A new preparative approach to HgPbP14 structure type materials: Crystal structure of Cu0.73(1)Sn1.27(1)P14 and characterization of M1-xSn1+xP14 (M = Cu, Ag) and AgSbP14

Article abstract of DOI:10.1002/zaac.200500338

A new preparative approach, using main group iodides as mineralization agents, was developed to prepare bulk quantities of HgPbP₁₄ [(M1)(M2)P₁₄] type polyphosphides containing group 11 cations on the M1 and main group elements Sn or formerly unseen Sb on the M2 position. The known (M1)²⁺:(M2)²⁺ combination of cations is extended with the combination (M1)¹⁺:(M2)³⁺ in AgSbP₁₄. A single crystal structure determination was performed for Cu _1-xSn_1+xP₁₄. Cu_0.73(1)Sn _1.27(1)P₁₄ crystallizes orthorhombically, space group Pnma (No. 62) with lattice constants a = 12.513(2) A, b = 9.800(1) A, c = 10.445(1) A, V = 1280.8(3) A³ and Z = 4. Small differences in the cell parameters between the single crystal and powder diffraction experiments of CuSnP₁₄ are probably due to a small homogeneity range. Tetravalent tin postulated beside divalent tin for isostructural Au_0.64Sn_1.36P₁₄ could not be detected by ¹¹⁹Sn-Moessbauer spectroscopic experiments for the copper tin polyphosphide. An ionic description like [(M1⁺) _1-x(M2⁴⁺)]²⁺ [(M2²⁺)(P ⁰)₁₀(P^1-)₄]^2- with x = 0.33 according to a Zintl-Klemm concept has to be substituted by a more covalent description of [(M1⁺)_1-x(M2²⁺) _x]^(1+x)+- [(M2²⁺)(P₁₄)] ^(1+x)- for the copper tin polyphosphide.

Full text of DOI:10.1002/zaac.200500338

Articles
DOI: 10.1002/zaac.200500338
A New Preparative Approach to HgPbP Structure Type Materials: Crystal
1
4
Structure of Cu
Sn
P and Characterization of M Sn P (M ؍
0
.73(1) 1.27(1) 14
1؊x 1؉x 14
Cu, Ag) and AgSbP14
Stefan Lange, C. Peter Sebastian, and Tom Nilges*
Münster/Deutschland, Institut für Anorganische und Analytische Chemie der Universität and NRW Graduate School of Chemistry
Received August 18th, 2005.
Abstract. A new preparative approach, using main group iodides
as mineralization agents, was developed to prepare bulk quantities
of HgPbP14 [(M1)(M2)P14] type polyphosphides containing group
probably due to a small homogeneity range. Tetravalent tin postu-
lated beside divalent tin for isostructural Au0.64Sn1.36 14 could not
be detected by Sn-Mössbauer spectroscopic experiments for the
P
1
19
1
1 cations on the M1 and main group elements Sn or formerly
copper tin polyphosphide.
2ϩ
2ϩ
An ionic description like [(M1 )1Ϫx(M24ϩ ]2ϩ [(M22ϩ)(P0)10(P1Ϫ ]2Ϫ
ϩ
unseen Sb on the M2 position. The known (M1) :(M2) combi-
) )
x
4
1ϩ
3ϩ
nation of cations is extended with the combination (M1) :(M2)
in AgSbP14. A single crystal structure determination was perfor-
med for Cu1ϪxSn1ϩx 14 crystallizes ortho-
rhombically, space group Pnma (No. 62) with lattice constants a ϭ
with x ϭ 0.33 according to a Zintl-Klemm concept has to be substi-
ϩ
2ϩ (1ϩx)ϩ
tuted by a more covalent description of [(M1 )1Ϫx(M2
)
x
]
-
2
ϩ
(1ϩx)Ϫ
P14. Cu0.73(1)Sn1.27(1)
P
[(M2 )(P14)]
for the copper tin polyphosphide.
˚ ˚ ˚ ˚
3
1
2.513(2) A, b ϭ 9.800(1) A, c ϭ 10.445(1) A, V ϭ 1280.8(3) A
and Z ϭ 4. Small differences in the cell parameters between the
Keywords: Covalent polyphosphides; Semiconductors; Mössbauer
single crystal and powder diffraction experiments of CuSnP14 are
spectroscopy
Introduction
tin, a non stoichiometric ratio of (M1):(M2) in combination
with a mixed Au /Sn occupancy on the M1 position was
ϩ
4ϩ
The orthorhombic HgPbP structure type (general formula
1
4
favoured according to an analysis of a single crystal and
1
(
M1)(M2)P ) contains [P ] tubes running along the crys-
ϩ
3ϩ
1
4
ϱ
14
EDX measurements. A vacancy model featuring Au , Au
tallographic b axis. The tubes show a pentagonal cross sec-
tion comparable with the tubes of Hittorf’s phosphorus [1].
In contrast to Hittorf’s phosphorus all tubes are oriented
parallel to each other and the Pb and Sn atoms are incor-
and vacancies on the M1 position was rejected due to struc-
ture chemical considerations. Except Au0.64Sn1.36P14 all
other group 11 materials, prepared from the elements or
from a Sn or Pb flux, were only characterized by EDX
analyses due to very low yields.
One major problem observed for all materials is the ex-
tremely fibrous morphology and the tendency to cleave,
leading to low quality crystals for single crystal structure
determinations and semiconductor applications. This prob-
lem is not solved up to now.
2
Ϫ
porated to the tubes. A [(M2)P ] unit results if the Zintl
1
4
Klemm concept is applied to discuss partial charges. The
parallel tubes are interconnected by the M1 atoms and ad-
ditional PϪP van der Waals interactions.
The HgPbP14 structure type was reported for (M1)PbP14
[
2, 3] and (M1)SnP [4Ϫ6] (M1 ϭ Zn, Cd, Hg, Cu, Ag,
14
ϩ
2
2ϩ
2ϩ
Au). In case of M cations (e.g. Cd and Hg ) single
crystal structure determinations were performed for the
semiconducting materials. An ideal composition of
Herein we report a new preparative approach to HgPbP14
type materials with high yields and improved crystal qual-
ity. Powder diffraction of bulk samples was used for phase
(
M1):(M2):P ϭ 1:1:14 was observed and a general formula
119
analysis and Sn-Mössbauer spectroscopy was applied to
2
ϩ
2ϩ
0
1Ϫ 2Ϫ
of (M1) [(M2) (P ) (P ) ]
was postulated. For
1
0
4
the copper tin polyphosphide in order to verify the realized
M1 ϭ Cu, Ag, Au, the gold polyphosphide is the only
structurally characterized material. From three different
structure models with various oxidation states of gold and
1ϩ
3ϩ
tin oxidation states. AgSbP , a (M1) /(M2) compound
14
with HgPbP14 type structure, is described here for the first
time.
Experimental Section
*
Dr. Tom Nilges
Institut für Anorganische und Analytische Chemie
Corrensstraße 30
D-48149 Münster
Tel.: (ϩ49)-251-83-36645
Fax: (ϩ49)-251-83-36002
Preparation
4
SnI was prepared by mixing tin (12 g, 0.10 mole, Heraeus,
99.999 %) and iodine (40 g, 0.16 mole, Chempur, resublimated) in
toluene (250 mL) [7]. The mixture was refluxed for approx. 30 minutes
until the violet colour of iodine disappeared. The hot solution was
E-mail: nilges@uni-muenster.de
Z. Anorg. Allg. Chem. 2006, 632, 195Ϫ203
© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
195

S. Lange, C. P. Sebastian, T. Nilges
decanted from the remaining tin. Orange SnI
cooling to room temperature. The crude product was recrystallized
from toluene and dried over molecular sieve in vacuum.
4
crystallized after
K X-rays, emitted concurrently by this source. Measurements were
performed at 78 K in the usual transmission setup. 100 mg of
CuSnP14 were placed in a thin-walled PVC container to give a
thickness between 10 and 15 mg Sn cm . Isomer shifts, quadru-
pole interactions and line widths were determined from the experi-
mental data using common fitting routines [10].
Ϫ2
SbI
mony (7 g, 0.057 mol, Chempur 99.9999 %) and iodine (14 g,
.055 mol, Chempur, resublimated) in 300 mL of toluene until the
colour of iodine disappeared [8]. The supernatant solution was fil-
tered from remaining antimony. Orange SbI crystallized upon
3
was prepared by refluxing a mixture of finely ground anti-
0
3
EDX and X-ray powder phase analyses
cooling to room temperature. The crude product was purified by
recrystallization from toluene and dried over molecular sieve in
vacuum.
The experimental data from EDX analyses are in good
agreement to the previously reported results for CuSnP14
but differ for the AgSnP14 phase [6]. All data are summar-
ized in Table 1. Keeping in mind the errors connected with
the semi quantitative method used, one can at least derive
some trends from the analyses. The copper tin polyphos-
phide has a (M1):(M2) ratio less than 1, while for both
silver materials this ratio is higher than 1. The values deter-
mined for the silver and antimony content of AgSbP14 can
be carefully interpreted as equimolar within the expected
errors and an ideal 1:1 composition is probable for
AgSbP . In case of AgSnP we find a more drastic devi-
CuSnP14, AgSnP14 and AgSbP14 samples were prepared from
1
9
9
:1:14 mixtures (starting compositions) of copper (Chempur, shot,
9.999 %) or silver (Chempur, powder, 99.9 %), tin (Heraeus, shot,
9.999 %) or antimony (Chempur, shot, 99.9999 %) and red phos-
phorus (Chempur, pieces, 99.999ϩ %). The total amount was about
g in all cases. 20 mg of SnI or SbI were added as a mineraliz-
ation agent.
1
4
3
All starting materials were sealed in evacuated silica ampoules (in-
ner diameter 8 mm, approximate length 6 cm), placed vertically
into a muffle furnace and heated to 820 K at a heating rate of 3 K
14
14
Ϫ1
ation from the 1:1 ratio with the tendency to a much higher
silver content.
min . The samples were kept at this temperature for 8 days, fol-
lowed by slow cooling to room temperature. The products were
grown as fibres on top of the reaction bulk, which sometimes also
contained binary phases or unreacted starting materials. In order
to produce phase pure samples of the (M1)(M2)P14 the following
procedure was used: Prior to the reaction both (M1) and (M2)
elements were molten together in an evacuated silica ampoule. The
resulting ingot was ground to a fine powder, which was then mixed
with powdered phosphorus and the mineralization agent.
Table 1 Results from EDX analyses. Each value /At-% represents
a set of approx. 8 measurements of different crystals from indepen-
dently prepared samples. At least one set was measured for each
material under discussion. Second set in parenthesis. Data of pre-
viously performed EDX analyses are given in italics for comparison.
CuSnP14
Cu0.73Sn1.27
P
14
AgSnP14
AgSbP14
(M1)(M2)P14
(theoretical)
(theoretical)
4 3
SnI or SbI were found in form of yellow to orange crystals well
separated from the products. Both compounds were removed mech-
anically or by refluxing of the products in toluene.
Cu: 6.3
Sn: 7.0
P: 86.7
Cu: 4.6
Sn: 7.9
P: 87.5
Ag: 8.8 (8.6)
Sn: 7.6 (5.5)
P: 83.7 (86.0) P: 84.0 (86.1) P: 87.5
Ag: 7.9 (7.7)
Sb: 6.7 (6.2)
M1: 6.75
M2: 6.75
EDX analyses
Cu: 6.8
Sn: 8.8
P: 84.4
Lit. [5]
Ag: 9.3
Sn: 11.2
P: 79.5
Lit. [5]
Semi-quantitative elementary analysis was performed with a Leica
420i scanning electron microscope (Fa. Zeiss) fitted with an elec-
tron dispersive detector unit (Fa. Oxford). Cu, Ag, Sn, Sb and GaP
were used as standards for calibration. A voltage of 20 kV was
applied to the samples. A number of approx. 8 measurements for
each set of crystals are averaged to one subset and at least one
subset was done for each material under discussion.
Crystals, grown on top of the bulk residue and bulk
phase (see Figure 1) of all samples were isolated and phase
analysis was performed in order to verify the presence of
the HgPbP14 structure type. Powder patterns of the crystals
and the bulk material of CuSnP14 and AgSbP14 could be
indexed and refined based on the orthorhombic HgPbP14
structure type in space group Pnma (No. 62). All crystallo-
graphic data are summarized in Table 2. No additional
crystalline phases were observed. We found slightly larger
cell parameters for bulk samples of CuSnP14 compared to
the single crystal data which is probably due to a small
homogeneity range of this material.
X-ray power diffraction
A Stoe StadiP powder diffractometer or Guinier cameras both
˚
operated with Cu Kα1 radiation (λ ϭ 1.54051 A) and fitted with a
linear 5° PSD (Fa. Braun) or image plate technology as detection
units were used to collect diffraction data in the 2θ range of 8Ϫ80°.
Silicon (StadiP) and Ͱ-quartz (Guinier) were used as external and
internal standards. Data readout of the image plates was done
using a BAS 1800 reader (Fa. Fuji) and the WinXpow program
package [9] was used for indexing and refinement purposes.
Crystals, grown on top of the bulk material in the case
of AgSnP14 were carefully separated from the bulk phase.
The bulk was identified as Ag SnP [11] and binary silver
3
7
and tin phosphides beside the expected polyphosphide. The
separated crystals were also checked by X-ray powder dif-
fraction in order to verify the isotypy to the copper and
gold polyphosphide (Figure 2). We found some “finger
Mössbauer spectroscopy
A Ca119mSnO
source was available for the 119Sn investigations and
3
a palladium foil of 0.05 mm thickness was applied to reduce the tin
196
zaac.wiley-vch.de
© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2006, 195Ϫ203

A New Preparative Approach to HgPbP14 Structure Type Materials
Figure 1 Scanning electron micrographs and photos of AgSnP14, CuSnP14 and AgSbP14 samples. Top row: AgSnP14 and AgSbP14 crystals
were separated from the bulk phase; CuSnP14 represents bulk material. Bottom row: Photos of the bulk materials.
Table 2 Results from X-ray powder diffraction of HgPbP14 struc-
ture type materials: All isostructural materials crystallize ortho-
rhombic in space group Pnma.
Reflection intensities were collected up to 2θ ϭ 60°. The
space group Pnma (No. 62) was derived from systematic
extinctions. One of the most critical points of this class of
materials, the crystal quality, was checked by precession
plots calculated from the IPDS data [12]. We found sharp
reflections and no hints for any diffuse scattering either due
to an orientation disorder or to a damage of the crystal.
Representative precession plots are presented in Figure 3.
A very small powder content as one might see in Figure 3,
˚
˚
˚
˚
Compound
a / A
b / A
c / A
V / A
Ref.
ZnSnP14
12.619(3) 9.794(2) 10.496(2) 1297(1)
12.817(3) 9.826(2) 10.775(2) 1357(1)
12.760(3) 9.824(2) 10.699(2) 1341(1)
[4]
[4]
[4]
CdSnP14
HgSnP14
Cu1ϪxSn1ϩx
a)
P
14 12.557(2) 9.821(2) 10.478(2) 1291.1(4) This work
a)
Cu0.73Sn1.27
P
14 12.513(2) 9.800(1) 10.445(1) 1280.8(3) This work
a)
Au0.66Sn1.34
P
14 12.595(2) 9.820(2) 10.562(2) 1306(1)
[5]
(0kl layer) is due to some bulk material attached to the
single crystal, which was not removed to avoid crystal dam-
age. Applying the HgPbP14 structure [6] as a starting model,
the structure was refined using the Jana2000 program pack-
age [13]. The copper tin polyphosphide crystallizes ortho-
rhombic, space group Pnma (No. 62), with Z ϭ 4 formula
units in the unit cell and lattice parameters of a ϭ
ZnPbP14
CdPbP14
HgPbP14
12.621(2) 9.811(1) 10.548(2) 1306(1)
12.865(2) 9.850(1) 10.815(2) 1371(1)
12.816(4) 9.852(2) 10.745(3) 1357(1)
[4]
[4]
[4]
AgSbP14
a)
12.816(3) 9.771(2) 10.806(3) 1353.2(9) This work
single crystal data
˚ ˚
3
12.513(2), b ϭ 9.800(1), c ϭ 10.445(1) A, V ϭ 1280.8(3) A .
print” regions around 2θ ϭ 30 and 50°, which can be ob-
served for all compounds under discussion but we were not
able to index all reflections of the diffractogram using the
orthorhombic HgPbP14 structure type. Known binary
phosphides, starting materials or known ternary phases
could not be identified. Whether those extra reflections are
due to a super structure or a completely different structure
type has to be investigated by single crystal structure deter-
mination, which is still in progress.
A numerical absorption correction based on an optimized
crystal shape derived from symmetry equivalent reflections
[14] did not improve the results from structure refinement
and was therefore rejected. The small absorption coefficient
Ϫ1
of μ ϭ 5.39 mm substantiates this finding. After a care-
r
ful analysis of fourier maps we found a scattering power on
the M1 position, which could not be satisfied by a full cop-
per occupancy alone. Due to the high scattering power ob-
served, a mixed tin/copper distribution restricted to a full
occupancy was applied to this position. The resulting small
and isotropic displacement parameter and a check of differ-
ence fourier maps revealed no hints for a split of this site.
We also checked the occupancy factor of the tin split posi-
tion M2 which is supposed to be 0.5 for a “full” tin occu-
pancy. The factor dropped to 0.479(3) after an unrestricted
refinement and increased slightly if a non harmonic descrip-
tion was applied to this position. The R values and residual
Crystal structure determination
A suitable crystal from a CuSnP14 reaction batch was sepa-
rated from the bulk phase with special care to the fact that
CuSnP14 is very sensitive to cleave on touching. Intensity
data were collected on a Stoe IPDS II diffractometer using
˚
MoKα radiation (λ ϭ 0.71073 A, graphite monochromator).
Z. Anorg. Allg. Chem. 2006, 195Ϫ203
© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
zaac.wiley-vch.de
197

S. Lange, C. P. Sebastian, T. Nilges
Figure 2 X-ray powder diffractograms of HgPbP14 type materials. Strong texture effects due to the extreme needle like morphology can
be observed. Diffractograms on the right hand side were calculated from single crystal data of Au0.66Sn1.34
P14 [6], Cu0.73(1)Sn1.27(1)P14 and
HgPbP14 [5] for comparison. “Finger print” regions which are characteristic for all materials are marked by ellipses.
electron densities did not change significantly if the occu-
pancy factor for this position was reset to 0.5. After a har-
monic, anisotropic treatment of all positions the refinement
converged at final R values of R ϭ 0.0456 for reflections
with I>2σ(I) and wR ϭ 0.0704 for all reflections. Taking
the mixed occupied M1 and the M2 split position into ac-
count a composition of Cu0.73(1)Sn1.27(1)P14 results. All rel-
evant crystallographic data are summarized in Tables 3Ϫ6.
troscopy of CuSnP14 material. In general Mössbauer spec-
troscopy is a very useful and reliable tool to determine for-
mal oxidation states. A total 1 Sn isomer shift range of
19
Ϫ1
119m
approx. δ ϭ Ϫ0.4 to 4.1 mm s (Ca
SnO source) has
3
IV
0
been found and the three formal oxidation states Sn , Sn
II
IV
and Sn [15] can be easily separated from each other. Sn
Ϫ1
is observed between Ϫ0.4 to 1.5 mm s
with predomi-
Ϫ1
nantly ionic compounds like fluorides around 0 mm s
and more covalent compounds like SnI4 or SnSe2 at
Further details of the crystal structure investigations are available
from Fachinformationszentrum Karlsruhe, D-76344 Eggenstein-
Leopoldshafen (Germany), on quoting the depository number
CSD 391335.
Ϫ1
1
.5 mm s . The whole range of Sn(0) materials is defined
Ϫ1
by isomer shifts from 1.8 to 2.9 mm s . Binary tin phos-
phides, e.g. Sn P , were reported at 2.9 mm s , respectively.
Predominantly covalent SnO (δ ϭ 2.7 mm s ) defines the
lower end of the isomer shift range of Sn materials while
Ϫ1
4
3
Ϫ1
II
Mössbauer spectroscopy
Ϫ1
ionic materials, e.g. SnCl , are localized at 4.1 mm s
.
2
In order to finally answer the question of a mixed Sn ϩ/Sn4ϩ
2
Mössbauer spectroscopy allows to differentiate between di-
valent and tetravalent tin. Examples are Sn F , Sn F and
1
19
distribution we have performed
Sn Mössbauer spec-
2
6
3 8
198
zaac.wiley-vch.de
© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2006, 195Ϫ203

A New Preparative Approach to HgPbP14 Structure Type Materials
Figure 3 Precession plots of Cu0.73(1)Sn1.27(1)P14 calculated from IPDS data.
14
Table 3 Crystallographic data for Cu0.73(1)Sn1.27(1)P .
Table 4 Atomic coordinates and isotropic displacement para-
˚
2
Compound
(g moleϪ1)
Cu0.73(1)Sn1.27(1)
630.23
P
14
meters /A for Cu0.73(1)Sn1.27(1)P .
14
M
r
Crystal size (mm)
Crystal system
Space group
0.28 x 0.02 x 0.01
orthorhombic
Pnma (No. 62)
12.513(2)
9.800(1)
10.445(1)
1280.8(3)
4
3.267(1)
5.394
1179
0.71073
Atom
Wyckoff-
Position
x
y
z
sof
U
iso
˚
a (A)
1
1
Cu1
Sn1
Sn2
P1
P2
P3
P4
P5
P6
P7
4c
4c
8d
4c
4c
8d
8d
8d
8d
8d
4c
4c
0.3245(1)
0.3245
/
/4
4
0.4986(1)
0.4986
0.73(1)
0.016(1)
0.016
˚
b (A)
0.27
0.5
1
1
1
1
1
1
1
˚
c (A)
0.6773(1)
0.1799(3)
0.4699(3)
0.0795(2)
0.4328(2)
0.3158(2)
0.3566(2)
0.0306(2)
0.4126(2)
0.0583(2)
0.2164(1)
/4
1/
0.4056(1)
0.3511(3)
0.3449(3)
0.4183(2)
0.2108(2)
0.8318(2)
0.6295(2)
0.8522(2)
0.9038(3)
0.7201(3)
0.037(1)
0.020(1)
0.017(1)
0.015(1)
0.014(1)
0.016(1)
0.020(1)
0.014(1)
0.015(1)
0.015(1)
˚
3
1
V (A )
Z
4
Ϫ3
0.0796(2)
0.0835(2)
0.0781(2)
0.0532(2)
ρ
x-ray (g cm
)
μ
x-ray (mmϪ1)
F(000)
˚
λ (A)
0.0780(2)
1
Diffractometer
Refinement method and program
IPDS II
Full matrix least squares on F ,
P8
P9
/
/4
4
1
1
2
1
Jana 2000 [13]
0Ϫ 180, 1.0
Ϫ17 Յ h Յ 17
Ϫ12 Յ k Յ 13
Ϫ14 Յ l Յ 14
12523
999
86
0.1160
ωϪrange (°), increment (°)
hkl range
No. of reflections
Independent reflections
No. of parameters
˚
2
Table 5 Anisotropic displacement parameters /A for
Cu0.73(1)Sn1.27(1)P .
14
R
int
R (I<2σI)
wR (I<2σI)
R (all)
0.0456
0.0677
0.0679
0.0704
1.38
2.37/Ϫ1.48
Atom
U
11
U
22
U
33
U
12
U
13
U
23
Cu1
Sn1
Sn2
P1
P2
P3
P4
P5
P6
P7
0.0167(7)
0.0167(7)
0.0386(7)
0.020(2)
0.021(2)
0.019(1)
0.015(1)
0.016(1)
0.020(1)
0.016(1)
0.019(2)
0.017(2)
0.015(1)
0.015(1)
0.029(2)
0.012(2)
0.012(2)
0.011(1)
0.010(1)
0.010(1)
0.024(1)
0.010(1)
0.011(2)
0.008(2)
0.015(1)
0.015(1)
0.043(1)
0.028(2)
0.017(2)
0.014(1)
0.018(1)
0.022(1)
0
0
Ϫ0.002(1)
Ϫ0.002(1)
Ϫ0.017(1)
0.007(2)
Ϫ0.002(2)
Ϫ0.003(1)
0.001(1)
0
0
wR (all)
GoF
0.004(1)
0
0
0.001(1)
0.001(1)
0.001(1)
Ϫ0.007(1)
0
0
0.001(1)
0.001(1)
0.001(1)
0.000(1)
Ϫ0.001(1)
0
Residual electron density (e AϪ3)
˚
the more covalently bonded Sn S . In case of Sn S two
0.004(1)
2
3
2 3
Ϫ1
0.017(1) Ϫ0.002(1)
Ϫ0.001(1)
Ϫ0.003(1)
0.001(1)
signals were found at isomer shifts of 1.15 mm s
and
0.017(1)
0.014(2)
0.019(2)
0.001(1)
0
0
Ϫ1
P8
P9
3
.5 mm s , clearly separated from each other. All references
0.001(1)
0
1
19
according the Sn isomer shift ranges are given in [15].
Z. Anorg. Allg. Chem. 2006, 195Ϫ203
© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
zaac.wiley-vch.de
199

S. Lange, C. P. Sebastian, T. Nilges
˚
Table 7 119Sn Mössbauer spectroscopic data recorded at 78 K for
14
Table 6 Selected distances /A of Cu0.73(1)Sn1.27(1)P .
Cu1ϪxSn1ϩx
P
14. Isomer shifts δ, electric quadrupole interactions
ΔEQ and experimental line widths Γ are summarized for the two
signals observed for Cu1ϪxSn1ϩx 14. A signal ratio, calculated from
the integrated signal intensities is also given.
Cu1/Sn1ϪP1
Cu1/Sn1ϪP2
Cu1/Sn1ϪP6
Cu1/Sn1ϪP6
2.376(4)
2.427(4)
2.397(2)
2.397(2)
0.658(2)
2.702(4)
2.691(4)
2.701(3)
2.204(3)
2.204(3)
2.200(3)
2.200(3)
2.278(3)
2.219(3)
2.214(3)
2.187(3)
2.207(3)
2.199(3)
2.206(3)
2.236(4)
P
a)
a), g)
Sn2ϪSn2
b)
δ/mm sϪ1
/mm sϪ1
Γ /mm sϪ1
Sn2ϪP1
Sn2ϪP2
Sn2ϪP6
P1ϪP3
signal
ΔE
Q
ratio /%
c)
1
2
2.61(9)
3.14(7)
1.75(7)
1.54(5)
0.91(3)
1.16(1)
29(1)
71(1)
a)
P1ϪP3
P2ϪP4
P2ϪP4
P3ϪP4
P3ϪP5
P4ϪP7
P5ϪP6
P5ϪP8
P6ϪP7
P7ϪP9
P8ϪP9
a)
d)
e)
e)
on M1 and 79 % on M2. All Mössbauer data are summar-
ized in Table 7.
f)
f)
Discussion
Detailed structure descriptions [4Ϫ6] have been given for
(M1)(M2)P14 and no further information can be added to
improve the structural knowledge in general. We therefore
like to answer some open questions left with the group 11
polyphospides.
a)
b)
c)
d)
1
e)
f)
1
Ϫ x, Ϫy, Ϫ1/
2
ϩ z
Ϫ x, Ϫy, 1.5 ϩ z
Occupancy factor 0.5
x, /
2
Ϫ y, z 1
/
2
2
1
1/
/
2
Ϫ x, Ϫy, /
2
ϩ z
g)
1
Ϫ x, Ϫy, 1 Ϫ z
1
1
Ϫ /
2
Ϫ x, Ϫy, /
2
ϩ z
Structure chemical aspects related to the bridging
cations in HgPbP14 type materials
Figure 5 gives an overview of the crystal structure under
discussion featuring parallel [(M2)P ] strands, connected
14
via [3ϩ1] coordinated M1 positions. The SnϪP bond dis-
˚
tance distribution of 2.376(4), 2 x 2.397(2) and 2.427(4) A
reflect the second order Jahn Teller distortion [16] for the
ϩ
10
majority component Cu (d ion) and can also be inter-
preted in terms of a lone pair effect of the minor compo-
nent Sn2 . Sn should accommodate a more “symmetri-
ϩ
4ϩ
cal” coordination inside the P unit around M1 due to the
4
lack of a lone pair. This type of coordination chemistry
Figure 4 Experimental (points) and fitted (lines) 119Sn Mössbauer
spectra of Cu1ϪxSn1ϩx 14 at 78 K. Two independent signals repre-
sented by solid lines were fitted to the Mössbauer curve showing
P
Ϫ1
isomer shifts of 2.61(9) and 3.14(7) mm s
.
For CuSnP14 two quadrupolar splitted signals with an
Ϫ1
isomer shift of δ ϭ 2.61(9) and δ ϭ 3.14(7) mm s have
1
2
been observed indicating the presence of two different tin
positions in a non spherical environment (Figure 4). A
covalent nature of the copper tin polyphosphide can be
derived from the value of the isomer shifts located at the
II
low end of the isomer shift range of Sn materials. The non
Ϫ1
existence of a signal within the range of 0 Ϫ 1.5 mm s
clearly proves the absence of tetravalent tin for the copper
tin polyphosphide. The signal ratio of δ :δ ϭ 29:71 % ag-
1
2
rees quite well, within the expected errors, with the tin dis-
tribution on the two cation positions derived from structure
Figure 5 Crystal structure section of Cu0.73(1)Sn1.27(1)P14. View
analysis. The δ signal is therefore assigned to tin localized
1
parallel the b-axis. Copper grey spheres, tin black spheres and phos-
phorus white spheres. Grey polyhedra centred by copper represen-
ting the connection between the parallel orientated [SnP14] units.
on the M1 and the δ2 signal to the M2 position. For
(Cu0.73(1),Sn0.27(1))SnP14 one can calculate a tin ratio of 21
200
zaac.wiley-vch.de
© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2006, 195Ϫ203

A New Preparative Approach to HgPbP14 Structure Type Materials
can be observed for the tin(II),tin(IV) sulfide Sn S [17]. We
the next neighbours, that are four phosphide positions, is
2
3
˚
found no experimental evidence for such a symmetrically
coordinated site in Cu1ϪxSn1ϩxP14 (see crystal structure de-
termination section).
optimised to end up at approx. 2.7 A, within the usual
˚
SnϪP bond distance range of 2.5Ϫ2.75 A [18Ϫ24]. It can
be concluded that y is responsible for the split and if y ex-
tends to a certain value, the M2 split position is generated
to optimize the bonding situation for the M2 cation. In the
˚
A M2 split position with a distance d ϭ 0.658(2) A was
˚
observed for Cu0.73(1)Sn1.27(1)P . Compared with 0.60 A for
14
˚
Au0.64Sn1.36P14 the split can be correlated to the distance y
case of M2 ϭ Sn for HgSnP14 a y distance of 5.75 A is
(
see Figure 6), defined by the bridging phosphide positions
short enough to result in attractive interactions between Sn
and the bridging P atoms (d(SnϪP) ϭ 2.94 A).
˚
of the P14 tube, surrounding the M2 position. In case of the
gold and copper tin polyphosphide the M2ϪP distance to
Chemical bonding in HgPbP14 type materials
Recent
M1)1Ϫx(M2)1ϩxP14 (M1 ϭ Cu, Ag, Au, M2 ϭ Sn, Pb)
polyphosphides left one unanswered question concerning a
results
dealing
with
nonstoichiometric
(
ϩ
3ϩ
2ϩ
4ϩ
mixed contribution of Au /Au
and/or Sn /Sn
in
Au1ϪxSn1ϩxP14 [6]. This mixed contribution is necessary to
allow charge balance, if the Zintl-Klemm concept is applied
to discuss the chemical bonding for this material.
In case of Cu0.73(1)Sn1.27(1)P14, which appeared to be iso-
structural to the gold compound, a much higher scattering
power than expected for a full copper occupancy was found
on the M1 position. From the crystal chemical point of
ϩ
2ϩ
2ϩ
2ϩ
ϩ
view one has to consider a Cu /Sn , Cu /Sn , Cu /
3ϩ
2ϩ
ϩ
4ϩ
Cu /Sn or a Cu /Sn mixing. If the chemical bonding
is discussed on the basis of the Zintl-Klemm concept, as
done so in the case of Au0.64Sn1.36P14, an overall charge of
2
ϩ must be considered for the mixed occupied M1 position.
2ϩ
The occurrence of Cu in tetrahedral coordination is very
unlikely and the instability of trivalent copper definitely
points towards a Cu /Sn or Cu /Sn mixing. Focusing
carefully, and with respect to the accuracy of single crystal
X-ray diffraction methods, on the composition of
ϩ
4ϩ
ϩ
2ϩ
Cu0.73(1)Sn1.27(1)P , that value is rather close to the ideal
14
ϩ
4ϩ 2ϩ
case of Cu0.66Sn1.33P14, caused by a [(Cu0.66) ,(Sn0.33)
]
mixing on the M1 position. It has to be stated at this stage,
that the low copper content compared with tin is in good
agreement with the results from EDX analyses and also in
agreement with both mixing models. Therefore, a clear de-
cision between both models can only be made by additional
experiments like Mössbauer spectroscopy.
Two signals, assigned to two different divalent tin posi-
tions were determined by Mössbauer spectroscopy.
As a consequence, the postulated model with a mixed
ϩ
4ϩ
Cu /Sn occupancy on the M1 position must therefore
be rejected. The absence of tetravalent tin requires a more
covalent description and the classical view of a Zintl-
Klemm type description for Cu1ϪxSn1ϩxP14 like
ϩ
4ϩ 2ϩ
2ϩ
0
1Ϫ 2Ϫ
[
(M1) ,(M2)
]
[(M2) (P ) (P ) ] has to be substi-
10 4
tuted by
[
a
more covalent one leading to
(M1 )1Ϫx,(M2 ) ](1ϩx)ϩ[(M22ϩ)(P )]
ϩ
2ϩ
(1ϩx)Ϫ
. This coval-
x
14
ent description is in good agreement to the semiconducting
behaviour expected for this class of materials. Bond chemi-
cal calculations and electrochemical experiments are
planned to verify the actual charges and the semiconduct-
ing properties.
Figure 6 View perpendicular to the b axis in various (M1)(M2)P14
type materials. Distances /A and angles /°. The b axis can be
separated into two distances x and y, defined between the “roof”
P-positions, bridged by the cation positions (M1 ϭ (Cu,Sn),
˚
(Au,Sn), Hg) and (M2 ϭ Sn).
Z. Anorg. Allg. Chem. 2006, 195Ϫ203
© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
zaac.wiley-vch.de
201

S. Lange, C. P. Sebastian, T. Nilges
Structure considerations for AgSbP14
found for all HgPbP14 type tin polyphosphides. This fact
agrees well with the short b-axis observed and the estimated
Antimony phosphorus materials with SbϪP bonds showing
a covalent-ionic nature are very rare in the field of purely
inorganic compounds. Very recently some ab initio calcu-
lations have been reported on nanolaminated M SbP (M ϭ
Ti, Zr, Hf), where MP layers are interleaved by Sb [25].
˚
y distance of 5.7 to 5.8 A postulated for AgSbP .
14
Whether AgSbP14 can be roughly described in terms of
the Zintl-Klemm concept like the stoichiometric HgPbP14
2
materials, featuring trivalent antimony, or if it has to be
discussed in a more covalent manner, like the copper tin
polyphosphide, has to be clarified after single crystal struc-
ture determination, Mössbauer spectroscopy and bond
chemical calculations.
Barely characterized SbP and BiSbP , prepared by a solu-
2
tion precipitation route using trimethylsilylphosphine and
antimony or bismuth halide were also reported, but struc-
tures could not be discussed due to the lack of crystalline
material [26]. Focussing on the metal-organic sector some
organo-metal cage compounds with attractive SbϪP bond
Conclusion
˚
interaction within the range of 2.49Ϫ2.55 A have been re-
HgPbP14 type materials have been subject of intense scien-
tific research for more than 50 years. Starting with M1 ϭ
Zn, Hg, Cd and M2 ϭ Sn, Pb interesting semiconducting
properties were found. Followed by the preparation of non-
stoichiometric polyphosphides with M1 ϭ Cu, Ag and Au
and M2 ϭ Sn, Pb some questions concerning the mixed
valence of cations and the crystal structure arose. Today
those questions are answered due to improved synthesis
procedures and additional Mössbauer spectroscopic experi-
ments.
The HgPbP14 structure type was found for the copper tin
and the silver antimony polyphosphide under discussion by
X-ray powder diffraction phase analysis and the crystal
structure was determined for the copper tin polyphosphide
from single crystal data. Tetravalent tin postulated for the
group 11 tin polyphosphides could not be observed in case
of the copper tin polyphosphide and the bond chemical de-
scription according the Zintl-Klemm concept was replaced
by a more covalent description. The first silver antimony
polyphosphide continues the ongoing scientific progress in
this class of semiconducting materials, representing one
very rare example of a covalent SbϪP interaction in a
purely inorganic environment. AgSbP14 is a promising can-
didate to offer a completely new insight in the structure
chemistry of pnictogens involving both antimony and phos-
phorus. So far, to our knowledge, SbϪP interactions
characterized by crystal structure determination are re-
ported for metalorganic materials only.
HgPbP14 type compounds have some characteristic struc-
ture chemical aspects to be highlighted: The bridging phos-
phide ions at the so called “roof” positions within the tubes
are the most flexible part of the present polyphosphide ma-
terials and are therefore responsible for the chemical varia-
bility of these materials. In contrast, the non bridging part
of the polyphosphide tube is rigid and the bond distances
are not significantly influenced by the type of cations in-
volved in the bridging process and the 3D stacking of the
tubes. Those aspects are supposed to be main features of
this class of materials. Whether the flexibility will be high
enough to introduce other than the so far used cations is
one challenging aspect of our future work.
ported [27, 28]. Those results are based on single crystal
structure determinations and therefore deliver the only re-
liable structure information for materials with SbϪP inter-
actions.
Unfortunately, we were not able to find suitable crystals
of AgSbP14 to perform single crystal structure determi-
nation so far. After an optimized adaptation of our syn-
thetic route to the special necessities of this new class of
materials we will report this in the early future. Neverthe-
less, from a combination of all present results, it is possible
to predict some of the structural features of AgSbP . EDX
1
4
analysis points towards a composition close to an equimo-
lar cation ratio and AgSbP14 is isostructural to the other
HgPbP14 type compounds (see Figure 2). Trivalent anti-
mony, featuring a lone pair cation, definitely fits to the M2
position in so far, that this position is usually occupied by
2
ϩ
2ϩ
other lone pair cations (Sn and Pb ).
˚
For AgSbP14 an axis length of b ϭ 9.771(2) A has been
derived from X-ray powder diffraction experiments re-
presenting the shortest b-axis compared with all other
known HgPbP14 type materials (see Table 2). The trivalent
antimony cation, substituting a divalent cation on the M2
position, is expected to shorten the PϪP distance of the
adjacent bridging phosphide positions. As one might see in
Figure 6 the b axis is determined by the P14 tubes. The type
of the M1 and M2 cation is also responsible for the elong-
ation of this axis or, to be precise, for the distances x (M1)
and y (M2) between the bridging phosphide ions, bonded
to the cations. Therefore the combination of x and y has to
be taken into account to estimate the length of the b-axis.
In Cu0.73(1)Sn1.27(1)P14 as well as Au0.64Sn1.36P14 and
HgSnP14 the PϪP distance x, determined by the bridging
M1 cation, is dependent on the M1ϪP distance and inde-
pendent on the PϪM1ϪP angle, which is approx. 107° in
all three cases. Variation of the M1 charge obviously does
not change the bond angle of this bridge. Calculating with
˚
an AgϪP distance of approximately 2.50 Ϫ 2.55 A averaged
from selected silver phosphides [29Ϫ33] and assuming a
comparable PϪM1ϪP bond angle for AgSbP14 the respect-
˚
ive PϪP distance will be located in the range of 4.0 to 4.1 A.
This gives one the possibility to estimate a y distance of 5.7
˚
to 5.8 A comparable or slightly shorter than for HgSnP .
14
Acknowledgement. C.P.S thanks the NRW Graduate School of
Chemistry for a PhD stipend. We are indebted to H. J. Göcke and
F. Schappacher for the work at the scanning electron microscope.
˚
Assuming a SbϪP bond distance of 2.5 A for AgSbP , this
distance is shorter than the optimal SnϪP distance of 2.7 A
14
˚
202
zaac.wiley-vch.de
© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2006, 195Ϫ203

A New Preparative Approach to HgPbP14 Structure Type Materials
References
[15] P. E. Lippens, Phys. Rev. B. 1999, 60, 4576 and references
herein.
[1] H. Krebs, I. Pakulla, G. Zürn, Z. Anorg. Allg. Chem. 1955,
[
[
16] J. K. Burdett, O. Eisenstein, Inorg. Chem. 1992, 31, 1758.
17] R. Kniep, D. Mootz, U. Severin, H. Wunderlich, Acta Crys-
tallogr. 1982, B38, 2022.
278, 274.
[
[
2] H. Krebs, Th. Ludwig, Z. Anorg. Allg. Chem. 1958, 294, 257.
3] K. E. Pachali, W. Ott, H. Thurn, Acta Crystallogr. 1975, A31
Suppl., 69.
[
[
[
[
18] O. Olofsson, Acta Chem. Scand. 1970, 24, 1153.
19] P. C. Donohue, Inorg. Chem. 1970, 9, 335.
20] J. Gullman, O. Olofsson, J. Solid State Chem. 1972, 5, 441.
21] J. Y. Pivan, R. Guerin, J. Padiou, M. Sergent, J. Solid State
Chem. 1988, 76, 26.
[
[
4] U. D. Scholz, W. Jeitschko, J. Solid State Chem. 1987, 67, 271.
5] M. Eschen, J. Wallinda, W. Jeitschko, Z. Anorg. Allg. Chem.
2002, 628, 2764.
[6] H. Thurn, H. Krebs, Acta Crystallogr. 1969, B25, 125.
[
[
[
22] J. Gullman, J. Solid State Chem. 1990, 87, 202.
23] W. Hönle, H. G. von Schnering, Z. Kristallogr. 1980, 153, 339.
24] M. M. Shatruk, K. A. Kovnir, A. V. Shevelkov, I. A.
Presnyakov, B. A. Popovkin, Inorg. Chem. 1999, 38, 3455.
25] D. Music, Z. Sun, J. M. Schneider, Phys. Rev. B 2006, 71,
[7] G. Brauer, Handbuch der Präparativen Anorganischen Chemie,
3
rd. ed., Vol. 2, F. Enke, Stuttgart 1978, p. 759.
8] G. Brauer, Handbuch der Präparativen Anorganischen Chemie,
rd. ed., Vol. 1, F. Enke, Stuttgart 1975, p. 591.
[
[
3
[
[
9] Stoe and Cie GmbH, WinXPow, Ver. 1.08, Darmstadt (Ger-
092102.
many), 2000.
26] G. C. Allen, C. J. Carmalt, A. H. Cowley, A. L. Hector, S.
Kamepalli, Y. G. Lawson, N. C. Norman, I. P. Parkin, L. K.
Pickard, Chem. Mater. 1997, 9, 1385.
[
10] Wissenschaftliche Elektronik GmbH, Wissoft 2003, Ver.
.0.1.14; R. A. Brand, Normos Moessbauer fitting program,
Wissenschaftliche Elektronik GmbH, Starnberg (Germany)
004.
1
[27] D. E. Hibbs, M. B. Hursthouse, C. Jones, R. C. Thomas, J.
2
[11] M. M. Shatruk, K. A. Kovnir, A. V. Shevelkov, B. A. Popov-
Chem. Soc., Dalton Trans. 1999, 2627.
kin, Angew. Chem. 2000, 112, 2561; Angew. Chem. Int. Ed.
[28] S. J. Black, M. D. Francis, C. Jones, Chem. Commun., 1997,
305.
2000, 39, 2508.
[
12] Stoe & Cie GmbH, X-area program package, Ver. 1.25, Darm-
[29] M. H. Möller, W. Jeitschko, Inorg. Chem. 1981, 20, 833.
[30] M. H. Möller, W. Jeitschko, Z. Anorg. Allg. Chem. 1982, 491,
225.
[31] A. Mewis, Z. Naturforsch. 1979, 34b, 1373.
[32] C. Tomuschat, H. U. Schuster, Z. Naturforsch. 1981, 26b, 1193.
[33] R. O. Demchyna, H. Borrmann, S. I. Chykhrii, Yu. B.
Kuz’ma, Yu. N. Grin, Z. Kristallogr. Ϫ New Cryst. Struct.
2002, 217, 161.
stadt (Germany), 2004.
[13] V. Petricek, M. Dusek, The crystallographic computing system
JANA 2000, Institute of Physics, Praha (Czech Republic),
2000.
[14] Stoe and Cie GmbH, XRed Ver. 1.10 and XShape Ver. 2.05,
Programmes for numerical absorption correction, Darmstadt
(
Germany), 2004.
Z. Anorg. Allg. Chem. 2006, 195Ϫ203
© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
zaac.wiley-vch.de
203