, H. G.; Menke, H. Angew. Chem., Int. Ed. 1972,
1
1, 43–44. (b) Menke, H.; von Schnering, H. G. Z. Anorg. Allg. Chem.
1
973, 395, 223–238.
(
5) (a) Shatruk, M. M.; Kovnir, K. A.; Shevelkov, A. V.; Presnyakov,
I. A.; Popovkin, B. A. Inorg. Chem. 1999, 38, 3455–3457. (b) Shatruk,
M. M.; Kovnir, K. A.; Lindsj o¨ , M.; Presniakov, I. A.; Kloo, L. A.;
Shevelkov, A. V. J. Solid State Chem. 2001, 161, 233–242. (c) Kovnir,
K. A.; Shatruk, M. M.; Reshetova, L. N.; Presniakov, I. A.; Dikarev,
E. V.; Baitinger, M.; Haarmann, F.; Schnelle, W.; Baenitz, M.; Grin,
Yu.; Shevelkov, A. V. Solid State Sci. 2005, 7, 957–968. (d) Zaikina,
J. V.; Kovnir, K. A.; Sobolev, A. V.; Presniakov, I. A.; Prots, Y.;
Baitinger, M.; Schnelle, W.; Olenev, A. V.; Lebedev, O. I.; Van
Tendeloo, G.; Grin, Y.; Shevelkov, A. V. Chem. Eur. J. 2007, 13,
1
20 h (sample 2).
X-ray Powder Diffraction. All samples were analyzed by X-ray
powder diffraction in a transmission alignment using a Huber G670
image plate camera, Cu KR radiation, λ ) 1.540598 Å. The unit
cell parameters were calculated from least-squares fits using LaB
cubic, a ) 4.15692 Å) as an internal standard utilizing the program
1
6
(
13
package WinCSD. Lattice parameters were obtained from the
refinement of a set of 54 reflections common for all samples.
Rietveld refinement was performed using the JANA2000 program
5
090–5099. (e) Kovnir, K. A.; Sobolev, A. V.; Presniakov, I. A.;
Lebedev, O. I.; Van Tendeloo, G.; Schnelle, W.; Grin, Y.; Shevelkov,
A. V. Inorg. Chem. 2005, 44, 8786–8793. (f) Kishimoto, K.; Arimura,
S.; Koyanagi, T. Appl. Phys. Lett. 2006, 88 (1-3), 222115.
6) (a) Nesper, R.; Curda, J.; von Schnering, H. G. Angew. Chem., Int.
Ed. Engl. 1986, 25, 350–352. (b) Kovnir, K. A.; Abramchuk, N. S.;
Zaikina, J. V.; Baitinger, M.; Burkhardt, U.; Schnelle, W.; Olenev,
A. V.; Lebedev, O. I.; Van Tendeloo, G.; Dikarev, E. V.; Shevelkov,
A. V. Z. Kristallogr.-New Cryst. Struct. 2006, 221, 527–532. (c)
Kishimoto, K.; Akai, K.; Muraoka, N.; Koyanagi, T.; Matsuura, M.
Appl. Phys. Lett. 2006, 89 (1-3), 172106.
14
package.
(
Scanning Electron Microscopy (SEM) and Wavelength
Dispersive X-ray Spectroscopy (WDXS). Selected samples were
analyzed by means of SEM. In order to investigate the phase
distribution, metallographic studies were performed first. The
samples were analyzed by light and electron optical microscopy.
The preparation of the samples, including embedding in conductive
resin and diamond polishing using dried hexane as a lubricant, was
(
(
7) Kovnir, K. A.; Uglov, A. N.; Zaikina, J. V.; Shevelkov, A. V.
MendeleeV Commun. 2004, 135–136.
8) Jaussaud, N.; Toulemonde, P.; Pouchard, M.; San Miguel, A.;
Gravereau, P.; Pechev, S.; Goglio, G.; Cros, C. Solid State Sci. 2004,
1
5
carried out in an inert atmosphere of an argon-filled glovebox.
6
, 401–411.
9) Reny, E.; Yamanaka, S.; Cros, C.; Pouchard, M. Chem. Commun. 2000,
505–2506.
The surface of the samples was further examined in partly polarized
light (ZEISS Axioplan2 optical microscope) using differential
interference contrast.
(
2
(
(
(
10) Chemistry, Structure, and Bonding of Zintl Phases and Ions; Kau-
zlarich, M. S., Ed.; VCH Publishers: New York, 1996.
11) Slack, G. A. In CRC Handbook of Thermoelectrics; Rowe, D. M.,
Ed.; CRC Press LLC: Boca Raton, FL, 1995.
(13) Akselrud, L. G.; Zavalij, P. Y.; Grin, Yu.; Pecharsky, V. K.;
Baumgartner, B.; W o¨ lfel, E. Mater. Sci. Forum 1993, 133-136, 335–
351.
(14) Petricek, V.; Dusek, M.; Palatinus, L. Jana2000. The Crystallographic
Computing System. Institute of Physics: Praha, Czech Republic, 2000.
(15) Schnelle, W.; Burkhardt, U.; Ramlau, R.; Niewa, R.; Sparn, G.
Scientific Report MPI CPfS 2003, 38–43.
12) (a) Zaikina, J. V.; Kovnir, K. A.; Demtschyna, R.; Burkhardt, U.; Prots,
Y.; Borrmann, H.; Schwarz, U.; Shevelkov, A. V. Book of Abstracts,
The 10th European Conference on Solid State Chemistry 2005, P096,
1
62. (b) Kishimoto, K.; Koyanagi, T.; Agai, K.; Matsuura, M. Jpn.
J. Appl. Phys. Part 2 2007, 46, L746-L748.
Inorganic Chemistry, Vol. 48, No. 8, 2009 3721

Zaikina et al.
Microstructures were homogeneous or showed the majority phase
with irregular-shaped grains up to 200 µm. The minority phases in
the case of non-single-phase samples were observed in material
contrast images collected by the BSE (back-scattered electron)
detector of an electron microprobe Cameca SX100 with tungsten
cathode.
a shifted 14-order Chebyshev polynomial function. A pseudo-Voigt
function was applied to generate the profile shape. The following
neutron scattering lengths were used: bTe ) 5.80 fm, b ) 5.13
P
2
0
fm, bSi ) 4.15 fm.
Physical Property Measurements. The magnetic susceptibility
ꢀ ) M/H was measured in external magnetic fields µ H of 0.01,
0
The composition of the ternary phase was determined by WDXS
carried out with the electron microprobe. The intensities of the X-ray
lines P K , Si K , and Te L were determined on two wavelength
R R R
dispersive spectrometers equipped with focusing monochromator
crystals made of TAP (thallium acid phthalate) and PET (pen-
taerythritol). An electron beam current of 10 nA was adjusted for
0.1, 1, 3.5, and 7.0 T over a temperature range of 1.8-400 K using
a MPMS XL-7 SQUID magnetometer (Quantum Design). Rather
small polycrystalline samples (m ) 48-87 mg) were placed in
polyethylene tubes and were held in place by short pieces (m ≈15
mg) of the same material. Corrections for the diamagnetism of the
individual sample holders were applied.
suitable count rates at the acceleration voltages of 25 kV (Te L
and 15 kV (P K , Si K ). The signals were compared to the
intensities of the reference materials SiP
corrections and averaging over 10 points, an overall value of Σ )
00.2(2) wt % was found. The homogeneity and chemical composi-
R
)
Electrical resistivity measurements were performed on the same
samples with a standard dc (direct current) four-point method
between 4 and 320 K. Contacts were made with a silver-filled epoxy
paste. Because of the small size of the pieces and resulting problems
in the determination of the contact geometry, the inaccuracy of the
absolute resistivity is estimated to be (30%.
R
R
16
2
and GeTe. After matrix
1
tion of the reference materials were verified by X-ray powder
diffraction and chemical analysis, respectively.
All sample handling for physical properties measurements was
Chemical Analysis. Samples for the neutron diffraction experi-
ments and standards for WDXS were characterized by chemical
analysis using an inductively coupled plasma optical emission
spectrometry method as implemented in a Varian Vista RL
spectrometer. For sample dissolution, a microwave-assisted acid
done under inert atmosphere.
NMR Spectroscopy. 31P magic-angle spinning (MAS) and static
wide line (SWL) NMR experiments were done using a Bruker
AVANCE spectrometer with B ) 11.74 T equipped with a standard
0
Bruker MAS probe. The sample was mounted in a 4 mm zirconium
dioxide rotor for spinning frequencies below and above 15 kHz. A
Hahn echo sequence was used in all experiments with an interpulse
delay of 100 µs in SWL and was synchronized with rotor frequency
in MAS experiments. The durations for the π-pulses were 3 and
200 µs in nonselective and selective excitation experiments,
respectively. A repetition time of 2 s was sufficient to ensure full
recovery of the longitudinal magnetization. The signal shift was
digestion (mixture of 3 mL of aqua regia (HCl:HNO
3
, 3:1) and
0
.05 µL of HF) was applied. All values are the average of at least
three replicates. The precise determination of composition of the
samples in the Si-P-Te system was hampered by the simultaneous
presence of volatile (phosphorus and tellurium) and sparingly
soluble (silicon) components, which resulted in the inaccurate
overall composition deviating from 100%.
Thermal Behavior. Differential scanning calorimetry (DSC) and
thermal gravimetric investigations were performed with the Netzsch
DSC 409C equipment in a quartz glass crucible sealed under
vacuum or in an open quartz crucible (static air), correspondingly,
in the temperature range of 300-1550 K with a heating rate of 5
K/min. After cooling with the same rate, the samples were examined
by X-ray powder diffraction.
3 4
referred to H PO .
Results and Discussion
Crystal Structure Description and Comparison with
Other Clathrates. The details of the X-ray single-crystal
diffraction experiments and a brief crystal structure descrip-
Neutron Powder Diffraction. Neutron powder diffraction
patterns of the samples were recorded using the high-resolution
tion of the clathrate I Si46-x
x y
P Te (y ) 7.35, 6.98, 6.88; x e
2
y) have been already published in our recent short com-
1
7
powder D2B diffractometer at the Institute Max von Laue-Paul
Langevin (Grenoble, France). A polycrystalline sample (∼5 g) was
placed in a cylindrical vanadium can. The [331] reflection from
the Ge monochromator crystal was used to produce a monochro-
matic neutron beam of wavelength 2.3985 Å. Data were collected
at room temperature in a 2θ angular range from 10° to 155° with
a step size of 0.05°. The total counting time was approximately
21
munication. However, it is worthwhile to outline some
basic details concerning the atomic arrangement of the title
compound. The ideal crystal structure of the clathrate I has
8
the composition G E46, where the guest atoms G are enclosed
in large cages of the four-connected host framework consist-
ing of 46 E atoms (Figure 1). The framework atoms occupy
three crystallographic positions, 6c, 16i, and 24k, of the space
group Pm 3j n (no. 223), while the guest atoms reside in the
1
8
4
h. The raw data were merged using the LAMP program.
Diffraction patterns were analyzed by the Rietveld refinement
1
9
method using the GSAS software package. The only peak
belonging to the vanadium can was subtracted. For each sample
the profile parameters, background parameters, zero correction, and
lattice constant were refined first. The background was fitted using
2
a and 6d positions situated in the center of the 20- and
2
4-vertex polyhedral cages, respectively, alternating in the
unit cell in the 2:6 ratio.
The variations of the ideal G E46 composition due to the
8
formation of vacancies in either guest positions or host
framework in order to achieve the Zintl composition or
because of the host-guest matching are described in the
(
16) (a) Kanatzidis, M. G.; P o¨ ttgen, R.; Jeitschko, W. Angew. Chem., Int.
Ed 2005, 44, 6996–7023. (b) Compound GeTe with the accurately
defined composition was donated by Dr. V. I. Shtanov (Chemistry
Department, Lomonosov Moscow State University).
17) High-resolution two-axis diffractometer D2B. 2007. Institut Laue-
Langevin, Grenoble, France; http://www.ill.eu/d2b/.
3
,5a-d,7
jn or
Deviations from the space group Pm3
literature.
(
(
(
even superstructure formation resulting in the loss of the
cubic symmetry were also observed for several clathrates
18) The Large Array Manipulation Program (LAMP); http://www.ill.fr/
data_treat/lamp/lamp.html.
19) (a) Larson, A. C.; Von Dreele, R. B. General Structure Analysis System
(
8
2
GSAS). Los Alamos National Laboratory Report (LAUR) 2000,
(20) Sears, V. Neutron News 1992, 3, 26–37.
(21) Zaikina, J. V.; Kovnir, K. A.; Schwarz, U.; Borrmann, H.; Shevelkov,
A. V. Z. Kristallogr. – New Cryst. Struct. 2007, 222, 177–179.
6-748, 1–224. (b) Toby, B. H. J. Appl. Crystallogr. 2001, 34, 210–
13.
3722 Inorganic Chemistry, Vol. 48, No. 8, 2009

x y
Cationic Clathrate I Si46-xP Te
1
b by tellurium is associated with a difference in their
coordination (Figure 1). Each atom, Te(1) or Te(2), is
surrounded by 20 framework atoms located in the 8-fold and
1
2-fold positions. The distances between Te and the atoms
in the 8-fold position are relatively short and lie in the range
of 3.16-3.19 Å. In the case of a fully occupied Te(1), the
8-fold position is either solely or mainly occupied by silicon,
while it is merely occupied by phosphorus atoms in the case
of the vacancy-containing Te(2) position. Thus, the symmetry
reduction from Pm 3j n to Pm 3j leads not only to a segregation
of tellurium vacancies in one of the positions in the center
of the 20-vertex polyhedron, but also to the preferred
occupation of the polyhedral vertexes by either silicon or
phosphorus atoms. Within the framework, the interatomic
distances fall in the range of 2.30-2.34 Å, typical for the
single covalent bonds Si-Si and Si-P. The distances
between the positions that are fully occupied by silicon are
slightly larger than those occurring between the positions
with a mixed Si/P occupation, while the distances of ∼2.20
Å corresponding to the single covalent P-P bond are not
observed in the structure. Such a separation of phosphorus
atoms from each other could be an additional reason for the
lowering of the symmetry and for the change of the space
group from Pm 3j n to Pm 3j . This crystal structure model was
x y
Figure 1. Crystal structure of clathrate I Si46-xP Te . 20-vertex polyhedra
for Te(1) and Te(2) and 24-vertex polyhedron for Te(3) are emphasized.
Scheme 1. Wyckoff Site Transformation from the Space Group Pm 3j n
to the Pm 3j Space Group for Clathrate I Structure
proposed on the basis of X-ray single-crystal diffraction data
with sin(θ)/λ > 1, data/parameters > 50, that allowed to
reliably distinguish between the positions occupied exclu-
sively by phosphorus or silicon atoms.
Several examples of clathrates I, all based on silicon, which
contain vacancies in the guest positions are known. They
2
3
3d,23
are the anionic clathrates Rb6.15(2)
0
1.85Si46
,
K
8-xSi46,
24
25
superconducting Ba8-x
cationic clathrate Si40
all these clathrates does not lead to a change/lowering of
the symmetry. Thus, to the best of our knowledge, Si46-x Te
y
0
x
Si46 and Na
x 6
Ba 0
2-xSi46, and the
7
6 6.5
P I 01.5. The vacancy formation in
P
x
is the first example of a clathrate I compound that exhibits
guest vacancy ordering resulting in the change of the space
group from Pm 3j n to Pm 3j .
due to either the ordering of vacancies or the separation/
3a,4,5b,d,8,22
orderingofdifferentatomsatdistinctframeworkpositions.
The solid solution Si46-x Te is a new cationic clathrate
P
x
y
Two other examples of cationic clathrates having tellurium
I. It crystallizes in space group Pm 3j (no. 200), which is a
subgroup of the space group Pm 3j n corresponding to the ideal
8
6c
as guest atoms, Si38Te16 and Ge30
recently. The crystal structure of Si38Te16, or more precisely
Si38Te ]Te , determined by means of X-ray single-crystal
P
16Te
8
,
were reported
clathrate I structure type. The host framework is composed
of silicon and phosphorus atoms, while tellurium atoms
occupy the guest positions. Upon the transformation from
the space group Pm 3j n to the space group Pm 3j , both guest
[
8
8
diffraction corresponds to the Si-Te host framework trapping
8
Te atoms inside the cages. The authors suggested a change
of the space group from Pm 3j n to P 4j 3n with the concomitant
splitting of the 16-fold position into two 8-fold sites. One
of those positions is exclusively occupied by silicon, while
another one is filled by tellurium. Tellurium guest atoms were
set as fully occupied positions in the center of 20- and 24-
vertex cages. However, the reported value of the atomic
displacement parameters (ADP) for the Te atom in the
smaller 20-vertex cage was almost the same as that refined
for the Te atom in the large 24-vertex cage. This most likely
and framework positions split into crystallographically dif-
ferent sites (Scheme 1). The guest position 2a turns into two
one-fold positions, 1a and 1b. The splitting of the framework
positions 16i and 24k of the space group Pm 3j n leads to the
1 2
positions 8i and 8i and 12j and 12k of the space group
Pm 3j , correspondingly. The obvious reason for the change
of the space group is the vacancy segregation in one of the
tellurium guest positions in the center of the 20-vertex cage,
1
b, while the occupancy of the 1a position is always larger.
Such a difference in the occupation of the positions 1a and
(
23) Ramachandran, G. K.; McMillan, P. F.; Dong, J.; Sankey, O. F. J.
Solid State Chem. 2000, 154, 626–634.
(
22) (a) Carrillo-Cabrera, W.; Budnyk, S.; Prots, Y.; Grin, Y. Z. Anorg.
Allg. Chem. 2004, 630, 2267–2276. (b) Dubois, F.; F a¨ ssler, T. F. J. Am.
Chem. Soc. 2005, 127, 3264–3265. (c) D u¨ nner, J.; Mewis, A. Z. Anorg.
Allg. Chem. 1995, 621, 191–196.
(24) Fukuoka, H.; Kiyoto, J.; Yamanaka, S. Inorg. Chem. 2003, 42, 2933–
2937.
(25) Kawaji, H.; Iwai, K.; Yamanaka, S.; Ishikawa, M. Solid State Commun.
1996, 100, 393–395.
Inorganic Chemistry, Vol. 48, No. 8, 2009 3723

Zaikina et al.
indicates a partial occupation of the tellurium position in the
center of the smaller 20-vertex cage, since usually clathrate
I compounds feature almost 2 times larger and anisotropic
ADP for the guest atoms in the larger 24-vertex cage
compared to the ADP for the guest atom in the smaller cage.
Further, the relatively high ADP for silicon in one of the
8
-fold positions compared to the other silicon framework
atoms may point to the vacancies in this position and, thus,
may lead to a deviation from the composition [Si38Te ]Te
The crystal structure and composition of the second
tellurium-containing clathrate Ge30 were determined
by Rietveld refinement of X-ray powder diffraction data.
However, the composition was not confirmed by any
additional method. The reported crystal structure of
8
Ge30P16Te corresponds to a Ge-P host framework, where
8
8
.
P16Te
8
6
c
Figure 2. Dependence of the cubic unit cell parameter, a, upon Te content,
x y
y, in Si46-xP Te determined by Rietveld refinement (black rhombs) and by
WDXS (red rhombs). The linear fit is represented by dashed line. The esd’s
estimated standard deviation) in determination of the unit cell parameters
(
by powder XRD do not exceed 0.0002 Å, which is smaller than the symbols
used. The esd’s for the Te content determined by Rietveld refinement are
considered to be 0.1 per clathrate formula due to the underestimation of
this value given by Rietveld method.
phosphorus atoms preferably occupy the 16i position of space
group Pm 3j n, while tellurium atoms serve as guests. On the
basis of the X-ray powder data, the authors suggested a
partial occupation of 85 and 95% for the tellurium atoms in
the center of 20- and 24-vertex cages, respectively. The
confirmed that the actual tellurium content of the single-
phase samples was close to the starting composition (Figure
2). The precise determination of the cubic unit cell parameter
refined composition corresponds to [Ge30.1
vacancies in the tellurium guest positions. Homogeneity
ranges for the tellurium-containing clathrates [Si38Te ]Te
and [Ge30 were not reported.
Synthesis, X-ray Powder Diffraction Data, and
Homogeneity Range. As discussed above, Si46-x Te shows
P15.9]Te7.4 with
x y
of the clathrate I Si46-xP Te revealed that a varied from
9.9789(2) to 9.9724(1) Å, linearly decreasing with increasing
tellurium content from y ) 6.6(1) to y ) 7.5(1) (Figure 2).
The samples with the starting composition ynom < 6.5 and
8
8
P16]Te
8
y
nom > 7.5 were not single-phase and exhibited no unit cell
P
x
y
parameter changes within three esd compared to the corre-
sponding single-phase border compositions. The sample with
the nominal composition ynom > 7.5 contained impurities of
Te vacancies in the smaller cage. A different number of
vacancies and a corresponding different number of tellurium
guest atoms cause a broad homogeneity range. We investi-
gated different compositions of the clathrate Si46-xP Te by
x y
means of X-ray powder diffraction.
The starting composition was calculated according to the
Zintl concept, assuming that each isolated tellurium guest
atom carries a formal charge of -2, while the four-bonded
phosphorus atoms of the framework have a formal charge
of +1. Further, taking into account that the vacancies are
formed in the guest tellurium positions but not in the host
Si
tellurium content (ynom ) 6.3) constituted a mixture of the
clathrate I phase with an impurity of clathrate III Si172-x Te
y
2 3 2
Te and phosphide SiP , while the sample with the low
P
x
2
6
(
x ) 2y, y > 20).
These findings are compatible with a homogeneity range
for the solid solution Si46-x Te at 1375 K of 6.5(1) e y e
.5(1). The linear decrease of the unit cell with increasing
tellurium content, y, apparently corresponds to an increase
of the phosphorus content in Si46-x Te , x ) 2y. This is in
P
x
y
7
P
x
y
x y
framework, the composition Si46-xP Te (x ) 2y) could be
deduced with a phosphorus content being twice as large as
the tellurium amount. Samples with the nominal composi-
agreement with the smaller covalent radius of P compared
27
to Si (rcov(Si) ) 1.17 Å, rcov(P) ) 1.10 Å), and according
to the Zintl concept an increase of the tellurium content
necessarily leads to an increase of the phosphorus content.
Thus, the increasing amount of Te results in a compression
of the Si-P host framework. It should be noted that a
combination of the light elements Si and P constituting the
host framework leads to the smallest value of the unit cell
parameter among all cationic clathrates I and is comparable
tions Si46-x
)
P
x
Te
2y) were synthesized by two-step annealing at 1375 K.
The synthesis temperature was chosen just below the
decomposition temperature of Si46-x Te
y
(y ) 6.3, 6.5, 6.75, 7.00, 7.35, 7.5, 7.75, x
P
x
y
(∼1460 K, see
below) and the synthesis temperature of the binary Si38Te16
8
clathrate. However, the formation of the target clathrate
phase also occurred at lower temperatures (925 K) similar
7 7
with that of the anionic clathrate I K B Si39 (a ) 9.952(1)
7
to the synthesis temperature of Si40
P
6
I
6.5, indicating that the
3b
Å).
formation of the clathrate is also possible via the gas phase
without the melting of silicon.
The X-ray powder diffraction data show that the Si46-xP Te
x y
samples with the nominal composition 6.5 e ynom e 7.5
correspond to the pure clathrate I phase. The actual tellurium
content, y, was determined by WDXS analysis, as well as
The change of the space group from Pm 3j n (no. 223),
corresponding to the ideal clathrate I structure type, to Pm3j
(no. 200) resulted in the appearance of weak but noticeable
reflections with l * 2n in the [hhl] zone (Figure 3). A
complete elimination of these reflections does not occur even
by Rietveld refinement, assuming the composition Si46Te
taking into account the similarity of the X-ray scattering
factors for Si and P. Such a combination of the methods
y
(26) Zaikina, J. V.; Kovnir, K. A.; Haarmann, F.; Schnelle, W.; Burkhardt,
U.; Borrmann, H.; Schwarz, U.; Grin, Yu.; Shevelkov, A. V. Chem.
Eur. J. 2008, 14, 5414–5422.
(27) Emsley, J. The Elements; Clarendon Press: Oxford, U.K., 1991.
3724 Inorganic Chemistry, Vol. 48, No. 8, 2009

x y
Cationic Clathrate I Si46-xP Te
Figure 3. X-ray powder diffraction pattern for Si46-x
x y
P Te with composition y ) 7.25(5), x ) 14.10(5) (WDXS). The inset shows the representative peaks
with l * 2n in the [hhl] zone marked with asterisk.
for clathrate I phase present in the samples with ynom close
to 6 or 8. Calculation of the theoretical X-ray powder
diffraction patterns revealed that reflections with l * 2n
reflected mainly the difference in the occupation of the 1a
and 1b positions. The intensity of these reflections is highest
for y ) 7 when the position 1a is fully occupied and the
position 1b is vacant, and the difference in the occupation
is maximal. Thus, Si46-xP Te always crystallizes at 1375 K
x y
in the space group Pm 3j because the difference in occupations
of the 1a and 1b is always present.
SEM, WDXS, and Homogeneity Range. The micro-
structure images obtained in back-scattering electron (BSE)
contrast for the samples Si31.1(1)
P14.6(1)Te7.40(3) and
Si32.00(5)P14.10(5)Te7.25(5) (WDXS determined) are shown in
Figure 4 as a representative example. It is clearly visible
that the samples are the single phases. The tellurium content
determined by WDXS is in line with that obtained from
Rietveld refinement using X-ray powder diffraction data
(
Figure 2). The P content determined by WDXS is plotted
against the Te amount (Figure 5). The ratio x to y in Si46-
Te is close to the Zintl composition, x ) 2y shown by a
xP
x
y
solid line. However, the slight but noticeable deviations
toward the phosphorus-deficient compositions are evident.
Moreover, the increase of the tellurium content is ac-
companied by an increasing phosphorus amount, which
causes a compression of the host framework and a concomi-
tant decrease of the unit cell parameter (Figures 2 and 5).
Neutron Powder Diffraction. Two single-phase clathrate
samples (1 and 2) characterized by X-ray powder diffraction,
chemical analysis, SEM, and WDXS (Table 1) were used.
Crystal structure refinement was performed in the space
group Pm 3j (Table 2). After the refinement of the profile
Figure 4. BSE image for (a) Si31.1(1)
P14.6(1)Te7.40(3) and (b)
14.10(5)Te7.25(5) (composition according to WDXS).
Si32.00(5)
P
were refined. Each framework position was then set as jointly
occupied by silicon and phosphorus. In order to exclude high
correlation of the site occupancy factors of the framework
positions with mixed Si/P occupations, the total phosphorus
amount was fixed to 14.6(1) atoms per unit cell for both
samples, which corresponds to the composition determined
by the WDXS and chemical analysis. ADPs for all frame-
parameters, background parameters, zero point correction,
and lattice parameter, the atomic coordinates for all the
framework positions, as well as for the 6g guest position,
Inorganic Chemistry, Vol. 48, No. 8, 2009 3725

Zaikina et al.
Table 2. Data Collection and Structure Refinement Parameters for
Crystal Structure Determination Experiments from Neutron Powder
a
Diffraction Data
sample 1
sample 2
refined composition
diffractometer
Si31.3P14.7Te7.3(2)
Si31.3P14.7Te7.4(1)
D2B
temperature, K
295(2) K
2.3985
wavelength, λ, Å
range 2θ, degree
step size 2θ, degree
space group
10-155
0.05
Pm 3j (no. 200)
b
lattice parameter, a, Å
9.978(1)
3.2
4.1
9.976(4)
3.1
3.9
profile R factor (R
weighted profile R
factor (Rwp
p
)
)
structure R factor (RF2
)
4.0
1.99
2.5
1.88
pseudo-Voigt
2
ꢀ
profile function
background function
14 coefficients shifted Chebyshev function
34
Figure 5. Dependence of the P content upon the Te content in Si46-x
P Te
x y
a
The esd in the Rietveld method is usually underestimated. To get
determined by WDXS analysis.
more reliable values the following equation was used: esdnew ) esdRietveld
b
N
points/Nhkl
.
The parameters are taken from Rietveld refinement. Those
which are determined on the basis of X-ray powder data with internal
standard are shown above in Table 1.
a
Table 3. Crystallographic Data for the Si46-x
P Te
x y
Samples 1 and 2 ;
space group Pm3j
X-ray single-
2
1
sample 1
sample 2
crystal data
Guest Atoms
sof Te in 1a (0,0,0)
sof Te in 1b
0.9(1)
0.4(1)
1
1
0.4(1)
0.347(2)
(
U
1/2,1/2,1/2)
Te in 1a ) 1b
iso
0.003(1)
0.011(2)
0.017(1)
0.00658(2)/
.00727(9)
0.01249(1)
0
Ueq Te in 6g (x,1/2,0) 0.018(1)
Host Framework
sof Si/P
6
f
0.685/0.315(5) 0.70/0.30(2) 0.61/0.39(4)
0.920/0.080(6) 0.90/0.10(1) 0.89/0.11(3)
Figure 6. Neutron diffraction Rietveld plot for the refinement of the sample
14.7Te7.4. Experimental data (points), calculated profile (line), and
difference profile (lower part) are shown.
8i1
8i2
12j
2
, Si31.3
P
0/1
0/1
0/1
0.667/0.333(2) 0.67/0.33(1) 0.71/0.29(2)
1
2k
1/0
0.006(1)
1/0
0.004(1)
1/0
0.0069(7)
34
x y
Table 1. Compositions of Si46-xP Te Samples 1 and 2 Determined by
Uiso of Si/P
b
Different Methods
a
The esd in the Rietveld method is usually underestimated. To get
sample 1
9.9720(1)
sample 2
9.9747(1)
more reliable values the following equation was used: esd ) esdRietveld
new
b
iso
N
points/Nhkl. Averaged U is given.
a
unit cell parameter, Å
WDXS
Si31.3(1)
P
14.7(1)Te7.45(3)
14.7(1)Te7.4(1)
Si31.2(1)
Si30.8(2)
P
P
14.8(1)Te7.40(5)
14.7(1)Te7.40(3)
and preferentially populates the 12j position. It should be
noted that low intensity of reflections with l * 2n in the hkl
zone defining the transition from the space group Pm 3j n to
the space group Pm 3j led to a rather inaccurate determination
chemical analysis
Rietveld refinement
a
Si30.7(3)
P
b
Si46Te7.3(1)
Si46Te7.4(1)
b
X-ray powder data with internal standard. All framework atoms set
as Si.
work atoms were constrained to be equal and were isotro-
pically refined. For the Te positions the site occupancy factors
and ADPs (anisotropic for 6g, isotropic and constrained to
be equal for 1a and 1b) were allowed to vary in a separate
series of the structure refinement.
of the atomic occupancies.
Summing up, the distribution of silicon and phosphorus
atoms within the framework, as well as the vacancies at the
Te site as refined using neutron powder diffraction data, is
in qualitative accordance with the crystal structure model
suggested on the basis of the X-ray single-crystal data
measured up to high sin(θ)/λ values.
The final refinement revealed that the Te position 1b is
partly occupied while the Te positions 6g and 1a remain
fully occupied within 1esd. The resulting Si to P distribution
is similar to that determined by the X-ray single-crystal
NMR Spectroscopy. To study different environments of
3
1
the atoms locally, P solid-state NMR experiments for the
single-phase sample with the Si32.5(5)P13.2(1)Te6.65(5) composi-
2
1
refinement. In particular, the 8i
the cage around the tellurium 1a position, is preferentially
occupied by silicon while the 8i position, which is next to
1
position, which builds up
tion (determined by WDXS) were performed. The featureless
asymmetric line shape of the signal did not change upon
magic-angle spinning (Figure 7, top). This evidenced that
the signal is composed of shift distributions resulting from
minute changes in the local environment, thus causing a
variation of the spectral parameters (e.g., isotropic shift, shift
2
the vacancy-containing 1b Te position, is populated by
phosphorus exclusively (Table 3 and Figure 1). Likewise,
there is a distinct difference in the 12-fold positions oc-
cupancy where silicon exclusively occupies the 12k position
3726 Inorganic Chemistry, Vol. 48, No. 8, 2009

x y
Cationic Clathrate I Si46-xP Te
Figure 7. 31P NMR signals of Si32.5
temperature and B ) 11.74 T. The SWL and MAS signals with rotation
frequencies of 7.5 and 15 kHz are normalized and separated by an arbitrary
was obtained from frequency-
P13.2Te6.6 measured at ambient
0
offset (top). Spin-spin relaxation time T
2
dependent selective excitation experiments (bottom).
anisotropy, or asymmetry). The distribution of shifts is a
characteristic feature for disordered compounds and is
frequently observed, for instance, in glasses where the first
coordination sphere determines the center of gravity of the
NMR signal. The local disorder in the second and higher
Figure 8. Thermal behavior of clathrate I Si46-x
in evacuated sealed quartz ampule for the sample Si32.30(5)
and (b) TG experiment in open quartz crucible in static air atmosphere for
x y
P Te for (a) DSC experiment
P13.70(5)Te6.90(1),
2
8
the sample Si32.7(1)P13.3(1)Te6.70(5).
coordination spheres causes the broadening of the signal.
In clathrates, the mechanism causing the broadening of the
signal contributions seems to be similar. The signal shift of
the center of gravity of the individual signal contributions
stems from the next-neighbor coordination. The broadening
results from substitutional disorder and/or partially vacant
sites in the cages and/or the framework. In the investigated
8
enhanced in comparison to Si30P16Te .
21
According to the X-ray single-crystal determination, the
estimated ratio of the signal components is 1.4:3.8:8;
however, the individual signal contributions could not be
resolved. Therefore, evidence for different P sites in
Si32.5(5)P13.2(1)Te6.65(5) is obtained by spin-spin relaxation
sample Si32.5(5)P13.2(1)Te6.65(5), the vacancies in the framework
experiments. These are sensitive to the local environment
of the nuclei since the homonuclear dipole-dipole coupling
can be reduced to extracted signal contributions by the
application of the selective excitation. The spin-spin
can be excluded, leaving only substitutional disorder in the
framework and partially vacant sites in the small cages as
3
1
the origin for the broadening of the P NMR signal.
The frequency range of the NMR signal of
relaxation time T increases by a factor of 2-4 upon the
2
Si32.5(5)P13.2(1)Te6.65(5) (from -300 to 600 ppm) is typically
application of selective excitation compared to nonselective
excitation and shows characteristic frequency dependence
observed for chemical shielding in nonconducting diamag-
2
9
netic compounds (-400 to 600 ppm). This indicates that
the signal shift is only slightly influenced by the interaction
of the P nuclei with the conduction electrons; in other words
because of the low concentration of charge carriers, the
Knight shift is far from being dominant.
(Figure 7, bottom). The low values indicate small average
distances of the interacting spins as observed at low
frequency. The high values observed at large frequencies
point out a larger average distance of the interacting nuclei.
This can be expected for the P atoms neighboring a partially
filled small cage, indicating that the signal broadening is
influenced by the change of the local environment. The
The frequency distribution is significantly larger than that
in other cationic type I clathrates Sn17Zn
7
P
22
X
8
with X )
5
c
Br, I (from 100 to 230 ppm) and type III clathrates Si172-
2
6
31
change from low to high values of T is due to the varying
2
x
P
x
Te
y
(-300 to 200 ppm). The P NMR signal of sample
3
0
amount of the different signal contributions. The crystal
structure analysis reveals that the phosphorus atoms tend to
avoid each other within the framework.
Thermal Behavior. A typical result of a DSC experiment
is shown in Figure 8 for the sample with the compostion
Si32.30(5)P13.80(5)Te6.95(1). Analysis of the data (Table 4) revealed
Si30
P16Te
8
prepared by chemical transport at significantly
lower temperatures (800 K) covers a much smaller frequency
range (-250 e δ e 150 ppm). We attribute a much broader
frequency range for Si32.5(5)P13.2(1)Te6.65(5) to the large amount
of disorder present in the title compound. All findings
indicate that the disorder in the latter clathrate is significantly
changes of the unit cell parameters after the DSC experiment,
as well as the presence of admixtures, together with traces
(
28) Zhang, P.; Dunlap, C.; Florian, P.; Grandinetti, P. J.; Farnan, I.;
Stebbins, J. F. J. Non-Cryst. Solids 1996, 204, 294–300.
29) Multinuclear NMR; Mason, I., Ed.; Plenum Press: New York, 1987.
30) Philipp, F.; Schmidt, P. J. Cryst. Growth 2008, 310, 5402–5408.
of Si
2
Te
3
and white phosphorus on top of the DSC ampule.
Te partially decomposes in
(
(
All this indicates that Si46-x
P
x
y
Inorganic Chemistry, Vol. 48, No. 8, 2009 3727

Zaikina et al.
Table 4. Results of XRD, DSC, and WDXS for the Samples Si46-x
unit cell parameter, a/Å
before DSC after DSC
x y
P Te from Homogeneity Range
composition according to WDXS
endothermic effect temperature, K
phase content after DSC
experiments in sealed under vacuum quartz ampule
a
Si32.5(5)P13.2(1)Te6.65(5)
9.9789(2)
9.9783(2)
9.9750(1)
9.97168(8)
9.9724(1)
9.9878(2)
9.9726(1)
9.9838(2)
9.9783(2)
9.9798(2)
1493
1465
1453
1445
1479
clathrate I, clathrate III
clathrate I, Si, SiP
2
Si32.30(5)P13.80(5)Te6.95(1)
Si32.00(5)P14.10(5)Te7.25(5)
clathrate I, Si
clathrate I, Si
clathrate I, Si
2
Te
2
Te
2
Te
3
3
3
Si31.1(1)P14.6(1)Te7.40(3)
Si30.80(5)P15.10(4)Te7.50(2)
experiment in open quartz crucible in static air atmosphere
9.9789(2) 9.9787(1) 1295
crystallizes in P4 /mnm space group with unit cell parameters a ) 19.2304(3) Å and c ) 10.0602(3) Å. Cristobalite.
b
Si32.5(5)
P
13.2(1)Te6.65(5)
clathrate I Si, SiO
2
a
26
y
b
Clathrate III Si172
P
x
Te
2
vacuum. Increasing the unit cell parameter for the samples
after DSC clearly demonstrates that Si46-x Te becomes
P
x
y
tellurium and phosphorus depleted during the DSC experi-
ment. Apparently upon heating, the most volatile clathrate
components s tellurium and phosphorus s tend to evaporate,
x y
thus shifting the composition of the Si46-xP Te clathrate I
phase along the x = 2y line toward a higher silicon content.
The sample Si32.5(5)P13.2(1)Te6.65(5) is close to the Te-poorest
border composition according to the established homogeneity
Te . Upon heating in vacuum, it loses
tellurium and rearranges with the formation of clathrate III
range of Si46-x
P
x
y
2
6
Si172-x
P
x
Te
a higher Si/Te ratio but is still situated on the line x ) 2y.
Such thermal behavior of clathrate I Si46-x Te is similar to
that of binary alkali-metal silicides M Si (M ) Na, K, or
Rb). Upon heating, M Si loses some amount of alkali-metal
leading to a structural rearrangement by forming either
clathrate I M Si46 (x e 8) or clathrate II M
y
.
The composition of the clathrate III exhibits
P
x
y
4
4
4
4
x
x
Si136 (x e
2
,3d,23,31
2
4).
Heating of the samples in open crucibles in static air
atmosphere showed that Si46-x Te is stable up to 1295 K
Figure 8b). Further heating leads to the decomposition of
the clathrate I phase with the evaporation of tellurium and
phosphorus (Table 4). Clathrate III Si172-x Te exhibits an
even higher thermal stability in air since it shows no sign of
P
x
y
(
P
x
y
Figure 9. Electrical resistivity vs temperature for Si46-x
x y
P Te . Inset: ln(1/
F) vs 1/T of Si30.8(1)P15.3(1)Te7.50(2).
2
6
decomposition or oxidation below 1500 K. Therefore, the
Si-P-Te clathrates may serve as a basis for the creation of
new clathrate-based materials for high-temperature applica-
tions.
x y
Si46-xP Te are weakly field dependent in the whole temper-
ature range. This is due to traces of ferromagnetic impurities
with T .400 K probably originating from steel tools used
during the sample preparation and mounting. An extrapola-
C
Electrical Resistivity and Magnetic Properties. Si46-
x
P
x
Te
electrical conductivity depending on their composition
Figure 9). A typical behavior for thermally activated
semiconductors was observed only for Si30.80(5)
The value of the band gap E ) 0.12 eV was calculated from
linear fits to the corresponding dependence of ln(1/F) upon
/T. In turn, the samples with the compositions
y
samples exhibit different temperature behavior of the
tion of ꢀ(T,H) for 1/Hf0 for high fields (Honda-Owen
method ) yields corrected values ꢀ(T) for the susceptibility
3
2
(
(
Figure 10). Traces of Curie-paramagnetic impurities and
P15.10(4)Te7.50(2).
possibly paramagnetic point defects (vacancies and structural
disorder) lead to the observed upturns in ꢀ(T) toward low
temperatures.
g
1
The corrected susceptibility can be written as ꢀ(T) ) ꢀdia
Si32.5(5)P13.2(1)Te6.65(5) and Si32.00(5)P14.10(5)Te7.25(5) demonstrate
+
ꢀ
struct(T) + ꢀ (T). Here, ꢀdia is the diamagnetism of the
C
temperature dependences of conductivity typical for heavily
doped semiconductors, which may be caused either by slight
deviation from the Zintl (x ) 2y) composition or by closing
the gap originating from band broadening due to, e.g.,
stronger disorder.
closed-shell ions, which can be calculated as the sum of the
32
diamagnetic increments. It has to be remarked that the latter
values have most often been determined from room-
temperature measurements (T ≈300 K). ꢀstruct(T) is a possible
“
structural increment” caused by molecular ring currents on
The high-field magnetic susceptibilities ꢀ(T,H) ) M/H of
(
31) Reny, E.; Gravereau, P.; Cros, C.; Pouchard, M. J. Mater. Chem. 1998,
(32) Selwood, P. W. Magnetochemistry, 2nd ed.; Interscience: New York,
8
, 2839–2844.
1956.
3728 Inorganic Chemistry, Vol. 48, No. 8, 2009

x y
Cationic Clathrate I Si46-xP Te
Figure 10. Temperature dependences of the magnetic susceptibility for
Te samples corrected by the Honda-Owen method.
Si46-xP
x
y
Table 5. Components of Magnetic Susceptibility for Si46-x
P
x
Te
y
at 300
-1
a
K in emu mol
composition
b
dia, emu mol-1
struct, emu mol-1 , emu mol-1
ꢀ
ꢀ
ꢀ
C
-
-
-
6
6
6
-6
-6
-6
-6
Si32.5(5)
P
P
P
13.2(1)Te6.65(5) -515 × 10
-45 × 10
-96 × 10
-69 × 10
550 × 10
310 × 10
310 × 10
-6
-6
Si32.00(5)
Si30.80(5)
a
14.10(5)Te7.25(5) -554 × 10
15.10(4)Te7.50(2) -571 × 10
b
See text for details. Composition is given according to the WDXS
data.
33
the cage walls in the clathrate structure plus other intrinsic
contributions. The observed weak linear-in-temperature
contribution to ꢀ(T) is, therefore, included in this ꢀstruct(T)
Figure 11. Temperature dependences of electrical resistivity and magnetic
susceptibility (H ) 7 T) of crystal Si30.7(1)
P
15.2(1)Te7.55(3) (sample 2).
term. The last term ꢀ
impurity and defect contributions, which can be modeled by
a Curie law ꢀ ) C/T. From least-squares fits in the
temperature range of 50-400 K, the values for ꢀ (300 K)
dia + ꢀstruct(300 K) and equivalent fractions ꢀ of magnetic
C
(T) represents the above-mentioned
exhibit variation in the composition and does not contain
admixtures. The magnetic susceptibility of the investigated
crystal is temperature-dependent, with the room-temperature
C
0
-6
)
ꢀ
C
value being ꢀ (300 K) ) -700(100) × 10 emu/mol, which
0
species with S ) 1/2 (calculated from the Curie constant C)
are calculated, and the values for all three terms are given
in Table 5. The resulting values for ꢀstruct(300 K) are negative
is in good agreement with the behavior of the polycrystalline
samples. Large standard deviations stem from the small mass
of the crystal. The electrical resistivity is typical for heavily
doped semiconductors. Investigation of the thermoelectric
-
6
and rather small, on average -70 ( 30 × 10 emu/mol.
For clathrates, the absolute values of the structural increment
properties of Si46-xPxTe , including the high-temperature
y
-
6
are usually significantly higher, ≈-300 × 10 emu/mol
region, is currently in progress.
3
2
for the clathrate I compounds. For the clathrate III Si172-
Te with a Si-P framework and Te anions situated in part
xP
x
y
Conclusions
-6
26
of the cages, ꢀstruct is -700 × 10 emu/mol. The clathrate
III structure has 30 cavities per formula unit compared to
eight cavities per formula unit in the clathrate I. A value of
x y
The new cationic clathrate I Si46-xP Te is synthesized from
the elements. It features very high stability (up to 1460 K in
vacuum and 1295 K in air), and its cubic unit cell parameter
is one of the smallest among clathrates. The presence of
vacancies in the guest positions leads to the wide homogene-
ity range 6.6(1) e y e 7.5(1), x e 2y at 1375 K. Vacancies
order in one of the 20-vertex cages of the clathrate framework
that corresponds to a preferential coordination of guest Te
atoms by silicon atoms. Te atoms tend to avoid cages rich
in phosphorus, which results in a change of the space group
from Pm 3j n (ideal clathrate I structure type) to Pm 3j . The
ꢀ
struct for Si172-x
P
x
Te
y
normalized for eight cages is still -200
6
×
10- emu/mol. Therefore, other intrinsic paramagnetic
contributions to ꢀstruct may, thus, play an important role in
clathrate I Si46-x Te . Together with the pronounced tem-
perature dependence of the magnetic susceptibility in the
P
x
y
x y
whole temperature range, it points out that Si46-xP Te is not
a typical Zintl phase.
Similar properties were observed for a large crystal (1.5
×
2.55 × 0.7 mm) with the composition Si30.7(1)
P15.2(1)Te7.55(3)
presence of the sites with different coordination of phos-
phorus atoms was confirmed by neutron powder diffraction
selected from sample 2 (Figure 11). According to metallo-
3
1
graphic, SEM, and WDXS studies, the crystal does not
and P NMR spectroscopy. The compositional deviation (x
e 2y) from the Zintl concept (x ) 2y) leads to the properties
alien to the Zintl phase s the samples are heavily doped
semiconductors and temperature-dependent diamagnets.
(
33) (a) Cros, C.; Pouchard, M.; Hagenmuller, P. J. Solid State Chem. 1970,
, 570–581. (b) Paschen, S.; Tran, V. H.; Baenitz, M.; Carrillo-Cabrera,
2
W.; Grin, Y.; Steglich, F. Phys. ReV. B 2002, 65 (1-9), 134435. (c)
Zaikina, J. V.; Schnelle, W.; Kovnir, K. A.; Olenev, A. V.; Grin, Y.;
Shevelkov, A. V. Solid State Sci. 2007, 9, 664–671.
Acknowledgment. Authors wish to express their gratitude
to Dr. J. Hunger and Dr. E. Suard (Institute Laue-Langevin,
(
34) (a) Sakata, M.; Cooper, M. J. J. Appl. Crystallogr. 1979, 12, 554–
5
63. (b) Berar, J.-F; Lelan, P. J. Appl. Crystallogr. 1991, 24, 1–5.
Inorganic Chemistry, Vol. 48, No. 8, 2009 3729

Zaikina et al.
Grenoble, France) for their help in collecting and refinement
of the neutron data, Mr. T. Vogel and Ms. M. Eckert for the
metallographic and WDXS investigations, Dr. G. Auffer-
mann and Ms. A. V o¨ lzke for the chemical analysis, Dr. V. I.
Shtanov for donating a sample of GeTe, and Mr. R. Koban
for the measurement of physical properties. J.V.Z. thanks
the Max-Planck Society for a research fellowship. This
research is supported in part by the Russian Foundation for
Basic Research.
Supporting Information Available: A CIF file containing
information on two structural experiments. This material is available
free of charge via the Internet at http://pubs.acs.org.
IC8023887
3730 Inorganic Chemistry, Vol. 48, No. 8, 2009