The organometallic double metal cyanide [(Me2Sn)3{Co(CN)6}2·6H 2O]. A three-dimensional framework of infinite, stapled ribbons

Article abstract of DOI:10.1016/S0022-328X(00)00198-4

The admixture of aqueous Me₂SnCl₂-Me₃SnCl solutions to solutions of K₃[Co(CN)₆] in the molar ratios 3:0:2 and 1:1:1, respectively, affords precipitates of [(Me₂Sn)₃{Co(CN)₆}₂·6H ₂O] (1) and [(Me₂Sn)(Me₃Sn)Co(CN)₆] (2). Another synthesis of 2 is based upon the bromination of [(Me₃Sn)₃Co(CN)₆] in MeOH. The crystal structure analysis of 1 reveals stacks of infinite planar ribbons built up of trans-Me₂Sn(μ-NC)₄ and trans-Me₂Sn(μ-NC)₂(OH₂)₂ units as well as of trans-Co(CN)₂ fragments with terminal CN-ligands. Sn←OH₂(NC)₂ hydrogen bonds interlink adjacent ribbons. Multinuclear (¹³C, ¹⁵N, ⁵⁹Co, ¹¹⁹Sn) CPMAS solid-state NMR studies of 1 suggest some intra-ribbon dynamics, and help in proposing both a plausible architecture for anhydrous 2 and a better understanding of the nature of another precipitate 5 resulting from the admixture of a Me₂SnCl₂/Me₃SnCl solution to a solution of K₄[Fe(CN)₆] (1:2:1).

Full text of DOI:10.1016/S0022-328X(00)00198-4

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 604 (2000) 34–42
The organometallic double metal cyanide
[(Me2Sn)3{Co(CN)6}2·6H2O]. A three-dimensional framework of
infinite, stapled ribbons
Eric Siebel a, R. Dieter Fischer a,*1, Nicola A. Davies b, David C. Apperley b,
Robin K. Harris b,*2
a Institut fu¨r Anorganische und Angewandte Chemie der Uni6ersita¨t Hamburg, Martin-Luther-King-Platz 6, 20146 Hamburg, Germany
b Department of Chemistry, Uni6ersity of Durham, South Road, Durham DH1 3LE, UK
Received 24 November 1999; accepted 7 March 2000
Abstract
The admixture of aqueous Me2SnCl2–Me3SnCl solutions to solutions of K3[Co(CN)6] in the molar ratios 3:0:2 and 1:1:1,
respectively, affords precipitates of [(Me2Sn)3{Co(CN)6}2·6H2O] (1) and [(Me2Sn)(Me3Sn)Co(CN)6] (2). Another synthesis of 2 is
based upon the bromination of [(Me3Sn)3Co(CN)6] in MeOH. The crystal structure analysis of 1 reveals stacks of infinite planar
ribbons built up of trans-Me2Sn(m-NC)4 and trans-Me2Sn(m-NC)2(OH2)2 units as well as of trans-Co(CN)2 fragments with
terminal CN-ligands. Sn’OH2···(NC)2 hydrogen bonds interlink adjacent ribbons. Multinuclear (13C, 15N, 59Co, 119Sn) CPMAS
solid-state NMR studies of 1 suggest some intra-ribbon dynamics, and help in proposing both a plausible architecture for
anhydrous 2 and a better understanding of the nature of another precipitate 5 resulting from the admixture of a Me2SnCl2/
Me3SnCl solution to a solution of K4[Fe(CN)6] (1:2:1). © 2000 Elsevier Science S.A. All rights reserved.
Keywords: Double metal cyanide; Three-dimensional framework; Stapled ribbons
1. Introduction
(CN)6] (2). While the {Me3Sn}+ ion usually
adds two more NC-ligands to achieve trigonal
bipyramidal (tbp) configuration, the {Me2Sn}2+ ion is
known to prefer coordinative saturation in affording
a pseudo-octahedral {Me2Sn(NC)4} unit [4]. In analogy
to the well-known super-Prussian blue derivative
[(Me3Sn)3Co(CN)6] (3), wherein each of the six cyanide
nitrogen atoms occupies one axial coordination site of
a Me3Sn fragment [5], four cyanide N atoms of
usually different Co(CN)6 units are required to satisfy
the four still available sites of each Me2Sn fragment.
As an instructive oxygen-bridged example, the co-
ordination polymer [Me2SnMoO4] is known to span
a 3D framework of SnMe2O4 quasi-octahedra and
MoO4 tetrahedra [6]. In the following, we present the
crystal structure of 1 and demonstrate spectroscopically
For more than 25 years, double metal cyanide com-
plexes, reminiscent in their composition of Prussian
blue, have been known to catalyze the polymerization
of epoxides to high molecular weight polyether polyols
[1]. In particular, zinc hexacyanocobaltate(III),
[Zn3{Co(CN)6}2], in the presence of ZnCl2, diglyme and
water, leads to polymers with narrow molecular weight
distributions and exceptionally low levels of unsatura-
tion [2]. According to the current patent literature, the
search for continuously more efficient, double metal
cyanide catalysts is ongoing [3]. As two new
organometallic Prussian blue derivatives, and likewise
potential epoxide polymerization catalysts, we describe
here the supramolecular assemblies [(Me2SnIV)3-
{CoIII(CN)6}2·6H2O] (1) and [(Me2SnIV)(Me3SnIV)CoIII-
that
2
is actually
a
novel ‘alloyed’ polymer
(Me2Sn:Me3Sn), and not a mixture of 1 and 3. More-
over, another precipitate potentially formulated as the
1 *Corresponding author. Tel.: +49-40-4283831001; fax: +49-40-
likewise
‘alloyed’
new
coordination
polymer:
428382882.
2 *Corresponding author.
[(Me2Sn)(Me3Sn)2Fe(CN)6·3H2O] (5) will be described
0022-328X/00/$ - see front matter © 2000 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(00)00198-4

E. Siebel et al. / Journal of Organometallic Chemistry 604 (2000) 34–42
35
and critically inspected by solid-state magnetic reso-
nance spectroscopy.
solutions of Me2SnCl2, Me3SnCl and K3[Co(CN)6]
(1:1:1). During alternative attempts to prepare the
brominated homologue of 4 following the well estab-
lished bromination route of R3SnR% systems [7]:
2. Preparation of 1, 2 and 5
[(Me3Sn)3Co(CN)6]+3Br2
Immediately on addition of an aqueous solution of
Me2SnCl2 to an aqueous solution of K3[Co(CN)6] in the
molar ratio 3:2, a voluminous white precipitate is
formed (Eq. (1)):
“[(Me2SnBr)3Co(CN)6]¡+3MeBr
(2)
we also arrived exclusively at 2 and unreacted Br2.
Actually, 2 is obtained stoichiometrically in methanolic
solution according to Eq. (3):
3Me2SnCl2+2K3[Co(CN)6]
[(Me3Sn)3Co(CN)6]
(MeOH)
“[(Me2Sn)3{Co(CN)6}2·xH2O]¡
(1)
+Br2 ꢀꢀꢀꢀꢀꢀꢁ [(Me2Sn)(Me3Sn)Co(CN)6]
The elemental analysis of the white powder obtained
after filtration, washing (H2O, Et2O) and drying is more
consistent with the presence of two H2O molecules per
formula unit (x=2) than with x=6, as actually indi-
cated by the crystallographic results (vide infra). From
the initially clear filtrate, transparent, plate-like single
crystals suitable for crystallographic studies were de-
posited after a few days. Interestingly, the reaction of
Me2SnCl2 and K3[Co(CN)6] in the stoichiometric ratio
3:1 (instead of 3:2, see Eq. (1)) does not afford the
potential 3D-polymer [(Me2SnCl)3Co(CN)6] (4),
wherein each tin atom would suitably be ‘functional-
ized’ for various substitution reactions.
+Me3SnBr+MeBr
(3)
The potential quasi-homologue [(Me2Sn(Me3Sn)2Fe-
(CN)6·3H2O] (5) of 2 was prepared from Me2SnCl2,
Me3SnCl and K4[Fe(CN)6] (1:2:1) by co-precipitation in
H2O. Spectroscopic results (vide infra) indicate that
here the formation of [(Me3Sn)4Fe(CN)6] [8] is not
strictly circumvented. In principle, the latter polymer
and
a
new species to be formulated as
[Me2Sn)2Fe(CN)6·xH2O] could be expected instead of
5. Interestingly, in the absence of Me3SnCl (2:1
Me2SnCl2–K4[Fe(CN)6]), only a gel-like precipitate los-
ing about 30% of its initial weight on drying was
obtained, the filtration of which turned out to be
extremely difficult. A similar ‘cyanogel’ is reported to
result from SnCl4 and K4[Fe(CN)6] [9].
Analytically pure, and practically anhydrous,
[(Me2Sn)(Me3Sn)Co(CN)6] (2) was obtained, in strict
analogy to the preparation of 1 (Eq. (1)) from aqueous
Table 1
Crystal data and structure refinement parameters of 1
3. Crystal structure of 1
Empirical formula
Formula weight (g mol−1
Temperature (K)
Crystal system
Space group
a (pm)
C9H12N6O3CoSn1.5
489.12
153(2)
Orthorhombic
Pbmn (No. 62)
1038.0(0)
2419.2(9)
1485.6(4)
3731(2)
8
The low-temperature single-crystal structure analysis
of 1 confirms the presence of four well-positioned, and
two strongly disordered, interstitial water molecules per
formula unit. Crystal data and structure refinement
parameters are given in Table 1, and the basic structure
of this coordination polymer is shown in Fig. 1. As
expected, all tin atoms are quasi-octahedrally coordi-
nated, carrying their two methyl groups throughout in
trans-orientation. While there are three different
{Me2Sn} building blocks in the asymmetric unit, just
one {Co(CN)6} entity is present. Only one of the three
tin atoms adopts the initially expected Me2Sn(NC)4
pattern, while the composition of the other two
{Me2SnX4} units is {SnMe2(NC)2(OH2)2}. As four of
the six water molecules are coordinated to tin atoms,
two (trans-oriented) cyanide ligands of each Co atom
must remain terminal. The basic structure elements of 1
are about 1.5 nm broad (i.e. from Sn1 to Sn2 via Sn3)
and ca. 0.6 nm thick (i.e. from N3 to N5) ribbons of
infinite length. While the surfaces of these ribbons are
expected to be more hydrophobic owing to the specific
orientation of the methyl groups, their edges should be
)
b (pm)
c (pm)
V (106 pm3)
Z
Dcalc (g cm−3
)
1.742
Diffractometer/wavelength
Hilger & Watts Y290,
Mo–Ka
28.95
Absorption coefficient (cm−1
F(000)
[-Range (°)
)
1872
2.39B[B25.06
−15h512, −15k528,
−15l517
4349
Index range (°)
Reflections collected
Independent reflections
Absorption correction (DIFABS
Data/restraints/parameters
R indices (all data) R1/wR2
3451
)
Tmin/Tmax: 0.324/0.754
3351/48/213
0.0909/0.2119
0.0572/0.1487
1.765 and −1.418
Final R indices (I\2|) R1/wR2
Largest diff. peak and hole
(e×10−6 pm−3
)

36
E. Siebel et al. / Journal of Organometallic Chemistry 604 (2000) 34–42
to each other extending exclusively along the c axis, in
that between two equivalent staples of ribbons in alter-
nate positions (i.e. staples A and B) each of the termi-
nal cyanide ligands seems to be involved in two
OꢀH···NC bridges. Eight distinct OꢀH···N hydrogen
bonds are thus active per formula unit. The resulting
3D framework of 1 is comparatively compact, although
along c infinite, straight channels of a quasi-rectangular
cross section (ca. 0.6×0.6 nm) extend, which accom-
modate the two strongly disordered H2O molecules.
The SnꢀN and SnꢀO distances of 1 (Table 2) com-
pare well with those found in earlier described coordi-
nation polymers containing Me3Sn building blocks [10].
The CꢀNꢀSn angles adopt throughout values close to
180°, while the C(Me)ꢀSnꢀC(Me) angles of the virtually
trans-oriented methyl ligands deviate significantly from
180°. Even the various CꢀCoꢀC and CoꢀCꢀN angles of
the {Co(CN)6} unit (not listed in Table 2) reflect nota-
ble deviations from the ideal octahedral values. The
C(Me)ꢀSn distances vary between 206 and 218 pm,
being on average slightly shorter than C(Me)ꢀSn dis-
tances in Me3Sn building blocks.
Fig. 1. View on the fragment of one [(Me2Sn){Me2Sn(H2O)2}2-
{Co(CN)6}2] ribbon of 1 (extending along c), including the atomic
numbering scheme. Hydrogen atoms have been omitted. a=
Sn3···Sn3%: 763.3(3) pm, b=Co1···Co1%: 742.8(3) pm; c=Sn1···Sn3:
757.0(3) pm, d=Co1···Co1%: 745.0(3) pm.
The O···N distances of the OꢀH···NC hydrogen
bonds of 1 are only slightly longer than those reported
recently for an exemplary Me3Sn-containing assembly
[10]. The SnꢀOꢀN angles listed in Table 2 deviate only
weakly from 109° stating that the OꢀH···N fragments
should be almost collinear. ‘Unconventional’ CꢀH···N
hydrogen involving the tin-bonded methyl groups can
be ruled out, as all C···N distances are longer than 375
pm.
Fig. 2. Supramolecular architecture of 1: view along c, depicting in
the a/b plane the traces of five equivalent ribbons. Inter-ribbon
OꢀH···NC hydrogen bonds (see the text) are indicated by faint
straight lines. Methyl groups (bonded to Sn atoms) have been omit-
ted.
4. IR/Raman-spectroscopic and XRD-properties of 1, 2
and 5
Table 2
Selected interatomic distances (pm) and bond angles (°) of 1 a
The frequencies of some characteristic infrared (IR)
and Raman (Ra) active vibrations of 1, 2 (prepared
according to the two different routes) and 5 are listed in
Table 3. It is difficult to deduce from the w(CN) band
positions that 1 contains also terminal cyanide ligands.
For 2, the appearance of five w(CN) bands in total,
without any pairwise (IR/Ra)-coincidence, would sug-
gest the presence of {Co(CN)6} fragments of local D4h
symmetry. Although at least two Me2Sn fragments of 1
have notably different environments, no more than
three bands (IR+Ra) can be assigned to w(SnC) vibra-
tions. In contrast, 2 displays two intense w(SnC)s bands.
It is also noted that the spectra of 2 depend slightly on
the mode of the preparation. In the case of 5, the
comparatively larger number of Raman active w(CN)
bands is surprising. Interestingly, the host/guest system
[(nBu4N)0.5(Me3Sn)3.5Fe(CN)6·H2O] [11] was found to
behave similarly in displaying between 2026 and 3130
cm−1 even six Ra-active bands (and only four IR active
Sn1ꢀN1
Sn2ꢀN2
Sn3ꢀN4
Sn3ꢀN6
226.3(9)
221.6(9)
230.1(10)
225.5(13)
C1ꢀN1ꢀSn1
C2ꢀN2ꢀSn2
C4ꢀN4ꢀSn3
C6ꢀN6ꢀSn3
172.8(8)
175.8(8)
175.8(10)
175.1(10)
Sn1ꢀO2
Sn2ꢀO1
231.4(8)
236.1(8)
C11ꢀSn1ꢀC12
C21ꢀSn2ꢀC22
C31ꢀSn3ꢀC32
164.0(7)
168.5(6)
172.1(10)
O1···N3
O1···N5
O2···N3
O2···N5
278.3(14)
290.7(14)
286.1(14)
285.5(14)
Sn1···O2···N3
Sn1···O2···N5
Sn2···O1···N3
Sn2···O1···N5
N3···O1···N5
N5···O2···N3
113.5(4)
116.4(4)
115.3(4)
113.3(4)
98.3(4)
97.8(4)
a Dotted lines refer to OꢀH···N hydrogen bonds.
hydrophilic because of a considerable crowding of H2O
ligands.
The supramolecular architecture of 1 in total is de-
picted in Fig. 2. All of the ribbons are arranged parallel

E. Siebel et al. / Journal of Organometallic Chemistry 604 (2000) 34–42
37
Table 3
Selected infrared and Raman bands (w in cm−1) of 1, 2 and 5
a
Vibration
1
2
2 b
5
IR
2154
Ra
IR
Ra
IR
Ra
IR
Ra
w(CN)
2159
2172
2150
2166
2181
2201
2206
2151
2166
2172
2184
2201
2044
2065
2067
2092
2111
2136
2146
2181
2196
2190
was(SnC) c
ws(SnC) c
was(SnC) d
w(CoC) d
588
532
591
592
598
550
594 e
532
529 sh
528 sh
594 e
–
–
551
558
521
551
558
521
521
551 e
551 e
553e
448
477 e
492 e
442
472 e
483 e
440
474 e
482 e
a Prepared by co-precipitation.
b Prepared according to Eq. (3) (Sn:Br2=1:2).
c From Me2Sn group.
d From Me3Sn group.
e Very low intensity.
absorptions). An IR band at 598 cm−1 indicates the
presence of Me2Sn units.
The experimental powder X-ray diffractogram
(XRD) of 2 (Fig. 4) differs notably from that of 1 (Fig.
3), suggesting the presence of a new solid phase and not
of a mixture of 1 and 3. This view is further supported
by comparison with the reported [5] XRD of 3. The
XRD of 5 is dominated by one broad reflection cen-
tered at 2[=12° aside of a few sharp, but compara-
tively weak lines, suggesting essentially an amorphous
nature.
Fig. 3. XRDs of 1: (a) simulated, (b) experimental diffractogram.
5. Solid-state NMR spectra of 1, 2 and 5
As the experimental XRD of bulk 1 agrees very
satisfactorily with the XRD simulated from data of the
single-crystal X-ray study (Fig. 3), it seems legitimate to
discuss the results of a multinuclear (i.e. 13C, 15N, 59Co,
119Sn) CPMAS solid-state magnetic resonance study of
1 in terms of the asymmetric unit known from the
crystallographic structure analysis. While the latter in-
volves six nonequivalent methyl carbon and six cyanide
nitrogen atoms, no more than two 13C resonances (Fig.
5(a)) are found in the relevant chemical shift range
(Table 4). The relative intensity of the two methyl 13C
signals is 1:2 within experimental error, so we assign the
signal at lC=17.8 ppm to the methyl groups bonded to
Sn3 and that at lC=13.4 ppm to those attached to Sn1
and Sn2. The 15N spectrum was hard to attain since
even an accumulation of 67 700 free induction decays
Fig. 4. Experimental XRDs of 1 (a) and 2 (b). 2 was prepared by
co-precipitation.

38
E. Siebel et al. / Journal of Organometallic Chemistry 604 (2000) 34–42
Fig. 5. Carbon-13 CPMAS spectra at 75.43 MHz and ambient probe
temperature of (a) 1, and (b) 2. On the left is the 13CN region and on
the right (on a different frequency scale) the CH2 and CH3 regions.
Spectrometer operating conditions: contact time 5 ms (a) and 20 ms
(b); recycle delay 1.0 s; number of transients 1000 (a) and 600 (b);
spin rate 4940 Hz (a) and 4660 Hz (b).
over a period of nearly 50 h gave only a modest S/N of
ca. 4. However, two centerband signals of equal inten-
sity are seen at lN= −132.3 and −124.9 ppm. There
is some indication of a third, weaker, peak at lN= −
96.7 ppm, but of this we cannot be certain. Such a
situation would be consistent with terminal nitrogen
resonances being to higher frequency than those of
tin-coordinated nitrogens, although it would be ex-
pected [10] that OꢀH···NC hydrogen bonding would
displace the latter again towards lower frequencies.
Owing to the effect of site inequivalences and of cou-
pling to the quadrupolar 59Co and 14N nuclei, just one
very broad cyanide carbon resonance occurs, though
with substantial partially-resolved fine structure. In-
stead of three (asymmetric unit), no more than two
Fig. 6. Tin-119 CPMAS spectra at 111.79 MHz of (a) 1, simulated
(see text for parameters), (b) 1, experimental, and (c) 2, experimental.
Spectrometer operating conditions: contact time 5 ms (b) and 20 ms
(c); recycle delay 1.0 s (b) and 0.5 s (c); number of transients 55 000
(b) and 108 224 (c); spin rate 12.0 kHz (b) and 10.1 kHz (c). It is
likely that the relatively high-speed spinning causes the sample tem-
perature to rise significantly above ambient.
119Sn centerbands (Fig. 6) appear as singlets (with an
intensity ratio of 2:1 within experimental error) at
rather low frequency, although still within the chemical
shift range characteristic of hexa-coordinate tin [12].
One potential explanation of the constantly lower num-
Table 4
Chemical shifts (in ppm) of solid 1, 2 and 5 a
b
Sample
l(13C)
1J(119Sn, 13C) b,c
l(13C) d
l(15N)
l(119Sn)
1
13.4
17.8
1005
1190
ca. 130
−132.4
−124.8
−494
−410
2
−0.4
−1.95 e
18.2
555
ca. 130
−124.7 f
−121.4 f
−122.4 g
−118.8 g
−102
n.o.
1212
1166
−488
−502
21.0
5
−1.7
22.7
27.9
635
n.o.
n.o.
ca. 170
ca. −121
−31 h
−139 h
a All centerbands are singlets.
b From CH3 groups.
c In Hz, determined from weak accompanying doublet.
d From the CN groups.
e Weak-absent in samples obtained by bromination of 3.
f
1
ꢀ J(119Sn, 15Nꢀ=304 and 292 Hz for the lines at lN=−124.7 and −121.4 ppm, respectively.
g Minor peaks, possibly from an impurity.
h There is also a signal, with indeterminate centerband, in the Me2Sn region.

E. Siebel et al. / Journal of Organometallic Chemistry 604 (2000) 34–42
39
ber of actually observed resonance lines than of those
predicted from the asymmetric unit could be the as-
sumption of a rapid (on the NMR timescale) periodic
equilibration of the {Sn1(N1C1)2} and {Sn2(N2C2)2}
fragments (see Fig. 1). The then surviving 119Sn reso-
nances (Table 4) would only distinguish between Sn3
and Sn1/Sn2. The less intense of the two methyl 13C
lines has the higher J(119Sn, 13C) value (determined
from the detectable satellite doublet arising from the
presence of 119Sn nuclei). Interestingly, a tentative de-
termination of the C(Me)ꢀSnꢀC(Me) angles of 1 from
adopting Lockhart’s [13] empirical correlation of the
1J(119Sn, 13C) value with that angle (Eq. (4)) leads to
the values of:
Me3Sn systems, p has also been found [11,14] to be low
(B0.3). The 119Sn shielding anisotropies are n=952
and 865 ppm for the isotropic shifts lSn= −410.6 and
−494.0 ppm, respectively. These values are signifi-
cantly larger than those obtained for trigonal bipyrami-
dal SnMe3 groups for compounds 2 and 5 (see below)
and related cyanide-bridged intermetallic systems [5b].
A single 59Co-NMR band is found for 1 (Fig. 7(b)),
with a typical powder bandshape arising from second-
order quadrupolar effects. Spectral simulation (Fig.
7(a)) shows that liso(59Co)= −215 ppm, the quadrupo-
lar coupling constant is 11.6 MHz and the electric field
gradient asymmetry parameter is 0.19.
The solid-state magnetic resonance spectra of co-pre-
cipitated 2 strongly suggest the absence of both 1 and 3,
as none of the main 13C and 119Sn resonances of these
polymers were detected. The 119Sn spectrum of 2
1J(119Sn, 13C)+778
C(Me)−Sn−C(Me)angle=
(4)
10.7
167° for the averaged angle at Sn1/2 and of 186°
(presumably folding back to 174°) for the angle at Sn3.
Comparison with the crystallographically obtained data
(Table 2), i.e. 167° versus 166.2°, and 174° versus
172.1°, reveals a surprisingly good agreement. The spin-
ning sideband manifolds of the 119Sn spectrum have
been analyzed (Fig. 6) to give data on the shielding
tensors. For both peaks the shielding asymmetry is zero
within experimental error. This is reasonable for Sn3
but unexpected for Sn1/2 since the latter have cis bonds
to two oxygen atoms. For other trans-cyanide-bridged
(Fig. 6(c)) displays one signal at a chemical shift (lSn
−102 ppm) typical of Me3Sn, and a doublet (lSn
=
=
−488 and −502 ppm) at shifts typical of Me2Sn,
suggesting one trans-Me3SnX2 and two Me2SnX4 envi-
ronments (with equal abundances) in the asymmetric
unit. The Me3Sn:Me2Sn ratio is difficult to determine.
The Me2Sn signals show a wide spread of spinning
sidebands typical of such groups, but an analysis is not
feasible because of the low intensities of most side-
bands. The structural conclusion is confirmed by the
appearance (Fig. 5(b)) of two 13C signals for the methyl
groups of Me2Sn but only one for Me in Me3Sn. The
intensity ratio of the three peaks is 1:1:3 (within exper-
imental error) as expected and further confirmed by
solution 1H-NMR spectroscopy (see Section 7). Of
course, in many Me3Sn trigonal bipyramidal units of
cyanide-bridged metal compounds rapid internal rota-
tion about the NꢀSnꢀN axis renders the three methyl
groups equivalent.
The combined 119Sn- and 13C-NMR information
makes it clear that the two methyl groups in a given
1
Me2Sn unit are also equivalent. The values of ꢀ JSnCꢀ,
obtained from satellite peaks in the 13C spectrum, show
clearly that the CꢀSnꢀC group in Me2Sn must have
trans-oriented (probably at approximately 180°) methyl
groups. The cyanide 13C signal of 2, like that of 1,
consists of a complex band of partially-resolved lines
from the effects of inequivalences and of a coupling to
both 14N and 59Co. The 15N spectrum of one sample of
2 contains two major lines, of equal intensity, and with
119Sn satellites, plus two minor lines. A single, rather
broad 59Co signal is seen at lCo=ca. −260 ppm, with
spinning sidebands spread over 6000 ppm, but discern-
able second-order structure.
1
The 13C and solution H spectra of 5 confirm the
Fig. 7. Cobalt-59 direct polarization MAS spectra at 71.13 MHz of 1:
(a) simulated (see text for parameters), and (b) experimental. For the
latter the spectrometer operating conditions are: pulse angle 22.5°;
recycle delay 0.2 s; number of transients 20 000; spin rate 14.5 kHz.
It is likely that the relatively high-speed spinning causes the sample
temperature to rise significantly above ambient.
presence of both Me2Sn and Me3Sn entities, although
the difficulty of locating satellite doublets arising from
119Sn, 13C coupling for the former does not allow any
prediction of the C(Me)ꢀSnꢀC(Me) angles. The pair of

40
E. Siebel et al. / Journal of Organometallic Chemistry 604 (2000) 34–42
peaks assignable to Me2Sn carbons is at rather high
frequency (lC=27.9 and 22.7 ppm). In the 119Sn spec-
trum, one centerband typical of trans-Me3Sn(NC)2
fragments appears (l= −138 ppm), while a peak at
−30 ppm may be best assigned to either of the frag-
ments:{trans-Me3Sn(NC)(OH2)} or {trans-Me3Sn-
(OH2)2} [10]. The centerband 119Sn resonance of octa-
hedrally supplemented Me3Sn fragments is not clearly
detectable, although several weak lines form a spinning
sideband manifold within the expected range. Interest-
ingly, the two quite characteristic 119Sn centerbands of
[(Me3Sn)4Fe(CN)6] (6) [8] at −108 and +47 ppm
seem to appear as well, and gain in intensity during
the course of the NMR experiment. Uptake of
thermal energy owing to extended, rapid spinning
might be responsible for a more efficient separation of
the two ‘non-alloyed’ products 6 and [(Me2Sn)2-
Fe(CN)6·xH2O]. Only a poor-quality 15N spectrum of
5 could be obtained. This showed a broad signal at
lN ꢀ −121 ppm, with possible (but ill-defined) fine
structure.
c axis infinite, linear channels of approximately rectan-
gular cross sections (ca. 0.8×1.1 nm). It is, on the
other hand, true that too voluminous frameworks are
usually circumvented by the formation of more com-
pact architectures (e.g. by mutually interpenetrating
frameworks [17]).
Although the crystal structure of 1 differs notably
from that of Zn3{Co(CN)6}2·xH2O [18] and the struc-
tures of related members of the Prussian blue family
[19], both 1 and 2 are likely to also display catalytic
properties. Like the non-organometallic double metal
cyanides, these coordination polymers dissociate par-
tially into their Lewis basic and Lewis acidic compo-
nents. For instance, 2 and 5 are, like 1, completely
soluble in D2O–NaOD (pH: ca. 8–9), and these solu-
1
tion give rise to two distinct H-NMR signals of the
hydrated Me2Sn+ and Me3Sn+ ions the intensity ratio
of which helps determining independently the Me2Sn–
Me3Sn ratio of the sample in question.
The selective bromination of 3 affording quite unex-
pectedly 2, opens a new perspective for the bromination
of other organotin metal cyanide systems.
It seems to be more difficult ascribing a reasonable
structure to 5. Although the XRD of the practically
amorphous, faintly greenish precipitate does not display
any of the stronger reflections of 6, 119Sn resonances
clearly indicative of this species emerge in the solid-
state spectrum of 5 (vide supra). Another unexpected
feature is that 5 turns blue after more rigorous drying
(e.g. in vacuo). In D2O–NaOD solution, the blue spe-
cies gives rise to a significantly larger Me3Sn–Me2Sn
ratio than 3:1 (as expected for 5). Obviously, more
detailed information is required to clarify the nature of
5.
6. Conclusions
In aqueous solution, Me2Sn2+ and [Co(CN)6]3− ions
assemble spontaneously to the novel coordination poly-
mer [(Me2Sn)3{Co(CN)6}2·6H2O] (1) wherein, accord-
ing to the more structure-related formulation:
[(Me2Sn)(Me2Sn·2H2O)2{Co(CN)6}2·2H2O], only one of
the three Me2Sn units reaches hexacoordination exclu-
sively via CN“Sn bonding. The partial coordination
of water molecules to tin atoms leads to infinite, planar
ribbons carrying at their edges all Sn-coordinated H2O
molecules, which helps in achieving a three-dimensional
supramolecular architecture by extensive OꢀH···NC hy-
drogen bonding. A structurally related coordination
7. Experimental
polymer
containing
also
Me2Sn
units
is
[(Me2Sn)3(PO4)2·8H2O] [15], wherein infinite, ribbon-
like building blocks carry similarly H2O molecules at
their edges, although here the ribbon profiles are ori-
ented alternately horizontal and vertical. Another po-
tential homologue of 1 is the coordination polymer
[(Pr2Sn)3{Fe(CN)6}2·4H2O] (Pr=propyl) the crystal
structure of which is, however, still unknown [16].
While 1 in its hydrated form seems to be energetically
more favorable than an anhydrous framework, wherein
all coordination sites of the {Me2Sn} units would be
occupied by cyanide N atoms (12:12), a more Prussian-
blue-like 3D framework may readily be envisioned for
anhydrous 2. Thus, planar layers resulting formally
from desolvated ribbons of 1 after infinite expansion
along b (Fig. 1) might be regularly stacked and con-
nected via {Me3Sn} units by means of the still terminal
cyanide ligands oriented perpendicular to the planes.
The new framework would possess along the (original)
7.1. Methods and instrumentation
All operations could be carried out without special
protection from atmosphere. IR spectra were obtained
on a Perkin–Elmer IR 1720 instrument, and Raman
spectra on a Ramanow U-1000 spectrometer of Jobin
Yvon. X-ray powder diffractograms (XRD) were taken
on
a
Phillips X’PERT diffractometer (Cu–Ka)
1
equipped with a Ni filter, and solution H-NMR spec-
tra either on a Bruker AM 360 or a Varian Gemini 200
spectrometer.
Solid-state NMR: the solid-state NMR spectra were
recorded on a Varian VXR 300 spectrometer operating
at frequencies of 75.4, 11.9, 30.4 and 71.1 MHz for 13C,
119Sn, 15N and 59Co, respectively. Crosspolarization
with high-powder proton decoupling was used for all
spectra except 59Co where direct polarization was used.
The 113C and 15N spectra were recorded with a Doty

E. Siebel et al. / Journal of Organometallic Chemistry 604 (2000) 34–42
41
1
Scientific probe with 7 mm o.d. rotors, but for the 119Sn
and 59Co spectra a faster-spinning Doty Scientific probe
with 5 mm o.d. rotors was used. Acquisition conditions
are given in the figure captions. Chemical shifts are
reported, with the high-frequency positive convention,
in ppm relative to the signals for SiMe4, SnMe4,
NH4NO3 (nitrate line), and K3[Co(CN)6](aq.) for 13C,
119Sn, 15N, and 59Co, respectively. Shielding tensor
components are defined by ꢀ|ZZ−|isoꢀ]ꢀ|XX−|isoꢀ]
ꢀ|YY−|isoꢀ, with anisotropy n=|ZZ−|iso and p=
(|YY−|XX)/n. Analysis of spinning sideband manifolds
was carried out using either an in-house computer
program [20] or the STARS Varian software. Analysis of
the 59Co bandshape for quadrupolar effects used the
STARS software. The 90° pulse for the 59Co direct
polarization spectra was calibrated using an aqueous
solution of K3[Co(CN)6].
43.53%. The solution H-NMR spectrum of 2 (solvent:
D2O–NaOD, pH ca. 9) confirms independently a
Me3Sn–Me2Sn ratio of 1:1.
2 (by bromination of 3): a solution of Br2 (either 0.22
ml=4.2 mmol or 0.01 ml=0.2 mmol) in 10 ml of
methanol was added slowly and in the dark to a
suspension of [(Me3Sn)3Co(CN)6] (3) (either 1.0 g=1.4
mmol or 0.9 g=1.27 mmol), kept in a three-necked
flask equipped with a cooler and a dropping funnel.
After stirring over ca. 12 h, the mixture is briefly heated
up to the boiling point. After filtration, washing of the
residue with MeOH and Et2O and drying in vacuo,
white powders of 2 are obtained in quantitative yields.
Elemental analysis C11H15N6CoSn2: Anal Calc. C,
25.04; H, 2.87; N, 15.93; Co, 11.17; Sn, 45.00. Found:
24.61/24.70; H, 2.81/2.86; N, 16.00/15.62; Co, 11.28/
11.30; Sn, 45.39/45.36%.
The single-crystal X-ray study of 1 was carried out
on a Hilger & Watts Y290 four-circle diffractometer
equipped with a low-temperature device (Table 1). Posi-
tions of heavy atoms were determined by direct meth-
ods, and the positions of the lighter C-, N- and
O-atoms by means of difference Fourier and least-
squares calculations. All non-hydrogen atoms except
N1, N6, C1, C5, C6, C22, C31 and C32 were refined
anisotropically. The latter atoms could be refined
isotropically. H atoms could not be localized and were
also refined isotropically in keeping a fixed CꢀH dis-
tance of 96.0 pm.
Preparation of 5: a solution of Me2SnCl2 (2.19 g, 0.01
mol) and Me3SnCl (4.0 g, 0.02 mol) in 50 ml of H2O
was added under stirring to a solution of K4[Fe(CN)6]
(4.22 g, 0.01 mol) in 50 ml of H2O. The almost white,
paste-like solid obtained after filtration and washing
(H2O, Et2O) is transformed after drying (avoiding vac-
uum) into a faintly greenish powder that turns black at
350°C. Final yield: 7.1 g (95%). Elemental analysis
C14H30N6O3FeSn3 (x=3): Anal. Calc. C, 22.65; H,
4.07; N, 11.32; Fe, 7.52; O, 6.47; Sn, 47.97. Found: C,
21.31; H, 7.76; N, 11.51; Fe, 7.93; O, 7.85; Sn, 43.25%.
Note added in proof: briefly after the submission of our
manuscript, crystal structures of the anhydrous ‘homo-
logues’ of 1, [(R2Sn)3{Co(CN)6}2] with R=vinyl, n-
propyl and n-butyl, have been reported [22]. In these
coordination polymers, all four coordination sites of
each {R2Sn}2+ unit are in fact occupied by cyanide N
atoms (cf. Section 6).
7.2. Preparation of 1, 2 and 5
[(Me2Sn)3{Co(CN)6}2·xH2O] (1): a clear solution of
Me2SnCl2 (3.30 g, 0.015 mol) in 50 ml of H2O was
added slowly to a solution of K3[Co(CN)6] (3.32 g, 0.01
mol) in 50 ml of H2O. The white voluminous precipi-
tate which appears after up to 1 min is separated from
the filtrate, washed subsequently with H2O and Et2O
and dried finally in vacuo (oil pump). Yield: 7.6 g
(83%), decomposition temperatures: 200/360°C (green/
black). Elemental analysis C18H22N12O2Co2Sn3 (x=2):
Anal. Calc. C, 23.59; H, 2.42; N, 18.35; Co, 12.87; Sn,
39.28. Found: C, 23.65; H, 2.23; N, 17.65; Co, 12.87;
Sn, 38.42%.
[(Me2Sn)(Me3Sn)Co(CN)6] (2; co-precipitation): a so-
lution of Me2SnCl2 (2.19 g, 0.01 mol) and Me3SnCl (2.0
g, 0.01 mol) in 50 ml of H2O were added under
vigorous stirring to a solution of K3[Co(CN)6] (3.32 g,
0.01 mol) in 50 ml of H2O. After filtration of the
spontaneously formed white precipitate, subsequent
washing with H2O and Et2O and drying in vacuo (oil
pump), 4.8 g of analytically pure 2 (yield: 91%) are
obtained. Thermal decomposition 300°C (blue). Ele-
mental analysis C11H15N6CoSn2: Anal. Calc. C, 25.04;
H, 2.87; N, 15.93; Co, 11.17; Sn, 45.00. Found: C,
25.03; H, 2.92; N, 15.48; Co, 11.60; O, 0.50; Sn,
8. Supplementary material
Full details of the crystal structure determination of 1
have been deposited with the Cambridge Crystallo-
graphic Data Centre, CSD 407641 [21]. Copies of the
data can be obtained free of charge from the Director,
CCDC, 12 Union Road, Cambridge CB2 1EZ, UK
(Fax: +44-1223-336033; e-mail: deposit@ccdc.cam.
ac.uk or www: http://www.ccdc.cam.ac.uk).
Acknowledgements
This work was supported by the Deutsche
Forschungsgemeinschaft,
DFG
(Schwerpunktpro-
gramm: Nanoporo¨se Wirt/Gast-Systeme) and the
Fonds der Chemischen Industrie. The authors are

42
E. Siebel et al. / Journal of Organometallic Chemistry 604 (2000) 34–42
[10] P. Schwarz, E. Siebel, R.D. Fischer, N.A. Davies, D.C. Apper-
ley, R.K. Harris, Chem. Eur. J. 4 (1998) 919.
grateful to S. Samba for valuable technical assistance.
The extensive CPMAS NMR work was supported by
the UK EPSRC (through the National Solid-State
NMR Service based at Durham).
[11] P. Schwarz, S. Eller, E. Siebel, T.M. Soliman, R.D. Fischer,
D.C. Apperley, N.A. Davies, R.K. Harris, Angew. Chem. 108
(1996) 1611; Angew. Chem. Int. Ed. Engl. 35 (1996) 1525.
[12] Actually, relatively few quasi-octahedral R2Sn-derivatives with
l(119Sn) values around −400 ppm have been reported, e.g. (a)
W.F. Howard Jr., R.W. Crecely, W.H. Nelson, Inorg. Chem. 24
(1985) 2204. (b) A. Lycka, J. Holecek, B. Schneider, J. Straka, J.
Organomet. Chem. 389 (1990) 29.
[13] (a) T.P. Lockhart, W.F. Manders, J. Am. Chem. Soc. 109 (1987)
7015. (b) T.P. Lockhart, F. Davidson, Organometallics 6 (1987)
2478.
[14] R.K. Harris, D.C. Apperley, N.A. Davies, R.D. Fischer, Bull.
Magn. Reson. 15 (1993) 22.
[15] J.P. Ashmore, T. Chivers, K.A. Kerr, J.H.G. van Roode, Inorg.
Chem. 16 (1977) 191.
[16] A. Bonardi, C. Carini, C. Pelizzi, G. Pelizzi, G. Predieri, P.
Tarasconi, M.A. Zoroddu, K.C. Molloy, J. Organomet. Chem.
401 (1991) 283.
[17] S.R. Batten, R. Robson, Angew. Chem. 110 (1998) 1558; Angew.
Chem. Int. Ed. Engl. 37 (1998) 1460.
[18] D.F. Mullica, W.O. Milligan, G.W. Beall, W.L. Reeves, Acta
Crystallogr. Sect. B 34 (1978) 3558.
References
[1] R.A. Livigni, R.J. Herold, O.C. Elmer, S.L. Aggarwal, ACS,
Symp. Ser. No. 6 (polyethers) Washington, DC, 1975, p. 20.
[2] J.L. Schuchardt, S.D. Harper, Proceedings of the 32nd Annual
Polyurethane Technical/Marketing Conference (Oct. 14, 1989) p.
360.
[3] J. Hofmann, Patent WO 99/19063 (22.04.99). For a more de-
tailed survey of the development taking place in this field, the
IPC Group C 08 G65/10 should be consulted.
[4] (a) Polymeric Me2Sn(CN)2: J. Konnert, D. Britton, Y.M. Chow,
Acta Crystallogr. Sect.
B 28 (1972) 180. (b) Polymeric
Me2Sn{N(CN)2}2: Y.M. Chow, Inorg. Chem. 10 (1971) 1938.
[5] (a) K. Yu¨nlu¨, N. Ho¨ck, R.D. Fischer, Angew. Chem. 97 (1985)
863; Angew. Chem. Int. Ed. Engl. 24 (1985) 879. (b) U. Behrens,
A.K. Brimah, T.M. Soliman, R.D. Fischer, D.C. Apperley, N.A.
Davies, R.K. Harris, Organometallics 11 (1992) 1718.
[6] Y. Sasaki, H. Imoto, O. Nagano, Bull. Chem. Soc. Jpn. 57
(1984) 1417.
[7] See: S. Boue´, M. Gielen, J. Nasielski, J.-P. Lieutnant, R. Spiel-
mann, Bull. Soc. Chim. Belg. 78 (1969) 135.
[8] S. Eller, P. Schwarz, A.K. Brimah, R.D. Fischer, D.C. Apperley,
N.A. Davies, R.K. Harris, Organometallics 12 (1993) 3232.
[9] M. Heibel, G. Kumar, C. Wyse, P. Bukovec, A.B. Bocarsly,
Chem. Mater. 8 (1996) 1504.
[19] G. Malecki, A. Ratuszna, Powder Diffraction 14 (1999) 25.
[20] J. Ascenso, H. Bai, R.K. Harris, L.H. Merwin, Sideband fitting
program, University of Durham, 1985, 1987, 1991.
[21] For a recent discussion in the Internet on the justification of
empirical absorption corrections (like DIFABS), see: http://cryst-
sun1.unige.ch/stxnews/stx/discuss/dis-dif2.htm.
[22] T. Niu, A.J. Jacobson, Inorg. Chem. 38 (1999) 5346.
.
.