DOI: 10.1002/zaac.
200600308
Crystal Growth and Structure Refinement of K4Ge91)
Simeon Ponou and Thomas F. Fssler*
Mnchen/Germany, Department Chemie der Technischen Universitt Mnchen
Received November 14th, 2006.
Abstract. Deep-red moisture and air sensitive single crystals of
K4Ge9 were obtained by reacting GeO2 and elemental Ge with
metallic W and K at high temperature in a niobium ampoule. The
crystal structure of the compound was determined by single crystal
X-ray diffraction experiments. K4Ge9 crystallizes in a polar space
group R3c (No. 161), Z 4 with a 21.208(1) and c
25.096(2) A . The compound contains discrete Ge94 Wades nido-
clusters which are packed according to a hierarchical atom-to-
1 Introduction
The isolated clusters [E9] (E Si, Ge, Sn, Pb) are well
4
known in solution and have been stabilized with a variety
of counter ions and solvates [13]. In solid state, they are
found in simple binary intermetallic phases A4E9 [49] and
A12E17 [1012], but their crystal structures are made diffi-
cult by problems with crystal quality, multiple twinning,
structural disorder or large unit cells [1112].
No silicon phase with the composition A4Si9 is known.
In case of the germanides A4Ge9, reliable data are available
to the best of our knowledge only for the crystal structure
of Cs4Ge9 [5]. Unit cell parameters suggest Rb4Ge9 to be
isostructural with the Cs analogue, but the poor crystalline
quality of the single crystal always prevents the satisfactory
refinement of the data [5]. For the stannides, only the crys-
tal structure of K4Sn9 [9] have been reported whereas those
of the plumbides A4Pb9 (A K, Rb and Cs) are all de-
scribed [68].
Several unsuccessful attempts have been made to struc-
turally characterize the potassium analogue K4Ge9 whose
crystal structure is still missing. Only the unit cell param-
eters were reported but the crystal structure could not be
refined because of the poor crystalline quality of the single
* Prof. Dr. T. F. Fssler
Technische Universitt Mnchen
Department Chemie
Lichtenbergstrae 4
D-85747 Garching, Germany
Fax: 49 8928913186
e-mail:Thomas.Faessler@lrz.tum.de
1)
This work is supported by the European Union within the RTN
program (EU-project Nr. HPRN-CT 2002-00193).
Z. Anorg. Allg. Chem. 2007, 633, 393397 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
cluster replacement of the Cr3Si prototype and are separated by
K cations. Two independent [Ge9]4 clusters A (at Cr positions)
and B (at Si positions) are found with a ratio A:B 3:1. The B-
type cluster satisfactorily represents orientational disorder around
the three-fold axis.
Keywords: Alkali metal germanides; Zintl phases; Crystal structure
crystals [1213]. Herein, we report on the refinement of the
crystal structure of K4Ge9 from a single crystal that is
grown by a metallothermie-like reaction.
2 Experimental Section
Synthesis
The phase K4Ge9 was obtained by reacting elemental Ge
(Chempur, 99.9999 %) and GeO2 (Chempur, 99.999 %) with W
(Chempur, 2.8 powder, 99.95 %) and K (Merck, 99 %). The reac-
tion was originally chosen to produce the double-salt K10Ge18WO4
[14]. The simple phases K4Ge9 and K2WO4 were often obtained
instead. Single crystals of the binary phase K4Ge9 suitable for dif-
fraction experiment, were obtained from the reaction with the start-
ing composition W:GeO2:Ge:K 1:2:7:6. The mixtures (1.5-2 g)
were loaded in a niobium ampoule and heated at a rate of 2 K/min
to 600 C, held for 24 hours, then heated up again at 1 K/min to
950 C, held for 5 h and, subsequently cooled down to room tem-
perature at a rate of 0.1 K/min.
EDX analyses and X-ray powder diffraction
Semi quantitative EDX analysis of the single crystals was carried
out using a JEOL 5900LV electron microscope system operating at
20 kV and equipped with a LINK AN 10000 detector system for
EDX analysis. The phase analysis of the product was carried out
by X-ray powder diffraction using a Stoe STADI P2 diffractometer
(Ge(111) monochromator for Cu-K1 radiation: 1.54056 A )
equipped with a linear position sensitive detector (PSD) with
eff 40. The powder was finely ground in an agate mortar to a
homogeneous powder and filled into a glass capillary which was in
turn sealed using a red-hot platinum wire. The theoretical diagram
of K4Ge9 was calculated from the structure obtained from single
crystal analysis.
393
 S. Ponou, Th. F. Fssler

Table 1 Crystallographic data and details of structure refinement Table 2 Atom coordinates and equivalent isotropic atomic dis-
parameters for K4Ge9 placement parameters Ueq 1) for the compound K4[Ge9] (standard
deviation are in parentheses)
Chemical formula K4Ge9
Atom Wyck. s.o.f x y z 2
Ueq / A
Formula weight 809.71
Temperature 293(2) K Cluster A
Crystal size 0.20.20.25 mm3 Ge1 18b 1 0.8684(2) 0.0487(2) 0.2024(2) 0.058(1)
Crystal color and habit deep-red block Ge2 18b 1 0.7763(2) 0.0802(2) 0.1620(2) 0.052(1)
Diffractometer Oxford Xcalibur 3 Ge3 18b 1 0.7688(2) 0.9416(2) 0.1500(2) 0.055(1)
Ge4 18b 1 0.7896(2) 0.9357(2) 0.2588(2) 0.053(1)
Crystal system rhombohedral Ge5 18b 1 0.7937(2) 0.0666(2) 0.2756(2) 0.050(1)
Space group R3c (Nr. 161) Ge6 18b 1 0.6550(2) 0.9576(2) 0.1521(2) 0.054(1)
Unit cell dimensions a 21.208(1) A ; c 25.096(2) A
Ge7 18b 1 0.6638(2) 0.8632(2) 0.2136(2) 0.058(1)
Unit cell Volume 9775.4(8) A3 Ge8 18b 1 0.6775(2) 0.9434(2) 0.2944(2) 0.055(1)
Z 24 Ge9 18b 1 0.6715(2) 0.0368(2) 0.2316(2) 0.053(1)
Density calculated 3.30 g/cm3 Cluster B
Absorption coefficient (Mo K) 17.34 mm1 ( 0.71073 A ) Ge1 6a 1 0 0 0.0422(3) 0.063(2)
F(000) 8736 Ge2 18b 1 0.4094(2) 0.6798(3) 0.0565(2) 0.067(2)
2)
Theta range for data collection 3.3320.30 Ge3 18b 1/3 0.017(1) 0.906(1) 0.9879(5) 0.043(3)
Index range 20 h 19, 20 k 20, Ge4 18b 1/3 0.062(1) 0.949(1) 0.9732(6) 0.058(4)
2)
Ge5 18b 1/3 0.973(1) 0.890(1) 0.9927(6) 0.059(4)
19 l 24 2)
Ge6 18b 1/3 0.119(1) 0.041(1) 0.9780(6) 0.050(3)
Reflections collected 12106(R 0.076) Ge7 18b 1/3 0.096(1) 0.997(2) 0.9656(7) 0.063(5) 2)
Independent reflections 1780 (Rint 0.158)
Data completeness 99.4 % Cations
K1 18b 1 0.5999(6) 0.8059(5) 0.0740(4) 0.071(3)
Refinement method full-matrix least-squares on F2
K2 6a 1 0 0 0.2563(7) 0.071(5)
Parameters / restraints 168 / 1 K3 18b 1 0.7590(5) 0.0123(5) 0.0283(4) 0.072(3)
Goodness of fit on F2 1.169 K4 18b 1 0.6544(6) 0.1172(7) 0.1053(8) 0.136(7)
Observed reflections [I>2(I)] 1684 K5 18b 1 0.8453(6) 0.8412(6) 0.0975(4) 0.080(3)
Final R indices [I>2(I)] R1 0.069, wR2 0.196 K6 18b 1 0.8172(6) 0.2386(5) 0.2323(5) 0.083(3)
Final R indices (all data) R1 0.073, wR2 0.199
1)
Flack parameters 0.10(8) Ueq is defined as one third of the trace of the orthogonalized Uij tensor. 2)
Isotropic
Residual map [e/A 3] 1.12 / 1.30 parameter.

w 1/[2(Fo2)(0.1303P)2 132.17P] where P (Fo2 2Fc2)/3

Crystallographic studies
Several single crystals were selected in an argon filled glove-box
and mounted in a glass capillary which was subsequently sealed to
prevent contact with air. The crystals were first checked for singu-
larity and the best crystal was used for full data collection at room
temperature on the Oxford Xcalibur3 diffractometer with a CCD
detector with monochromated Mo-K radiation. Three series of
138 frames were collected with an exposure time of 45 sec each
with a crystal-detector distance of 50 mm. The reflections were col-
lected over the range 3.33 < < 25.02 and corrected for absorp-
tion (empirical) using the program CrysAlis RED (Oxford Diffrac-
tion Ltd). The cell parameters of the rhombohedral lattice were
determined and refined using all the integrated peaks. Data were
processed using the SHELXTL package [15]. Solution (direct
methods) and refinement (on F2) were performed through the pro-
grams SHELXS97 [15b] and SHELXL97 [15c] in the polar space
group R3c for a fairly large cell (ca. 212125 A 3), in agreement
with the lattice parameters reported by von Schnering and co-wor-
kers [13]. Eleven apparent Ge positions were assigned on the basis
of peak heights and six K positions judging from bond distances
and peak heights. Subsequent least-squares refinements and differ-
ence Fourier syntheses quickly located five other weak peaks that
have been assigned to defect Ge positions. The occupancies of these
positions were allowed to vary, and they converged successively to
approximately 1/3, in agreement with the expected composition.
The occupancies at those five positions were therefore fixed to 1/
3 and their ADPs (atomic displacement parameters) were refined
isotropically. To obtain acceptable residuals, the obverse/reverse
twinning which is typical in acentric crystals with rhombohedral
symmetry was applied. The final refinement on F2 yields the re-
sidual factors R1 0.073 and wR2 0.199 for all data [R1 0.069
and wR2 0.196 for I > 2(I)]. The Flack parameter [15d] of
0.10(8) indicates that the absolute structure is correct. The residual
Fourier map was essentially featureless. Relevant crystallographic
and refinement details are listed in Table 1 and, Table 2 contains
the atomic positions and equivalent displacement parameters.
Further details of the crystal structure investigations can be ob-
tained from Fachinformationszentrum Karlsruhe, 76344 Egg-
enstein-leopoldshafen, Germany, (fax: (49)7247-808-666; e-mail:
crysdata@fizkarlsruhe.de) on quoting the depository number
CSD-417133.
3 Results and Discussions
Some deep red crystals with irregular shape could be ex-
tracted from the reaction mixture. EDX analysis of those
crystals revealed the presence of only two elements K and
Ge with a K:Ge atomic ratio close to 1:2, indicating a
K/Ge binary phase. According to powder diffraction dia-
gram of the reaction product, the reaction yields a mixture
of K4Ge9 and K2WO4 corresponding to the equation:
W 2 GeO2 7 Ge 6 K K4Ge9 K2WO4
Neither GeO2 nor W metal could be detected in the X-
ray powder diagram of the product, meaning that the reac-
tion mechanism involves a metallothermic step in which
GeO2 is first reduced to elemental Ge by W and K. Further
reduction of elemental Ge by K metal results in the Zintl
phase K4Ge9.

394 www.zaac.wiley-vch.de 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Z. Anorg. Allg. Chem. 2007, 393397
Crystal Growth and Structure Refinement of K4Ge9

Table 3 Interatomic distances within Ge94 cluster in K4Ge9

(standard deviations are in parentheses)

Cluster A Cluster B disordered

Atomic pair
distance /A Atomic pair
distance /A

Ge1 Ge4 2.555(6) Ge5 Ge1 2.45(1)

Ge3 2.560(6) Ge4 2.54(2)
Ge2 2.568(6) Ge7 2.56(3)
Ge5 2.574(6) Ge2 2.66(2)
Ge2 Ge6 2.599(6) Ge1 Ge4 2.69(1)
Ge9 2.605(6) Ge3 2.59(1)
Ge3 2.878(6) Ge6 2.75(1)
Ge5 2.906(6) Ge7 2.83(2)
Ge3 Ge7 2.563(6) Ge2 Ge2 2.586(7) 2
Ge6 2.699(6) Ge7 2.80(2)
Ge4 2.778(6) Ge4 3.09(3)
Ge4 Ge7 2.582(6) Ge3 Ge6 2.60(2)
Ge8 2.620(6) Ge2 2.62(1)
Ge5 2.766(5) Ge7 2.51(3)
Figure 1 Projection of the crystal structure of K4Ge9 along the Ge5 Ge8 2.586(6) Ge4 Ge6 2.56(2)
c-direction. The ordered Ge9 clusters (type A) are shown as poly- Ge9 2.589(6) Ge2 2.51(2)
hedra. Black spheres are Ge positions in the disordered cluster Ge6 Ge9 2.517(6) Ge2 Ge2 2.586(7)
Ge7 2.607(6) Ge6 Ge2 2.52(2)
(type B) and K atoms are shown as isolated grey spheres. Ge7 Ge8 2.568(6) Ge7 Ge2 2.52(2)
Ge8 Ge9 2.585(6)

K4Ge9 crystallizes with the space group R3c and unit cell Ge6 Ge8 3.634(6) 1)
Ge6 Ge2 3.46(2) 1)

parameters a 21.208(1) and c 25.096(2) A . Its structure Ge7 Ge9 3.631(7) 1)

Ge3 Ge2 3.82(2) 1)

can be described as built up of isolated nine-atom clusters 1)

of germanium surrounded by potassium cations (Figure 1).
This is consistent with the ionic formulation as
(4K)[Ge9]4 in accordance with Wades rules electron pre-
cise nido-clusters with 22 skeleton electrons. There are two
crystallographically independent Ge9 clusters in the struc-
ture: type A cluster and type B cluster (disordered) in a
ratio A : B 3 : 1. The type A cluster adopts a monocapped
square antiprism with only small deviations from the ideal
C4v symmetry. The type B cluster approximately represents
threefold orientational disorder around the threefold axis.
This means that the site is statistically occupied by three
overlapping clusters with different orientations coinciding
only with four over nine atomic positions. This extensive
disorder is indicative of a lowering of local symmetry.
Alternative descriptions were also extensively checked. A
careful exploration of the reciprocal space to detect any ad-
ditional scattering features reveals neither classical super-
structure reflections (or even incommensurate satellites),
nor diffuse scattering. We therefore conclude that the trans-
lational cell is correct and that the disorder is best modelled
by using a split model.
Interatomic contacts (Table 3) within cluster A range
from 2.517(6) to 2.906(6) A with two types of effective bond
lengths, applying the bond classification according to Fig-
ure 2. Shorter bonding distances are found in group I
(2.555(6) to 2.574(6) A ), group III (2.563(6) to 2.699(6) A )
and group IV (2.517(6) to 2.607(6) A ). The bond distances
of group II are significantly longer and range from 2.766(5)
to 2.906(6) A . This is in agreement with the general pattern
of the bond distances distribution observed for all similar
nine-atom nido-cluster of the tetrel group [E9]4. The ratio
of the two diagonals d1/d2 for cluster A is 1.00, indicating
very small deviation from the ideal C4 symmetry of a mon-
ocapped square antiprism [1].
Corresponding to the diagonals of the basal (open) square.
Cluster B, on the other hand, is disordered around the
threefold axis giving three orientations which coincide with
four of its nine atomic positions. These fixed positions are
Ge1 at the special position 6a on the C3 rotation axis and,
Ge2 at the general position 18b with three positions that
are related by the C3 symmetry. The five other atomic posi-
tions (general position, 18b) are virtually repeated by the
symmetry operations of the threefold axis, leading to 1/3
occupancy. The orientational disorder is ascribed to the re-
sulting three possible symmetry equivalent orientations of
the cluster B. The analysis of the electron density distri-
bution around the disordered atoms suggests that some of
them are slightly offset from the ideal (symmetry equiva-
lent) positions. This indicates incomplete orientational or-
dering which is usually observed in the structure of different
phases featuring E9 nido-clusters. The ADPs of these atoms
(Ge3, Ge5, Ge6, Ge7) could be refined only iso-
tropically. Consequently, the three symmetry equivalent
orientations of the cluster B are just average orientations.
In the average model, the three orientations of cluster B are
constrained to be symmetrically equivalent, and corollary
the structure of the cluster in the three different orientations
is virtually the same (Figure 2).
The intra-cluster bonds distances (Table 3) in the dis-
ordered cluster B vary from 2.45(1) to 3.09(3) A with nearly
the same bond distribution pattern as for cluster A. The
shortest bond distances are found in group I (2.45(1) to
2.66(1) A ), group III (2.52(1) to 2.62(1) A ) and group IV
(2.51(3) to 2.59(1) A). The bond distances in group II that
involve five-bonded atoms are generally longer and range
from 2.69(1) to 3.09(3) A . The strong distortion of cluster
B is mirrored by the rather long distances like 3.09(3) A

Z. Anorg. Allg. Chem. 2007, 393397 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.zaac.wiley-vch.de 395

Figure 2 View of the type-A Ge9 cluster in K4Ge9 with labeling
of the bonds lengths in four groups (I, II, III, IV) according to
approximate C4 point group symmetry (a). Representation of the
type-B disordered Ge9 cluster (top) and the corresponding three
symmetry related orientations (around the three-fold axis) in the
crystal structure of K4Ge9. The Ge1 and Ge2 positions (dark
grey) coincide in all orientations; Ge5 caps one square face of the
square antiprism (b).
Figure 3 Coordination of the Ge9 cluster by K cations in K4Ge9 (up to 4.0 A ). Type-A Ge9 cluster with 16 closest K neighbours (a).
Type-B Ge9 cluster (disordered) with 13 closest K neighbours (b). Atoms with thermal ellipsoids at 50 % probability level; grey spheres
denote Ge atoms refined with isotropic ADPs.
396 www.zaac.wiley-vch.de 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
S. Ponou, Th. F. Fssler
at the waist (Ge2Ge4, group II) or even 2.75(1) A for
Ge1Ge6 (group III) which is longer than some bond
distances of group II. Also, the ratio of the basal square
diagonals for cluster B, d1/d2 1.10, diverges from the ideal
value of 1.00, indicating quite strong deviation from the
ideal C4 symmetry of the expected monocapped square
antiprism. Nevertheless, although the incomplete orien-
tational ordering leads to the strongly distorted cluster B,
the model of the disorder as three symmetrically related
orientations is fairly satisfactory as the resulting geometry
of the cluster remains close to that observed for similar
clusters in other phases.
For comparison, the GeGe intracluster distances in the
Cs homologue Cs4Ge9 [5] range from 2.52(1) to 3.01(1) A ,
for K4Ge9 9NH3 from 2.547(1) to 2.910(1) A , [16] and, for
the K-crypt salt crystallized from solution from 2.538(5) to
2.962(7) A [17].
The K cations are coordinated by 6 to 11 Ge atoms.
Relatively large ADPs are found for the least coordinated
K4 atom which is surrounded by only 6 Ge atoms. The Ge9
clusters A and B are surrounded by 16 and 13 closest K
neighbours, respectively (up to 4.0 A ), with several of the
cations capping the polyhedron faces (Figure 3). The
KGe interaction distances vary from 3.42(1) to 3.91(1) A
for cluster A, and from 3.11(2) to 3.95(2) A for cluster B.
The shortest intercluster distance is 3.656(5) A between two
clusters ot type A and is shorter than in Cs4Ge9
(3.956(8) A ) probably due to the bigger Cs cation size. The
shortest interaction distance between two K cations is
3.97(1) A.
Regarding only the packing of the Ge9 clusters, the crys-
tal structure of K4Ge9 represents a Cr3Si prototype packing
of Ge9 clusters, with two types of Ge9 clusters in a hier-
archical atom-to-cluster replacement structure in which Cr
atoms are replaced by the type-A and Si atoms by the dis-
ordered type-B Ge9 clusters, respectively. The K cations
occupy 16 of the 24 tetrahedral holes of the pseudo cubic
packing. The hierarchical relation can be written as:
Z. Anorg. Allg. Chem. 2007, 393397
Crystal Growth and Structure Refinement of K4Ge9
24[Cr]3[Si]1 8K16[Ge9]3[Ge9]1 ( represents tetra-
hedral hole). Also Cs4Ge9 and Rb4Ge9 adopt the Cr3Si
hierarchical packing of the clusters, however by lowering
the crystal symmetry which leads to four crystallograph-
ically independent clusters, e.g. the Cr-type positions are no
more symmetrically equivalent [5].
The orientational disorder of the type B Ge9 cluster in
K4Ge9 can be explained by the existence in the crystal
structure of three symmetrically equivalent orientations (re-
lated by a C3 axis) which are randomly adopted by the clus-
ter with the same frequency. The strong distortions of those
clusters may be ascribed to incomplete cluster ordering
which is likely due the weaker number of coordinating K
cations as compared to the type A cluster. Knowing that
many unsuccessful attempts have been made to obtain
single crystals of the K4Ge9 phase, the obtaining of single
crystals that are suitable for diffraction experiments using a
complex starting mixture seems to indicate that this reac-
tion medium is more adequate for the crystal growth of the
title compound.
References
[1] T. F. Fssler, Coord. Chem. Rev. 2001, 215, 377.
[2] J. D. Corbett, Chem. Rev. 1985, 85, 383.
[3] J. D. Corbett, Struct. Bonding 1997, 87, 157.
Z. Anorg. Allg. Chem. 2007, 393397 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[4] T. F. Fssler, S. Hoffmann, Z. Kristallogr. 1999, 214, 722.
[5] V. Que neau, S. C. Sevov, Angew. Chem. 1997, 109, 1818; An-
gew. Chem. Int. Ed. Engl. 1997, 36, 1754.
[6] E. Todorov, S. C. Sevov, Inorg. Chem. 1998, 37, 38893891.
[7] V. Que neau, S. C. Sevov, Inorg. Chem. 1998, 37, 13581360.
[8] S. Bobev, S. C. Sevov, Polyhedron 2002, 21, 641649.
[9] C. Hoch, M. Wendorff, C. Rhr, Acta. Crystallogr. 2002,
C58, i45i46.
[10] V. Que neau, E. Todorov, S. C. Sevov, J. Am. Chem. Soc. 1998,
120, 3263.
[11] W. Carrillo-Cabrera, R. Cardoso Gil, M. Somer, . Persil, H.
G. von Schnering, Z. Anorg. Allg. Chem. 2003, 629, 601608.
[12] C. Hoch, M. Wendorff, C. Rhr, J. Alloys Compd. 2003, 361,
206221.
[13] H. G. von Schnering, M. Baitinger, U. Bolle, W. Carrillo-Cab-
rera, J. Curda, Yu. Grin, F. Heinemann, J. Llanos, K. Peters,
M. Somer, Z. Anorg. Allg. Chem. 1997, 623, 1037.
[14] S. Ponou, Dissertation, T.U. Mnchen, 2006.
[15] (a) SHELXTL; Bruker Analytical X-ray Instruments: Madi-
son, WI, 1998. (b) G. M. Sheldrick, Acta Crystallogr., Sect. A
1990, 46, 467. (c) G. M. Sheldrick, SHELXL-97, Program for
the Solution and Refinement of Crystal structures; University of
Gttingen: Gttingen, Germany, 1997. (d) H. D. Flack, Acta
Crystallogr. 1983, A39, 876881
[16] C. Suchentrunk, J. Daniels, M. Somer, W. Carillo-Cabrera, N.
Korber, Z. Naturforsch. 2005, 60b, 277.
[17] C. H. E. Belin, J. D. Corbett, A. Cisar, J. Am. Chem. Soc.
1977, 99, 7163.
www.zaac.wiley-vch.de 397