A new family of binary layered compounds of platinum with alkali metals (A = K, Rb, Cs)

Article abstract of DOI:10.1002/zaac.200500325

By reacting platinum with alkali metals (A = K, Rb, Cs) a new family of binary alkali metal platinides has been synthesized and characterized by chemical analysis, X-ray powder diffraction, thermal analysis (DTA and DSC), and magnetic measurements. All three compounds exhibit similar XRD-patterns with strong reflections that can be indexed on the basis of a rhombohedral crystal system (K_xPt: a = 2.6462(1), c = 17.123(1); Rb_xPt: a = 2.6415(1) A, c = 17.871(1) A; Cs_xPt: a = 2.6505(1) A, c = 18.536(1) A; x < 1/2). The a lattice constant is independent on the alkali metal used and of value close to the Pt-Pt distance in NaPt₂ (2.645 A). The c parameter increases monotonically with the growing atomic radius of the alkali metal. The average structure of the alloys consists of cubic close packed layers of platinum atoms with layers of disordered alkali metals in between. For all compounds besides the strong reflections small satellites are observed which cannot be indexed together with the rhombohedral peaks in any rational 3-dimensional lattice. However, these satellites can be indexed as incommensurate modulations within the ab plane (found propagation vectors k = (0.1011, 0.2506, 0) for Cs_xPt, and k = (0.0168, 0.2785, 0) for Rb_xPt).

Full text of DOI:10.1002/zaac.200500325

DOI: 10.1002/zaac.200500325
A New Family of Binary Layered Compounds of Platinum with Alkali Metals
(A ؍ K, Rb, Cs)
Andrey Karpov and Martin Jansen*
Stuttgart, Germany, Max-Planck-Institut für Festkörperforschung
Received June 17th, 2005.
Abstract. By reacting platinum with alkali metals (A ϭ K, Rb, Cs)
a new family of binary alkali metal platinides has been synthesized
and characterized by chemical analysis, X-ray powder diffraction,
thermal analysis (DTA and DSC), and magnetic measurements. All
three compounds exhibit similar XRD-patterns with strong reflec-
tions that can be indexed on the basis of a rhombohedral crystal
the alloys consists of cubic close packed layers of platinum atoms
with layers of disordered alkali metals in between. For all com-
pounds besides the strong reflections small satellites are observed
which cannot be indexed together with the rhombohedral peaks in
any rational 3-dimensional lattice. However, these satellites can be
indexed as incommensurate modulations within the ab plane
(found propagation vectors k ϭ (0.1011, 0.2506, 0) for CsxPt, and
k ϭ (0.0168, 0.2785, 0) for RbxPt).
system (KxPt:
a
ϭ
2.6462(1),
c
ϭ
17.123(1); RbxPt:
a ϭ
˚
˚
˚
2.6415(1) A, c ϭ 17.871(1) A; CsxPt: a ϭ 2.6505(1) A, c ϭ
1
˚
18.536(1) A; x < /2). The a lattice constant is independent on the
alkali metal used and of value close to the PtϪPt distance in NaPt2
˚
(2.645 A). The c parameter increases monotonically with the gro-
Keywords: Alkali metals; Binary alloys; Crystal structure; Interme-
wing atomic radius of the alkali metal. The average structure of
tallic phases; Platinum
Introduction
however the compositions were not clarified. Recently, we
synthesized the first compound in the CsϪPt system Ϫ
Cs2Pt, which remarkably represents an ionic compound ex-
hibiting an open band gap [11]. The relativistic stabilization
of the 6s2 electronic configuration and atom polarizations
are the factors favoring the stability of this binary intermet-
allic, which were not covered by Miedema due to their com-
plexity. In addition, the processes involved in adsorption of
alkali metals on platinum surfaces, which act as promoters
in many catalyticreactions, have been intensively studied
[12]. Several ordered surface structures in NaϪPt(111) and
KϪPt(111) systems have been observed with alkali metals
occupying threefold hcp (hexagonal closed packed) holes
[13, 14]. In the present paper we report on a family of iso-
typic bulk binary phases in the AϪPt systems (A ϭ K,
Rb, Cs).
Binary intermetallics is one of the most widely investigated
classes of inorganic compounds. The Pauling database con-
tains more than 45000 entries covering approximately 2500
binary systems for nonradioactive metals [1]. Several con-
cepts have been developed trying to predict possible exist-
ence of an intermetallic compound in a binary system, ac-
counting for different atomic properties of the constituents
(size, atomic number, cohesive energy, electrochemical
properties, valence electron number) [2]. One concept with
a very good agreement between theoretically predicted and
experimental findings is the Miedema’s model which calcu-
lates enthalpies of formation of an alloy using two atomic
properties Ϫ the chemical potential and the electron density
at the boundary of the Wigner-Seitz atomic cell [3]. Accord-
ing to this model, among the binary alkali metal Ϫ plati-
num systems, only in LiϪPt and NaϪPt systems stable
compounds might be formed. This prediction coincided for
a long time with experimental data. In the LiϪPt system
four bulk compounds (Li2Pt [4], LiPt [4], LiPt2 [5], and
LiPt7 [6Ϫ8]) have been synthesized and characterized, and
in the NaϪPt system one compound (NaPt2 [5, 9]) is
known. The first evidence for the existence of several com-
pounds in the CsϪPt system had been observed in [10],
Experimental Part
Starting materials
Rubidium and cesium were prepared from RbCl or CsCl, respec-
tively, by reduction with Ca, and distilled two times in vacuum [15].
Potassium (99 %, Merck, Germany) was purified by filtration of
the melt and subsequent distillation (twice) [15].
Platinum sponge (99.9 %, MaTeck, Jülich, Germany), which was
used in high temperature reactions with alkali metals (T ϭ 700 °C),
was dried and degassed by heating at 400 °C in vacuum
(pϳ2ϫ10Ϫ5 mbar). For low temperature reactions with cesium (at
200 < T < 400 °C), very active platinum powder with an average
particle size of 100 nm was prepared according to [16] as follows.
Typically, 0.5 g of K2PtCl4 was suspended in 50 ml of di-ethylene
* Prof. Dr. M. Jansen
Max-Planck-Institut für Festkörperforschung
Heisenbergstr. 1
D-70569, Germany
Fax: ϩ49-(0)711-689-1502
E-mail: M.Jansen@fkf.mpg.de
84
 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2006, 632, 84Ϫ90

A New Family of Binary Layered Compounds of Platinum with Alkali Metals
glycol (DEG) at intensively stirring, and the suspension was heated
Sample characterization
at 120 °C for 1 hour. Then the suspension was allowed to cool
down to room temperature and centrifuged. The mother solution
of KCl in DEG was decanted, while the residual black platinum
powder was washed twice and resuspended in ethanol, with inter-
mediate centrifugations. Finally, the platinum was dried overnight
at room temperature in vacuum of 10Ϫ3 mbar. The as prepared
platinum was studied by scanning electron microscopy equipped
with an integrated EDAX-EDX system, and not even traces of oxy-
gen, or other impurity elements were detected.
Chemical analyses. Chemical compositions of the samples were de-
termined using ICP-OES (Vista Pro, Varian, Darmstadt, Ger-
many). Heavy atom analyses were also performed using a scanning
electron microscope (XL 30 TMP, Philips, Holland, tungsten cath-
ode, 25 kV), equipped with an integrated EDAX-EDX system.
Laboratory X-ray Powder Diffraction. All samples were routinely
examined by X-ray powder diffraction. Powder patterns were col-
lected with a linear position-sensitive detector on a STADI P dif-
fractometer in Debye-Sherrer geometry (Stoe & Cie GmbH, Ger-
many, Ge-monochromated CuKα1 radiation, λ ϭ 1.5406 A, 5 < 2
Θ < 100 degrees, or MoKα1 radiation, λ ϭ 0.7093 A 4 < 2 Θ < 40
degrees, in steps of 0.01 degree), with the samples sealed in glass
capillaries of 0.2 mm diameter. The data were calibrated with re-
spect to an external Si-standard. To investigate thermal stability
of the products, high temperature X-ray diffraction patterns were
collected at temperatures up to 500 °C in steps of 50 °C (the accu-
racy of the temperature is 5 °C) on a STADI P diffractometer in
Debye-Sherrer geometry equipped with a heating tool (Stoe & Cie
Tantalum tubes, typically used as reaction containers, were cleaned
with dilute HF (ϳ5 %), thoroughly rinsed with water, and heated at
1100 °C in high vacuum for 12 hours (final vacuum at the maximal
temperature of ϳ2ϫ10Ϫ5 mbar). For low-temperature reactions of
cesium with platinum (at T < 400 °C), a platinum tube was used
as purchased.
˚
˚
Because of the extreme sensitivity to air and moisture of both, re-
agents and products, all operations were performed under strictly
inert conditions using Schlenk technique or in an Ar-filled glove-
box with humidity and oxygen levels both lower than 0.1 ppm (MB
150B-G-II, M.Braun, Germany).
˚
GmbH, Germany, Ge-monochromated MoKα1, λ ϭ 0.7093 A, 4 <
2 Θ < 40 degrees, in steps of 0.01 degree) with the specimens sealed
in quartz capillaries of 0.2 mm diameter.
High resolution X-ray powder diffraction. High resolution X-ray
powder diffraction data of the rubidium and cesium platinides were
collected at ambient conditions in transmission geometry with the
samples sealed in a 0.3 mm capillaries at beamline ID31 at the
European Synchrotron Radiation Facility (ESRF) (Si(111) mono-
chromator, 9-Ge(111) crystal analyzers, λ ϭ 0.32696 A, 9 scintil-
lation counters, 0.5° < 2 Θ < 38.0° in steps of 0.002° 2Θ).
Sample preparation
Potassium and rubidium platinides. For the syntheses of potassium
and rubidium platinides (KxPt and RbxPt, x < 1/2), the correspond-
ing alkali metals and platinum sponge were weighted out in differ-
ent atomic ratios (A:Pt ϭ 1:1 to 3:1, A ϭ K, Rb), and placed into
tantalum tubes, which were sealed under argon with an arc welder.
In order to prevent oxidation, the tantalum tubes were encapsu-
lated under argon in silica jackets, carefully dried with a flame. The
reaction mixtures were heated with a rate of 50 °C/h to 700 °C,
annealed at this temperature for two days, and then cooled down
to room temperature with a rate of 10 °C/h. Independently of the
atomic ratios used, unreacted alkali metals were contained in the
raw products at all times, and it was not possible to fill a capillary
for powder X-ray investigations, at this step. For characterisation
of the products, the excesses of alkali metals were removed by distil-
lation in vacuum at appropriate temperatures during appropriate
times (e.g. rubidium was completely removed by heating the mix-
tures at 130 °C overnight in a vacuum of 10Ϫ3 mbar, the excess of
potassium was removed by heating the mixtures at 250 °C over-
night in a vacuum of 10Ϫ3 mbar). The resulting powders are light
grey with small shiny crystallites.
˚
Thermal properties. Differential scanning calorimetry (DSC) was
performed with a computer-controlled DSC sensor (DSC 404 C
Pegasus, Netzsch GmbH, Germany). Differential thermal analysis
(DTA) and thermo-gravimetry were conducted on a computer-con-
trolled thermal analyzer (STA 409, Netzsch GmbH, Germany).
Powder samples (typically 20 mg) were placed in platinum (for
DSC) or corundum (for DTA-TG) crucibles with lids, heated to
600 °C with a rate of 10 °C/min, and then cooled down to room
temperature with the same rate. The whole process was run under
argon.
Magnetic measurements. Magnetisation, exemplary of rubidium
and cesium platinides, was measured using a SQUID magnet-
ometer (MPMS 5.5, Fa., Quantum Design, USA) in the tempera-
ture range 5 Ϫ 330 K at H ϭ 1, 3 and 5 T. The specimens (109.1 mg
of Rb0.33Pt and 80.8 mg of Cs0.5Pt) were sealed in silica tubes under
helium. The raw data were corrected for the holder contribution.
A small field dependency was corrected by extrapolation of χ values
to 1/H Ǟ 0.
Cesium platinide. At the same conditions, reaction of cesium with
platinum following by distillation of cesium at 100 °C in vacuum of
10Ϫ3 mbar leads to the product, light-grey shiny CsxPt (x ϳ 1/2),
containing red transparent crystals of Cs2Pt as a by-product [11].
Heating the resulting mixture at higher temperatures (T > 150 °C)
in vacuum leads to decomposition of Cs2Pt, however, the product
phase CsxPt (x < 1/2) was now contaminated by elemental platinum.
CsxPt, free of Cs2Pt and Pt, was obtained by reaction of cesium
with self-prepared platinum powder at T < 400 °C. For example, a
mixture of cesium with platinum in a ratio 2.1:1 was placed under
argon in a Pt-crucible. The latter was closed mechanically with a
nipper, and encapsulated in a carefully dried Pyrex tube, long en-
ough that one end stuck out of the furnace. The tube was heated
at 300 °C for 2 weeks. In the Pt-crucible a black powder was ob-
tained, while the excess of cesium condensed in the cold part of the
glass tube. So it was possible to characterise the as obtained powder
without additional distillation of cesium.
Results and Discussion
General Properties
By reacting the constituting elements, a new family of bi-
nary alkali metal platinides with general composition AxPt
1
(A ϭ K, Rb, Cs, x < /2) has been discovered. They crys-
tallize as light grey shiny crystals in a shape of hexagonal
plates, some tens of micrometers in size, thus we have not
succeeded in picking suitable crystals for X-ray single crys-
tal investigations. The compounds hydrolyse rapidly when
Z. Anorg. Allg. Chem. 2006, 84Ϫ90
 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
zaac.wiley-vch.de
85

A. Karpov, M. Jansen
exposed to air. According to ICP-OES and EDX analyses
the chemical composition depends strongly on the respect-
ive temperature of distillation of the alkali metal excesses.
However, in all cases the maximum alkali metal content did
not exceed one half of the platinum content. The lowest
borders of the homogeneity range are not yet well defined.
Major difficulties arise from the dependency of the contents
on the p/T conditions at the removal of excesses of alkali
metals. For KxPt the lowest potassium to platinum ratio x,
observed so far, was 0.29, for RbxPt, x was more than 0.33,
while for CsxPt it was at all times close to 0.5. According
to DSC, DTA-TG and high temperature diffraction meas-
urements, the compounds are thermally stable in argon up
to 400Ϫ550 °C (the KϪPt phase up to 540 °C, the RbϪPt
phase up to 450 °C, and the CsϪPt phase up to 420 °C),
thereafter they melt incongruently evolving alkali metals.
In vacuum the thermal stability appears to shift to lower
temperatures. In several preliminary transmission electron
microscopy experiments, conducted at room temperature,
evaporation of the alkali metals as nanometer size crystal-
lites was observed, at exposing the substances to an electron
beam. All compounds are diamagnetic at room temperature
(330 K: χ(Rb0.33Pt) ϭ Ϫ7.61ϫ10Ϫ8 emu/g, χ(Cs0.5Pt) ϭ
Ϫ1.41ϫ10Ϫ7 emu/g). The magnetic susceptibility of the ru-
bidium sample shows a temperature dependency as a conse-
quence of its contamination with the elemental platinum.
the alkali metal (Fig. 4, Table 1), however appears to be
independent on the alkali metal content.
Placing the alkali metals in the octahedral holes of the
platinum ccp (cubic closed packed) sub-lattice Ϫ which
Fig. 1 Observed powder X-ray diffraction pattern (CuKα1 radi-
˚
ation, λ ϭ 1.5406 A) for K0.4Pt, and a calculated Le-Bail fit with
peak indexing on the basis of a rhombohedral crystal system (Rp ϭ
7.52 %, Rwp ϭ 11.31 %, GOF ϭ 1.64, all values as defined in
JANA2000 [22]). Shown are the observed (circles) and calculated
(solid line) data and the enlarged difference curve between observed
and calculated profiles (in an additional window). Vertical lines
indicate the Bragg reflection positions for K0.4Pt (top) and Pt (bot-
tom). Stars indicate not indexed satellite reflections. The high angle
part is enlarged for clarity.
Crystal Structure
All compounds exhibit virtually isotypic XRD-patterns.
The strong reflections for all three compounds can be inde-
xed on the basis of a rhombohedral crystal system (Fig.
1Ϫ3, Table 1). The parameter a of the resulting hexagonal
cells is almost independent on the type alkali metal incor-
˚
porated, and ϳ2.64 A of value which is too small as com-
pared to the smallest separations possible for the respective
˚
˚
alkali metal atoms (2r(K) ϭ 4.60 A, 2r(Rb) ϭ 4.94 A,
˚
2r(Cs) ϭ 5.34 A [17, p. 47]) or even the alkali metal ions
ϩ
ϩ
ϩ
˚
˚
˚
(2r(K ) ϭ 2.76 A, 2r(Rb ) ϭ 3.04 A, 2r(Cs ) ϭ 3.34 A, all
values for CN (coordination number) ϭ 6, [17, p. 49]).When
comparing the length of the a lattice constants with the
shortest PtϪPt distance of elemental platinum (d(PtϪPt) ϭ
˚
2.77 A [18]), it can be noted that it is significantly shorter.
Assuming a full occupation of the platinum positions the
reduction of a as compared to d(PtϪPt) in elemental plati-
num could be ascribed to additional, covalent interactions,
similarly to those observed in the one-dimensional homo-
Fig. 2 Observed high resolution synchrotron powder X-ray diffrac-
tion pattern (λ ϭ 0.32696 A) for Rb0.33Pt, and a calculated Le-Bail
˚
fit with peak indexing on the basis of a rhombohedral crystal sys-
tem (Rp ϭ 10.94 %, Rwp ϭ 15.95 %, GOF ϭ 3.22, all values as
defined in JANA2000 [22]). Shown are the observed (circles) and
calculated (solid line) data and the enlarged difference curve be-
tween observed and calculated profiles (in an additional window).
Vertical lines indicate the Bragg reflection positions for Rb0.33Pt
(top) and Pt (bottom). Stars indicate not indexed satellite reflec-
tions. The high angle part is enlarged for clarity.
˚
atomar platinum chains of BaPt (d(PtϪPt) ϭ 2.71 A [19]).
˚
A similar contraction to d(PtϪPt) ϭ 2.645 A has been ob-
served in NaPt2 (Laves phase), where platinum atoms are
coordinated by six platinum atoms and six sodium atoms
in the shape of an icosahedron [9]. The c lattice parameter,
and thus the corresponding inter-layer distance c/3, in-
creases monotonically with the growing atomic radius of
86
zaac.wiley-vch.de
 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2006, 84Ϫ90

A New Family of Binary Layered Compounds of Platinum with Alkali Metals
˚
˚
˚
corresponds to the position most distant from Pt Ϫ one can
deduce the maximal possible space (in atomic radii, rmax
available within the alkali metals in the platinides
(rmax(AA Pt) ϭ
3a/3)2 ϩ (c/6)2 Ϫ a/2; rmax(KK Pt) ϭ
1.91 A, rmax(RbRb Pt) ϭ 2.03 A, rmax(CsCs Pt) ϭ 2.12 A).
0.33
0.5
)
The minimal alkali metal radii can be estimated assuming
completely disordered layers of the alkali metals between
the platinum layers (rmin(A
)
ϭ
(c/3
Ϫ
a)/2;
1.66 A,
Ί
(Ί
AxPt
x
0.4
˚
˚
rmin(KK
)
ϭ
1.53 A, rmin(RbRb
)
ϭ
0.4Pt
0.33Pt
˚
rmin(CsCs Pt) ϭ 1.76 A). The ranges of “allowed” (within
0.5
the Pt sublattice) alkali metal radii thus fall into the range
between the respective ionic radii and atomic radii in met-
als. As is quite obvious, for all compounds a significant re-
duction of the alkali metal as compared to the atomic radii
in metals is observed, suggesting a considerable degree of
their ionization. Noteworthy, the arithmetic means of the
estimated range of tolerable alkali metal radii in the platin-
ides is getting closer to the ionic radii when moving from
potassium through rubidium to cesium, in full accordance
with the trend of their first ionization potentials. A similar
conclusion can be drawn comparing the volume of Cs0.5Pt
3
˚
per formula unit (VCs Pt ϭ 37.56 A , assuming Z ϭ 3) with
0.5
the volume increments as reported by Biltz [20] for the ionic
3
3
˚
˚
ϭ 41.99 A ).
0.5Pt
and metallic radii (VCs
ϭ 22.04 A , VCs
ϩ0.5Pt
By inspecting the full-widths of half-maxima (FWHM)
of the powder reflections a strong anisotropy can be de-
tected. The 00l and hh0 peaks have three to six times
smaller full-widths at the same 2Θ as compared to the
peaks with one non-zero h index (Fig. 5). This is giving
strong evidence for a significant disorder of this layered
structure within the ab plane.
Besides the rhombohedral reflections discussed above, a
series of peaks with low intensities has been observed in all
samples. Their positions and relative intensities are listed in
Table 2. In trying to index them together with the rhombo-
hedral reflections, we have failed finding any rational 3-di-
mensional lattice. In a second approach, we have tried to
assign these peaks to a propagation vector using the pro-
gram SUPERCELL of the program package FULLP-
ROF2000 [21] assuming incommensurate modulations in
the ab plane (the reciprocal component z* of the propa-
gation vector was fixed to 0). The best solution we found
Fig. 3 Observed high resolution synchrotron powder X-ray diffrac-
tion pattern (λ ϭ 0.32696 A) for Cs0.5Pt, and a calculated Le-Bail
˚
fit with peak indexing on the basis of a rhombohedral crystal sys-
tem (Rp ϭ 16.86 %, Rwp ϭ 22.02 %, GOF ϭ 2.73, all values as
defined in JANA2000 [22]). Shown are the observed (circles) and
calculated (solid line) data and the enlarged difference curve be-
tween observed and calculated profiles (in an additional window).
Vertical lines indicate the Bragg reflection positions for Cs0.5Pt.
Stars indicate not indexed satellite reflections. The intensity scale
is interrupted between 4500 and 16000 counts, for clarity.
Table 1 Lattice parameters of the alkali metal platinides for the
rhombohedral substructures.
3
3
˚
˚
˚
˚
˚
Formula a /A
c /A
0.333·c /A V /A
0.333·V /A
K0.4Pt
Rb0.33Pt
Cs0.5Pt
2.6462(1) 17.123(1) 5.708(-)
2.6415(1) 17.871(1) 5.957(-)
2.6505(1) 18.536(1) 6.179(-)
103.839(5) 34.613(-)
107.991(2) 35.997(-)
112.768(2) 37.589(-)
Table 2 d values and relative intensities of the reflections which
were not indexed on the basis of the rhombohedral subcell or were
assigned to the impurity of the elemental platinum.
K0.4Pt
Rb0.33Pt
Cs0.5Pt
˚
˚
˚
d /A
I /%
9.6
5.8
5.0
6.4
4.1
3.9
4.7
7.2
3.8
3.9
5.6
d /A
I /%
1.39
1.22
0.93
1.41
0.88
0.54
0.55
0.80
0.67
0.53
d /A
I /%
1.07
2.50
3.30
0.85
2.08
1.36
0.70
0.70
1.11
0.64
0.59
0.52
0.45
0.44
2.726
2.490
2.043
1.863
1.638
1.549
1.495
1.411
1.210
1.116
1.102
3.156
3.018
2.789
2.606
2.387
1.851
1.770
1.562
1.464
1.451
3.390
3.209
3.076
2.858
2.667
2.448
1.881
1.800
1.611
1.513
1.497
1.389
1.194
1.169
Fig. 4 The dependence of the a (left y-axis) and c/3 (right y-axis)
lattice constants of AxPt on the alkali metal.
Z. Anorg. Allg. Chem. 2006, 84Ϫ90
 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
zaac.wiley-vch.de
87

A. Karpov, M. Jansen
Fig. 6 A Le-Bail fit for Rb0.33Pt according to the 5-dimensional
indexing (propagation vector k ϭ (0.0168, 0.2785, 0), Rp ϭ 7.99 %,
Rwp ϭ 10.29 %, GOF ϭ 2.08, all values as defined in JANA2000
[22]). Shown are the observed (circles) and calculated (solid line)
data and the enlarged difference curve between observed and calcu-
lated profiles (in an additional window). Vertical lines indicate the
Bragg reflection positions for Rb0.33Pt on the basis of a rhombo-
hedral crystal system (top), incommensurately modulated satellite
reflections (middle) and Pt (bottom). Satellite reflections with non-
zero intensities are labelled with five indices.
Fig. 5 FWHM versus diffraction angle 2Θ for Rb0.33Pt. The ob-
served XRD-profiles of 009 and 104 reflections for Cs0.5Pt are
shown in the insert window, demonstrating the difference between
their shapes.
for Cs0.5Pt was k ϭ (0.1011, 0.2506, 0). For Rb0.33Pt the
best k was found to be (0.0168, 0.2785, 0) c.f. Fig. 6. No
solutions have been found with propagation vectors along
the c or a axes only (keeping h,k ϭ 0, or k,l ϭ 0, respec-
tively). Attempts to find any propagation vector for K0.4Pt
from laboratory powder X-ray diffraction data failed prob-
ably due to an insufficient resolution of the satellite peaks.
The different propagation vectors found for the compounds
are more likely an effect of the different alkali metal radii
rather than of the different stoichiometry. Attempts to de-
termine the supercell using electron diffraction have failed,
so far, due to the low stability of the samples under an elec-
tron beam.
tances from the platinum atoms as compared to the sum of
their ionic or metallic radii. Hence these space groups can
be ruled out. The structure solutions in the non-centrosym-
metrical space groups, however, give plausible estimations
for the alkali metal radii as discussed above. Fig. 8 shows a
Rietveld plot for rubidium platinide using a model of dis-
ordered rubidium layers centered between the ccp platinum
layers (space group R3m, Pt at 3a (0,0,0), Occ. (occupation
factor) ϭ 1, Rb1 at 18c (0,1/3,1/6), Occ. ϭ 0.056, Rb2 at
3a (0,0,1/6), Occ. ϭ 0.056, Rb2 at 3a (1/3,2/3,1/6), Occ. ϭ
0.056, Rb2 at 3a (2/3,1/3,1/6), Occ. ϭ 0.056, total chemical
composition Rb0.5Pt, Fig. 9). Besides differences related to
the satellite reflections, a quite good coincidence between
the calculated and measured XRD-patterns was thus
achieved.
At analyzing the average crystal structures we placed
platinum atoms at the unit cell’s origin (0,0,0). When mov-
ing from cesium through rubidium to potassium the differ-
ence between the calculated XRD-patterns of the platinum
sublattice and measured XRD-patterns becomes lower, be-
cause of the decreasing scattering contribution of the alkali
metals as compared to those of platinum. The main dis-
agreement is observed for the ratio of intensities of the
(0,0,3) to (0,0,6) reflections (c.f. Fig. 7). The subsequent dif-
Conclusions
¯
ference Fourier syntheses in several space groups (R3, R3,
¯
R3m, R3m) did not reveal any maxima with electron densi-
New binary alloys of platinum with alkali metals (A ϭ K,
Rb, Cs) have been synthesized. In a first approximation,
the crystal structure of the compounds can be described
consisting of ccp platinum layers with a PtϪPt distance of
ties comparable to the scattering factors of the correspond-
ing alkali metal. This points out partial occupation of crys-
tallographic positions by alkali metals in rhombohedral
space groups. Optimizing occupation factors we found the
best coincidence between calculated and measured XRD-
patterns with alkali metals building disordered bilayers at
˚
ϳ2.64 A separated by disordered and partly occupied alkali
metal layers. A significant reduction of the PtϪPt separ-
ation as compared to those in the elemental platinum
˚
¯
z ϭ 1/6
1/24 in centrosymmetrical space groups (R3,
(2.77 A) can be assigned to covalent interactions as earlier
¯
R3m), or a monolayer at z ϭ 1/6 in non-centrosymmetrical
space groups (R3, R3m). In the case of centrosymmetrical
space groups the alkali metals fill space at too small dis-
in the case of BaPt [19]. The estimated alkali metal radii
are considerably reduced as compared to the atomic alkali
metal radii, thus suggesting ionization of the alkali metals
88
zaac.wiley-vch.de
 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2006, 84Ϫ90

A New Family of Binary Layered Compounds of Platinum with Alkali Metals
Fig. 7 A Rietveld fit of “Rb0.5Pt” using a model with only platinum
atoms at the unit cell’s origin (R(all) ϭ 11.53 %, Rw(all) ϭ 12.78 %,
Rp ϭ 17.45 %, Rwp ϭ 22.90 %, GOF ϭ 4.62, all values as defined
in JANA2000 [22]). For labelling see Fig. 2.
Fig. 9 Crystal structure of “Rb0.5Pt” as refined by Rietveld profile
fitting (black spheres: platinum atoms; light grey spheres: rubidium
atoms; black lines: unit cell-edges).
with transfer of part of the charge to the platinum. Hence,
the alloys can be considered as further members of platin-
ides Ϫ compounds consisting of negatively charged plati-
num ions.
However, some problems with the superstructure solu-
tion, arising from a series of small satellite reflections, re-
main not settled yet. The peaks have been indexed using
modulation vectors within the ab plane but the modulated
structure has not been solved. Unfortunately due to the low
stability of the samples in vacuum our preliminary attempts
to make transmission electron microscopy have failed, so
far. The possible way to overcome this problem might be in
conducting TEM experiments at cooling with liquid nitro-
gen. Another approach to solve the modulated structure
might be to grow single crystals under high pressure Ϫ this
procedure was already successfully applied for the growth
of single crystals constituted of one component with a low
melting point and a second with a high melting point, e.g.
NaPt2 [9].
Acknowledgement. The authors would like to thank Claus Mühle
for the assistance at the experimental work, Marie-Luise Schreiber
for the chemical analysis, Priv. Doz. Dr. Robert E. Dinnebier and
Dr. Andy Fitch for performing the high resolution X-ray powder
diffraction measurements, Eva Brücher for the magnetic measure-
ments.
References
[1] Pauling File, Inorganic Materials Database and Design Sys-
tem, Binary Edition, Ed.-in-chief P. Villars, Version 1.0, Re-
lease 2002/1.
[2] P. Villars in Intermetallic Compounds Principles and Practice,
Eds. J. H. Westbrook, R. L. Fleischer, John Wiley Sons, Chich-
ester, U.K., 1994, Vol. 1, p. 227Ϫ275.
[3] F. R. de Boer, R. Boom, W. C. M. Mattens, A. R. Miedema,
A. K. Niessen in Cohesion in Metals, Transition Metal Alloys,
Eds. F. R. de Boer, D. G. Pettifor, Cohesion and Structure,
North-Holland Physics Publishing, Amsterdam, 1988, Vol. 1.
[4] W. Bronger, B. Nacken, K. Ploog, J. Less-Common Met. 1975,
43, 143.
[5] C. P. Nash, F. M. Boyden, L. D. Whittig, J. Am. Chem. Soc.
1960, 82, 6203.
[6] W. Bronger, W. Klemm, Z. Anorg. Allg. Chem. 1962, 319, 58.
[7] O. Jr. Loebich, C. J. Raub, J. Less-Common Met. 1980, 70,
P47.
Fig. 8 A Rietveld fit of “Rb0.5Pt” using a model with platinum
atoms at the unit cell’s origin and rubidium atoms forming dis-
ordered layers in the middle between the platinum layers (R(all) ϭ
5.12 %, Rw(all) ϭ 4.57 %, Rp ϭ 12.80 %, Rwp ϭ 17.05 %, GOF ϭ
3.45, all values as defined in JANA2000 [22]). For labelling see
Fig. 2.
Z. Anorg. Allg. Chem. 2006, 84Ϫ90
 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
zaac.wiley-vch.de
89

A. Karpov, M. Jansen
[8] W. Bronger, G. Klessen, P. Müller, J. Less-Common Met. 1985,
109, L1.
[9] K.-J. Range, F. Rau, U. Klement, Acta Crystallogr. C 1989,
45, 1069.
[15] G. Brauer, Handbuch der Präparativen Anorganischen Chemie,
Bd. 2, Ferdinand Enke Verlag, Stuttgart, 1978, p. 938Ϫ942.
[16] L. K. Kukihara, G. M. Chow, P. E. Schoen, NanoStructured
Mat. 1995, 5, 607.
[10] M. Jansen, A.-V. Mudring, Diploma Thesis, A.-V. Mudring,
Rheinische Friedrich-Wilhelms-Universität, Bonn, 1995.
[11] A. Karpov, J. Nuss, U. Wedig, M. Jansen, Angew. Chem. 2003,
115, 4966; Angew. Chem. Int. Ed. 2003, 42, 4818.
[12] H. P. Bonzel in Physics and Chemistry of Alkali Metal Adsorp-
tion, Eds. H. P. Bonzel, A. M. Bradshaw, G. Ertl, Elsevier,
Amsterdam, 1989.
[17] U. Müller, Anorganische Strukturchemie, B. G. Teubner,
Stuttgart, 1996.
[18] H. E. Swanson, E. Tatge, J. Res. Nat. Bur. Stand. 1951, 46, 318.
[19] A. Karpov, J. Nuss, U. Wedig, M. Jansen, J. Am. Chem. Soc.
2004, 126, 14123.
[20] W. Biltz, Raumchemie der festen Stoffe, Verlag von Leopold
Voss, Leipzig, 1934.
[13] J. B. Hannon, M. Giesen, C. Klinker, G. Schulze Icking-
Konert, D. Stapel, H. Ibach, J. E. Müller, Phys. Rev. Lett.
1997, 78, 1094.
[21] J. Rodriguez-Carvajal, FULLPROF Version Juli 2001, CEA/
Saclay, France, 2001.
[22] V. Petricek, M. Dusek, Jana2000, Version 21/10/2003. The
crystallographic computing system, Institute of Physics,
Praha, Czech Republic, 2000.
´
[14] S. More, A. P. Seitsonen, W. Berndt, A. M. Bradshaw, Phys.
Rev. B 2001, 63, 075406.
90
zaac.wiley-vch.de
 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2006, 84Ϫ90