Polar self-assembly: Steric effects leading to polar mixed-ligand coordination cages

Article abstract of DOI:10.1002/chem.200501007

We present a highly unusual example of self-assembly, specifically a polar, mixed-ligand cage which forms in preference to symmetrical homoligand products, and which suggests that steric effects can be exploited to obtain novel non-uniform polyhedral cage

Full text of DOI:10.1002/chem.200501007

DOI: 10.1002/chem.200501007
Polar Self-Assembly: Steric Effects Leading to Polar Mixed-Ligand
Coordination Cages
[a]
Jianyong Zhang, Philip W. Miller, Mark Nieuwenhuyzen, andStuart L. James*
Abstract: We present a highly unusual
example of self-assembly, specifically a
polar, mixed-ligand cage which forms
in preference to symmetrical homo-
ligand products, and which suggests
that steric effects can be exploited to
obtain novel non-uniform polyhedral
cages. In particular, reaction between
the bulky tripodal triphosphine 2,4,6-
from trigonal-pyramidal geometry, the
basal vertices of which are defined by
three bulky triphosphines, L1, and the
apical vertex by the non-bulky trini-
trile, L2. There is apical elongation
amounting to 19% in comparison to
the ideal uniform tetrahedron. The
cage also encapsulates an SbF6 anion.
19F NMR spectra in solution for the
analogous PF6 complex [Ag4(L1)3(L2)-
(PF6)]3+, 1-PF6, confirm that one anion
is also encapsulated in solution. The
isolated due to slow decomposition in
solution. The selective formation of
these mixed-ligand cages is discussed in
terms of ligand–ligand and ligand-in-
cluded anion steric repulsions, which
we propose prevent the formation of
the competing hypothetical homo-
ligand tetrahedral structure [Ag4(L1)4-
(SbF6)]3+, and thus favour the mixed
ligand cage. “Cage cone angles” for L1
and L2 are estimated at 1158 and 1018,
respectively. Variable-temperature
31P NMR spectroscopy shows that com-
plex 1-SbF6 and the related previously
reported partial tetrahedral complex
[Ag4(L1)3(anion)]3+ undergo dynamic
twisting processes in solution between
enantiomeric C3-symmetry conforma-
tions.
tris(diphenylphosphino)triazine,
L1,
the non-bulky tripodal trinitrile 2,4,6-
tris(cyanomethyl)trimethylbenzene, L2
À
and
silver
hexafluoroantimonate,
synthesis of the analogous CF3SO3
AgSbF6, in a 3:1:4 ratio gives the
mixed-ligand aggregate [Ag4(L1)3(L2)-
(SbF6)]3+, 1-SbF6, instantly as the only
product in quantitative yield. The X-
ray crystal structure of complex 1-SbF6
is consistent with the suspected solu-
tion-state structure. The cage derives
complex, [Ag4(L1)3(L2)(OTf)]3+
,
1-
OTf, in solution is also described, al-
though 1-PF6 and 1-OTf could not be
Keywords: anions · cages · phos-
phane · self-assembly
Introduction
icantly widen the scope of this chemistry. For example, polar
shapes would have different guest inclusion characteristics,
and, as we have previously shown, their directionality can
be exploited to assemble still larger, discrete coordination
clusters of clusters.[2] One strategy to obtain polar shapes is
to take a uniform polyhedral structure and formally replace,
in a controlled and selective way, one or more of the bridg-
ing ligands with a different type of bridging ligand. Howev-
er, when mixtures of different bridging ligands are allowed
to react with labile metal ions in one pot, normally homo-
ligand aggregates are obtained instead (i.e. the ligands “self-
select”).[1,3] To our knowledge, only a small number of ex-
ceptions to this general behaviour have been described.[4] In
one case a non-polar mixed-ligand aggregate was obtained
selectively due to the template effect of an appropriately
shaped guest molecule.[4a] Most relevant to our work de-
scribed herein, other mixed-ligand, but non-polar, structures
have been obtained selectively by combining bulky and non-
bulky ligands. The principle of this approach, as illustrated
Although coordination-based self-assembly has been dem-
onstrated as an effective way to prepare large, multimetallic
cages and rings, these structures are normally restricted to
being highly symmetrical and non-polar.[1] For example,
large coordination cages reported to date are normally
based on uniform polyhedra, that is, those in which all the
vertices are identical, such as the Platonic and Archimedean
classes.[1] Ways to selectively obtain polar cages would signif-
[a] Dr. J. Zhang, Dr. P. W. Miller, Dr. M. Nieuwenhuyzen,
Prof. S. L. James
Centre for the Theory and Application of Catalysis
School of Chemistry, Queenꢀs University Belfast
David Keir Building, Stranmillis Road, Belfast BT9 5 AG
Northern Ireland (UK)
E-mail: s.james@qub.ac.uk
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
2448
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 2448 – 2453

FULL PAPER
by the study of mixed-component trimers by Severin
et al.,[4b] is that coordinatively saturated structures based
only on the bulky component are sterically disfavoured, and
this forces the system to form aggregates based on both
bulky and non-bulky building blocks (i.e. it prevents self-se-
lection). Examples of cages prepared by using this approach,
which contain a mixture of ligands, but which are still non-
polar, are the cylinders reported by Lehn et al.,[4c] and a
polar tetrapalladium mixed-ligand capsule reported recently
by Kobayashi et al.[4d] The most systematic use of this ap-
proach is by Schmittel et al.[4e] who have employed phen-
athrolines with bulky substituents to favour the formation of
heteroleptic complexes (the HETPHEN approach) and thus
have rationally prepared non-regular, but still non-polar,
polygons such as rectangles and ladders.[4e]
In recent years we have exploited bridging di- and tri-
phosphines as novel building blocks for the self-assembly of
discrete coordination cages as well as infinite coordination
polymers (metal-organic frameworks).[2,5–7] Triarylphos-
phines are unusually bulky ligands for coordination-based
self-assembly and have given rise to some interesting results.
For example, most recently we have found that the triazine-
cored tripodal ligand tris(diphenylphosphino)triazine, L1,
reacts with silver ions to give, selectively, the unsaturated
partial tetrahedral bowl-shaped structure [Ag4(L1)3(X)]3+
(X=OTf, SbF6), which has a metal-to-ligand ratio of 4:3.[2,6]
This compound is unusual in being an unsaturated, open,
polar structure. Interestingly, the expected “conventional”
Figure 1. Synthesis and X-ray crystal structure of the mixed-ligand cage
complex [Ag4(L1)3(L2)(SbF6)]3+ 1-SbF6 (Ag: dark blue; P: orange; N:
light blue; Sb: pink, F: green). Hydrogen atoms are omitted for clarity.
complete tetrahedral structure [Ag4(L1)4(X)]3+
, with a
metal-to-ligand ratio of 4:4, was not observed even when ad-
ditional L1 was added. This unusual behaviour can be ascri-
bed to significant ligand–ligand and/or ligand–anion steric
repulsion, which is exacerbated by the presence of bulky
phenyl groups on the ligands.
We report here the addition of a non-bulky tripod to this
“incomplete” structure as a way to form, with high selectivi-
ty, a closed, mixed-ligand cage. This closed cage is notewor-
thy because in contrast to most coordination cages it is not
based on a uniform polyhedron, and furthermore it is polar.
It also provides further insight into the role of sterics in sta-
bilizing unusual self-assembled cages.
13.7 ppm, with 1J
P =532 Hz. The 19F spectrum shows
109Ag,31
one encapsulated PF6 anion (d=À74.5 ppm) and three free
1
PF6 anions (d=À73.8 ppm). The H NMR signals of L2 are
shifted in the complex (d=3.99 ppm, CH2; d=2.38 ppm,
CH3) compared to the free ligand (d=3.74 and 2.46 ppm).
Compound 1-PF6 decomposed slowly in solution and X-ray-
quality crystals could not be obtained. However, NMR data
suggested that similar capped structures, 1-SbF6 and 1-OTf,
were formed when AgSbF6 or AgOTf were used in place of
AgPF6, although the 31P signals for the AgP2N centres were
broader for both of these complexes. We relate this broad-
ness to a dynamic twisting of the structure between enantio-
meric C3-symmetry conformations as described below. Crys-
tals of 1-SbF6 were obtained by layering a CDCl3/CD3NO2
solution with benzene, and were solved by X-ray crystallog-
raphy (Figure 1), which confirmed the suspected mixed-
ligand cage structure with an encapsulated SbF6 anion. The
Ag4(L1)3 fragment of 1-SbF6 is not greatly distorted com-
Results andDiscussion
When tris(diphenylphosphino)triazine L1,[2,6] and the non-
bulky tripodal trinitrile 2,4,6-tris(cyanomethyl)trimethylben-
zene, L2, are mixed with AgPF6 in the appropriate 3:1:4
ratio in CDCl3/CD3NO2, a single species is formed quantita-
tively, as indicated by 31P NMR spectroscopy, which corre-
sponds to the trigonal-pyramidal, or “capped bowl” struc-
pared to the previously characterised “free bowl” complex
[6]
[Ag3(L1)4(OEt2)(SbF6)]3+
.
However, incorporation of the
trinitrile as a lid does result in the SbF6 anion shifting fur-
ther away from the base of the cavity (AgP3···Sb=4.58 Š for
1-SbF6 and 4.42 Š for [Ag3(L1)4(OEt2)(SbF6)]3+ [5]). Within
1-SbF6 there are shorter F···Ag distances from the encapsu-
lated anion to the AgP2N centres (2.71, 2.83, 3.15 Š) than to
the AgP3 centre (shortest=3.29 Š), suggesting greater inter-
action with the former. The SbF6 anion lies in the same
plane as the three AgP2N centres with Ag-Sb-Agrim angles
ture [Ag4(L1)3(L2)PF6]3+, 1-PF6. For reference, the X-ray
À
crystal structure of the analogous SbF6
complex
[Ag4(L1)3(L2)SbF6]3+ 1-SbF6, which is described more fully
below, can be seen in Figure 1. The 31P NMR spectrum of 1-
PF6 shows the presence of an AgP3 centre at d=15.5 ppm,
with 1J
P =363 Hz, as well as AgP2N centres at d=
109Ag,31
Chem. Eur. J. 2006, 12, 2448 – 2453
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2449

S. L. James et al.
of 118.5, 120.0, 121.48 (sum=359.98). In its capping coordi-
nation mode, the trinitrile is slightly distorted at the CH2
groups (C···CH2···C angles are 107.6, 110.7, 111.08) to allow
it to coordinate simultaneously to the three Ag centres, and
the C=N-Ag angles are distorted from linearity at 163.2,
168.5 and 175.68. The angles at the AgP2N coordination cen-
tres reveal approximately planar coordination (average
sum=359.18) but with significant distortion from regular
trigonal-planar geometry.
angles of 1018 for L2 and 1158 for L1.[10] It is interesting
that for L1 the cage cone angle exceeds the theoretical max-
imum for the formation of a tetrahedral coordination cage
(1098).
A further factor which may be important in allowing the
complete closed cages to form is the greater reach of L2
compared to that of L1. In particular, each arm of L2 has
two additional carbon spacer atoms between the coordinat-
ing atom and the central ring, than in L1 (in which the P
centres connect directly to the triazine core). These addi-
tional spacers may be important in allowing sufficient room
for anion encapsulation in, and therefore stabilisation of, the
closed structures 1. In fact the closest distance between an F
atom of the included anion, and a C atom of the central ring
of L2 in 1-SbF6 is 3.31 Š (F14···C140). This is only slightly
The space-filling view of 1-SbF6 in Figure 2 shows how the
steric bulk of the three triphosphine ligands (L1=red,
green, and light brown) leaves only a small region for the
[9]
larger than the combined F and C van der Waals radii
(1.47(F) + 1.70(C)=3.17 Š). Figure 2 (bottom) illustrates
that there is still a reasonably tight fit for the anion. Accord-
ingly we investigated whether the use of smaller anions
might allow the previously unobserved homo-ligand tetrahe-
dral cage [Ag4(L1)4(anion)]3+ to form. Halides ClÀ, BrÀ or
IÀ, as their tetrabutyl ammonium salts, were added to equi-
molar solutions of AgSbF6 and L1. In addition, equimolar
solutions of AgNO2 and L1 were studied. In all cases room
temperature 31P NMR spectra showed broad singlets without
evidence of coupling to 107Ag and 109Ag. In each case, on
cooling the samples, complicated spectra consisting of large
numbers of peaks were observed, from which no structural
assignments could be made, and suggesting that mixtures of
several complexes were present. Although it cannot be
ruled out that these mixtures contained small amounts of
the symmetrical tetrahedral cages [Ag4(L1)4(anion)]3+
(anion=Cl, Br, I, NO2) they were clearly not formed with
any appreciable degree of selectivity. This observation sup-
ports the argument that ligand-ligand steric repulsion, rather
than ligand-included anion steric repulsion, is the reason for
the instability of the homo-ligand tetrahedral cage. Howev-
er, it should also be noted that with these small anions, we
did not observe partial tetrahedral structures [Ag4(L1)3-
(anion)]3+, which are the main species present under the
same conditions when larger anions (SbF6, OTf) are used.[6]
This therefore points instead to the complication that there
is actually an active templating effect of these smaller, more
nucleophilc anions. Through their stronger coordination to
the silver centres, they apparently stabilise instead mixtures
of other complexes. Anion templating effects are to be ex-
pected since we have previously observed them in adaman-
toid triphosphine–silver cages.[7]
Figure 2. The space-filling views showing: a) cage 1-SbF6 with “cage cone
angles” (trinitrile=light blue, triphosphines=red, green, brown); b) and
c) comparison of the sizes of L1 and L2, and d) part of the structure
showing how the extended reach of L2 allows room for the anion.
fourth ligand (L2=light blue) to complete the polyhedral
structure. We have previously found that L1 itself is unable
to form the hypothetical complete closed tetrahedral struc-
[6]
ture [Ag4(L1)4(SbF6)]3+
,
even on addition of excess L1.
The limited remaining space in the partial tetrahedral struc-
ture is clearly sufficient however for the sterically non-de-
manding trinitrile ligand L2. “Cage cone angles” (i.e. the
cluster cone angle[8] concept applied in this case to a coordi-
nation cage) give a rough comparison of the different ligand
sizes, even though the shape of L1 is rather irregular. They
were estimated for the two ligands by taking the included
Sb centre as the cage centroid, and representative angles to
the outermost diametrically opposed atoms in the ligands.
For L2 a CH3···Sb···CH2 distance was taken (where the CH3
and CH2 groups are mutually para), specifically
H13 K···Sb1···C133=70.48, and for L1 phenyl H atoms, spe-
cifically H20A···Sb1···H32A=94.88. Taking into account the
van der Waals radii of carbon and hydrogen (1.70 and
1.20 Š, respectively[9]) then gives more realistic cage cone
Whereas most polyhedral coordination cages have identi-
cal vertices (i.e. they are derived from uniform polyhedra),
the mixed-ligand cages 1-PF6, 1-SbF6 and 1-OTf described
here have non-identical vertices and so are formally derived
from a non-uniform polyhedron, the trigonal pyramid. The
effect of this lower symmetry on the interior dimensions of
1-SbF6 can be seen from the shape defined by the centroids
of the four capping ligands (Figure 3). This pyramid has an
average basal edge length of 6.53 Š (L1 to L1 distances)
2450
www.chemeurj.org
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 2448 – 2453

Polar Self-Assembly
FULL PAPER
17.0 ppm) are broadened sufficiently so that signals due to
the individual 107Ag and 109Ag isotopomers are not resolved.
To probe any dynamic behaviour of this structure, low-tem-
perature spectra were obtained. As the temperature is low-
ered, the broadness of these signals increases until from À10
to À308C only very broad features close to the baseline are
visible. At À508C new broad signals emerge from the base-
line at dꢀ21.5, 19.0, 11.8 and 9.5 ppm which appear from
their positions to show evidence of P-Ag coupling, although
again signals due to the 107Ag and 109Ag isotopomers are not
resolved. Overall, the spectra can be explained by a dynamic
twisting of the cage structure, in which enantiomeric C3-sym-
metry conformations interconvert quickly at higher temper-
atures, but slowly at low temperature. In this way, the two P
atoms at each AgP2N centre become inequivalent (PB and
PC in Figure 5) at low temperature.
Figure 3. The trigonal pyramid described by the centroids of the four li-
gands of complex 1, with distances in angstroms.
and average base-to-apex length of 7.79 Š (L1 to L2 distan-
ces) which amounts to a stretching of 19% compared to the
“ideal” tetrahedron.
We also studied the effect of varying the stoichiometry of
the reagents on the assembly of the cage. In principle, a
range of non-uniform structures can be accessed by further
formal replacement of the triphosphine for trintriles. In fact,
in this case, use of a 2:2:4 stiochiometry (L1:L2:Ag) gave
cage 1 as the only identifiable peaks in the 31P NMR spec-
trum. Increasing the amount of L1 to a stoichiometry of
4:1:4 or L2 to a stoichiometry of 3:2:4 also gave only cage 1
according to 31P NMR spectroscopy. In the latter case, the
free and coordinated L2 could not be distinguished by
1H NMR spectroscopy, suggesting that there was fast ex-
change on the NMR timescale. Lastly, the formation of 1
was also found to be independent of the overall concentra-
tion, that is, there was no evidence of higher oligomers 12,
or 14 at high concentration.
Figure 5. Interconversion by twisting between C3-symmetry conforma-
tions which renders the P centres at each AgP2N centre inequivalent in
the slow exchange regime. The capping trinitrile ligand has been omitted
for clarity.
Dynamic behaviour: As stated above, in contrast to the
sharp 31P NMR spectrum of 1-PF6, 1-SbF6 showed some
broadening in its 31P NMR spectrum at room temperature
(Figure 4). Whilst the signals due to the AgP3 centre are
sharp, and couplings to both 107Ag and 109Ag isotopes are re-
solved, the signals for the AgP2N centres (at dꢀ15.5 and
We also investigated the previously reported partial tetra-
hedral aggregate [Ag4(L1)3(OTf)](OTf)3 with regard to any
dynamic behaviour and analogous spectra were observed
(Figure 6). The room temperature spectrum in this case is
complicated by the coincidence of signals due to the AgP3
and AgP2 centres (at dꢀ16 ppm). However, as the tempera-
ture is lowered, only the signals due the AgP2 centres
become broad. At À708C the process is sufficiently slow
that signals due to the inequivalent P atoms are fully re-
solved, and this part of the spectrum then appears as the su-
perposition of two ABX spin systems (one for each Ag iso-
1
tope). JAg,P for each P atom in the AgP2 centres is 501 Hz,
2
and JP, P through the Ag centre is 81 Hz. We are unaware of
previous reports of 2J
coupling through a silver centre.
P, P
Spectra run at one degree intervals between À19 and
À238C gave a coalescence temperature of À228C, which
corresponds to a Gibbs free energy of activation for the
twisting process of 43.8 kJmolÀ1 from the Eyring equation.
A further point is whether the polarity of the structure is
manifested in the 19F spectrum of the included PF6 anion.
However, at low temperatures we saw no evidence for any
inequivalence of the F nuclei, indicating that tumbling of
anion was not frozen out or that the asymmetry does not
alter the chemical shifts of the individual F atoms.
Figure 4. Variable-temperature 31P NMR spectra of cage complex 1-SbF6.
Chem. Eur. J. 2006, 12, 2448 – 2453
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2451

S. L. James et al.
(300 MHz): d=7.61–7.04 (m, 96H; Ph), 6.43 (t, 12H; Ph), 3.97 (s, 6H;
CH2), 2.30 ppm (s, 9H; CH3); IR (Nujol): n˜ =2272 (n(CN)), 655 (n-
(SbF6)) cmÀ1
.
Synthesis of complex 1-PF6: Solid L1 (32.3 mg, 0.051 mmol) and L2
(4.0 mg, 0.017 mmol) were added to solution of AgPF6 (17.2 mg,
a
0.068 mmol) in CDCl3 (2 mL) and CD3NO2 (0.5 mL). The solution ob-
tained was used for NMR spectroscopy. 31P NMR (121.5 MHz, CDCl3/
CD3NO2, 3:1): d=15.5 (1J
P =363 Hz), 13.7 ppm (1J
P =532 Hz);
109Ag,31
109Ag,31
19F NMR (282 MHz, CDCl3/CD3NO2, 3:1): d=À73.8 (1J P =709 Hz,
19F,31
free PF6À), À74.5 ppm (1J P =726 Hz, ecapsulated PF6); 1H NMR
19F,31
(300 MHz, CD2Cl2/CD3NO2, 3:1): d=7.64–7.17 (m, 90H; Ph), 7.11 (t,
6H; Ph), 6.50 (t, 12H; Ph), 3.99 (s, 6H; CH2), 2.38 ppm (s, 9H; CH3).
193 K: 19F NMR (282 MHz, CD2Cl2/CD3NO2, 3:1): d=À73.3 (1J
=
P
19F,31
710 Hz, free PF6À), À76.3 ppm (1J P =726 Hz, encapsulated PF6);
19F,31
1H NMR (300 MHz, CD2Cl2/CD3NO2, 3:1): d=7.63–6.86 (m, 105H; Ph),
6.58 (s, 3H; Ph), 4.06 (s, 3H; CH2), 3.77 (s, 3H; CH2), 2.32 (s, 9H; CH3).
Synthesis of complex 1-OTf: The procedure was analogous to that of 1-
PF6. 31P NMR (121.5 MHz, CDCl3/CD3NO2, 3:1): d=15.8 (1J
=
109Ag,31
P
1
360 Hz), 12.7 ppm; H NMR (300 MHz): d=7.69–7.21 (m, 90H; Ph), 7.09
(t, 6H; Ph), 6.47 (t, 12H; Ph), 4.07 (s, 6H; CH2), 2.32 ppm (s, 9H; CH3);
19F NMR (282 MHz, CDCl3/CD3NO2, 3:1): d=À75.9 (free OTfÀ),
À74.9 ppm (coordinated OTfÀ).
Synthesis of L2: This was by a modified procedure of Podlaha et al.[11]
NaCN (0.892 g, 18.2 mmol) was dissolved in DMSO (30 mL) at 408C. A
solution of 1,3,5-tris(bromomethyl)-2,4,6-trimethylbenzene (2.394 g,
6.0 mmol) in DMSO (30 mL) was then added slowly over 2 h. The reac-
tion mixture was stirred for 20 h at 408C, then poured into ice-water
(40 mL) and extracted with Et2O and CH2Cl2. The combined organic
layers were washed with brine, dried and concentrated to give a colour-
less solid (0.44 g, 31%). Elemental analysis calcd (%) for C15H15N3: C
75.92, H 6.37, N 17.71; found: C 75.44, H 6.93, N 17.49; 1H NMR
(300 MHz, CDCl3): d=3.74 (s, 6H; CH2), 2.46 ppm (s, 9H; CH3).
Halide and nitrite reactions: A solution of L1 (63.3 mg, 0.10 mmol) in
CDCl3 (2 mL) was added to a suspension of AgNO2 (15.4 mg, 0.10 mmol)
in CD3NO2 (0.5 mL). The mixture was stirred for 30 min. A trace amount
of insoluble material was removed by filtration and the resulting solution
was examined by NMR spectroscopy (31P NMR (121.5 MHz, CDCl3/
CD3NO2, 3:1): d=15.5 ppm (br.) plus other minor peaks). Reactions with
halides performed similarly with equimolar quantities of AgSbF6, L1, and
NBu4X (X=halide) as the halide source.
Figure 6. Variable-temperature 31P NMR spectra for the partial tetrahe-
dral complex [Ag4(L1)3(OTf)](OTf)3.
Conclusion
A highly unusual, polar coordination cage has been found to
assemble selectively from a mixture of bulky and non-bulky
ligands and silver ions. Steric destabilisation of the compet-
ing hypothetical homo-ligand aggregate [Ag4(L1)4(anion)]3+
appears to be the key to the selective formation of this un-
usual structure. Clearly, the exploitation of steric effects pro-
vides a way to extend the self-assembly of polyhedral cages
toward less regular, even polar, structures, which are not ob-
tained using current “non-bulky” methodologies.
X-ray crystallographic data:[12, 13] (C163H149.50Ag4F24N17O11P9Sb4): Mr =
4175.71, monoclinic, space group P21/c, a=18.1599(16), b=30.704(3), c=
À3
31.043(3) Š, b=98.872(3)8, U=17103(3) Š
, ,
Z=4, m=1.242 mmÀ1
R
int =0.1116. A total of 197543 reflections were measured for the angle
range 2<2q<578, and 40379 independent reflections were used in the
refinement. The final parameters were wR2=0.3186 and R1=0.1288 [I>
2sI]. Some of the solvent molecules and SbF6 anions showed evidence of
disorder. The residual peaks in the final difference map are associated
with these disordered regions. The disordered anion was modeled as oc-
cupying two sites with occupancy of 50%. Attempts an modeling the dis-
order of the solvent molecules were unsuccessful and thus the solvents
were restrained and given full occupancy except for the ethanol molecule
which is disordered about an inversion centre and has an occupancy
factor of 50%.
Experimental Section
General: Reagents were purchased from Aldrich and used as supplied
unless otherwise stated. L1 was prepared as previously described.[6]
31P NMR spectra were recorded at 121.5 MHz and 278C unless otherwise
stated, and are referenced to external H3PO4 (aq) 85%.
Acknowledgement
Synthesis of complex 1-SbF6: Solid 2,4,6-tris(diphenylphosphino)triazine
(L1) (130.6 mg, 0.21 mmol) and 1,3,5-tris(cyanomethyl)-2,4,6-trimethyl-
benzene (L2)[11] (16.3 mg, 0.069 mmol) were added to a solution of
AgSbF6 (94.5 mg, 0.275 mmol) in CHCl3 (3 mL) and CH3NO2 (1 mL).
The resulting solution was layered with benzene to give block- and
needle-shaped crystals. Yield 215 mg, 89%. Elemental analysis (after
drying in vacuo) calcd (%) for [Ag4(L1)3(SbF6)4(L2)]: C 45.14, H 3.01, N
4.79; found: C 45.65, H 3.33, N 4.58; 31P NMR (121.5 MHz, CDCl3/
We thank The Leverhulme Trust for financial support.
[1] G. F. Swiegers, T. J. Malefetse, Coord. Chem. Rev. 2002, 225, 91–
121; D. L. Caulder, K. N. Raymond, Acc. Chem. Res. 1999, 32, 975–
982; P. J. Stang, Chem. Eur. J. 1998, 4, 19–27; M. Fujita, Chem. Soc.
Rev. 1998, 27, 417–425; C. J. Jones, Chem. Soc. Rev. 1998, 27, 289–
299; R. W. Saalfrank, B. Demleitner, “Transition Metals in Supra-
CD3NO2, 3:1, 258C): d=14.7 (1J
P =361 Hz), 12.9 ppm; 1H NMR
109Ag,31
2452
www.chemeurj.org
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 2448 – 2453

Polar Self-Assembly
FULL PAPER
molecular Chemistry” in Perspectives in Supramolecular Chemistry
vol. 5, (Ed.: J. P. Sauvage), Wiley-VCH, Weinheim, 1999, pp. 1–51; S.
Leininger, B. Olenyuk, P. J. Stang, Chem. Rev. 2000, 100, 853–908;
B. Moulton, J. Lu, A. Mondal, M. J. Zaworotko, Chem. Commun.
2001, 863.
James, Chem. Commun. 2002, 78; P. W. Miller, M. Nieuwenhuyzen,
S. L. James, Chem. Commun. 2002, 2008.
[6] P. W. Miller, M. Nieuwenhuyzen, J. P. H. Charrmant, S. L. James,
CrystEngComm 2004, 408.
[7] For examples of anions templating adamantoid silver–triphosphine
cages see: S. L. James, D. M. P. Mingos, A. J. P. White, D. J. Williams,
Chem. Commun. 1998, 2323; X. Xu, M. Niewenhuyzen, M. Cham-
bers, E. MacLean, S. J. Teat, S. L. James, Chem. Commun. 2002, 78.
[8] D. M. P. Mingos, Inorg. Chem. 1982, 21, 464–466.
[2] J. Zhang, M. Nieuwenhuyzen, J. P. H. Charmant, S. L. James, Chem.
Commun. 2004, 2808.
[3] For example see: C. Addicott, N. Das, P. J. Stang, Inorg. Chem.
2004, 43, 5335; R. Stiller, J.-M. Lehn, Eur. J. Inorg. Chem. 1998, 977;
A. Wu, L. Isaacs, J. Am. Chem. Soc. 2003, 125, 4831.
[9] A. Bondi, J. Phys. Chem. 1964, 68, 441.
[4] a) K. Kumazawa, K. Biradha, T. Kusukawa, T. Okano, M. Fujita,
Angew. Chem. 2003, 115, 4039–4043; Angew. Chem. Int. Ed. 2003,
42, 3909–3913; b) Z. Grote, R. Scopelliti, K. Severin, Angew. Chem.
2003, 115, 3951–3955; Angew. Chem. Int. Ed. 2003, 42, 3821–3825;
c) P. Baxter, J.-M. Lehn, A. DeCian, J. Fischer, Angew. Chem. 1993,
105, 92–95; Angew. Chem. Int. Ed. Engl. 1993, 32, 69; d) K. Kobaya-
shi, Y. Yamada, M. Yamanaka, Y. Sei, K. Yamaguchi, J. Am. Chem.
Soc. 2004, 126, 13896; e) see for example: M. Schmittel, A. Ganz,
Chem. Commun. 1997, 99; M. Schmittel, V. Kalsani, D. Fenske, A.
Wiegrefe, Chem. Commun. 2004, 490; M. Schmittel, V. Kalsani,
R. S. K. Kishore, H. Colfen, J. W. Bats, J. Am. Chem. Soc. 2005, 127,
11544–11545.
[10] To take account of van der Waals radii, two further angles were
added to the atom···Sb···atom angles. The tan of each additional
angle is (v/d) for each of the ligand atoms (where v=van der Waals
radius of atom, and d=distance from cage centroid (Sb1) to that
atom. A diagram is given in the Supporting Information.
[11] J. Podlaha, I. Cisarova, D. Alexander, P. Holy, T. Kraus, J. Zavada,
Collect. Czech. Chem. Commun. 2000, 65, 1587.
[12] G. M. Sheldrick, SHELXTL, An Integrated System for Data Collec-
tion, Processing, Structure Solution and Refinement, Bruker Analyt-
ical X-Ray Systems Inc., Madison, WI, 1998.
[13] SMART Software Reference Manual, version 5.054, Bruker Analyt-
ical X-Ray Systems Inc., Madison, WI, 1998.
[5] X. Xu, M. Nieuwenhuyzen, S. L. James, Angew. Chem. 2002, 114,
790–793; Angew. Chem. Int. Ed. 2002, 41, 764–767; E. Lozano, M.
Niewenhuyzen and S. L. James, Chem. Eur. J. 2001, 7, 2644; X. Xu,
M. Niewenhuyzen, M. Chambers, E. MacLean, S. J. Teat, S. L.
Received: August 18, 2005
Published online: December 12, 2005
Chem. Eur. J. 2006, 12, 2448 – 2453
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2453