Solvent-dependent self-assembly behaviour and speciation control of Pd 6L8 metallo-supramolecular cages

Article abstract of DOI:10.1002/chem.201304437

The C₃-symmetric chiral propylated host-type ligands (±)-tris(isonicotinoyl)-tris(propyl)-cyclotricatechylene (L1) and (±)-tris(4-pyridyl-4-benzoxy)-tris(propyl)-cyclotricatechylene (L2) self-assemble with Pd^II into [Pd₆L₈] ¹²⁺ metallo-cages that resemble a stella octangula. The self-assembly of the [Pd₆(L1)₈]¹²⁺ cage is solvent-dependent; broad NMR resonances and a disordered crystal structure indicate no chiral self-sorting of the ligand enantiomers in DMSO solution, but sharp NMR resonances occur in MeCN or MeNO₂. The [Pd ₆(L1)₈]¹²⁺ cage is observed to be less favourable in the presence of additional ligand, than is its counterpart, where L=(±)-tris(isonicotinoyl)cyclotriguaiacylene (L1a). The stoichiometry of reactant mixtures and chemical triggers can be used to control formation of mixtures of homoleptic or heteroleptic [Pd₆L₈] ¹²⁺ metallo-cages where L=L1 and L1a. All sorted: Stella octangular [Pd₆L₈]¹²⁺ metallo-cages with chiral propylated cyclotriveratrylene-type ligands exhibit homochiral self-sorting in MeCN or MeNO₂ but not in DMSO. Furthermore, differences in the favourability of the metallo-cage with propylated or methylated ligands can be used to control speciation in solution from heteroleptic to predominantly single homoleptic to dual homoleptic metallo-cages.

Full text of DOI:10.1002/chem.201304437

DOI: 10.1002/chem.201304437
Full Paper
&
Supramolecular Chemistry
Solvent-Dependent Self-Assembly Behaviour and Speciation
Control of Pd6L8 Metallo-supramolecular Cages
James J. Henkelis, Julie Fisher, Stuart L. Warriner, and Michaele J. Hardie*[a]
Abstract: The C3-symmetric chiral propylated host-type li-
gands (Æ)-tris(isonicotinoyl)-tris(propyl)-cyclotricatechylene
(L1) and (Æ)-tris(4-pyridyl-4-benzoxy)-tris(propyl)-cyclotricate-
chylene (L2) self-assemble with PdII into [Pd6L8]12+ metallo-
cages that resemble a stella octangula. The self-assembly of
the [Pd6(L1)8]12+ cage is solvent-dependent; broad NMR res-
onances and a disordered crystal structure indicate no chiral
self-sorting of the ligand enantiomers in DMSO solution, but
sharp NMR resonances occur in MeCN or MeNO2. The
[Pd6(L1)8]12+ cage is observed to be less favourable in the
presence of additional ligand, than is its counterpart, where
L=(Æ)-tris(isonicotinoyl)cyclotriguaiacylene (L1a). The stoi-
chiometry of reactant mixtures and chemical triggers can be
used to control formation of mixtures of homoleptic or het-
eroleptic [Pd6L8]12+ metallo-cages where L=L1 and L1a.
tangled [M4L4] cube.[10] The largest metallo-cages involving
CTV-type ligands are the [Pd6L8]12+ stella octangula assemblies
that occur with ligands that have 4-pyridyl groups appended,
such as L1a, L1b and L2a (Scheme 1).[9] The crystal structure
of the previously reported [Pd6(L1a)8]12+ stella octangula[9b]
has six PdII cations arranged in an octahedron with the eight
L1a ligands taking up the octahedron’s faces giving a 3 nm
sized cage assembly (Scheme 1). The pyramidal aspect of the
ligands gives the cage a spiked appearance (Scheme 1) similar
to a stella octangula, which is the first stellation of an octahe-
dron. Although crystals of [Pd6(L1a)8]·12NO3 are racemic, each
stella octangula cage is homochiral, being composed of only
one of the two L1a enantiomers.[9b] An analogous, octomeric
CTV-based cube, assembled through labile covalent bonds, has
been reported by Warmuth.[11]
Introduction
The self-assembly of metallo-supramolecular assemblies from
multifunctional ligands and transition metal cations is well es-
tablished and has yielded a variety of cage-like species.[1] These
may have hollow interiors where additional guest molecules or
ions may be bound, and their ability to act as host assemblies
means that a number of metallo-cage systems are being devel-
oped as nano-scale hosts and reaction vessels, with applica-
tions including the trapping of reactive species,[2] enabling un-
usual reactivities and catalysis[3] and templating nanoparticle
formation.[4] Metallo-cages, also known as coordination cages,
are examples of host assemblies where the individual molecu-
lar or ionic components do not necessarily have host-function
themselves, which is a distinction from molecular hosts, which
are individual molecules capable of binding guests. A theme of
our research is the self-assembly of metallo-cages that utilise
ligand-functionalised molecular hosts, in particular those based
on cyclotriveratrylene (CTV). Other types of molecular hosts
have also been employed to this effect, most particularly from
the calixarene family.[5]
The previously reported [Pd6L8]12+ stella octangula cages are
only soluble in dimethylsulfoxide (DMSO), which limits their
potential as nano-scale hosts. In a bid to improve the solubility
of these cages, we targeted the propylated ligands, tris(isonico-
tinoyl)-tris(propyl)-cyclotricatechylene (L1), and tris(4-pyridyl-4-
benzoxy)-tris(propyl)-cyclotricatechylene
(L2),
shown
in
Cyclotriveratrylene is a relatively rigid and pyramidal-shaped
host with an open upper rim.[6] Both CTV and its chiral ana-
logue cyclotriguaiacylene (CTG) can be converted into extend-
ed-armed host molecules through upper rim functionalisation.
Metallo-supramolecular assemblies of ligand functionalised
CTV-analogues include single-cage and double-cage catenating
[M3L2] capsule-like metallo-cryptophanes,[7] [M4L4] and [M6L4]
tetrahedra,[8] [M6L8] stella octangula assemblies,[9] and a self-en-
Scheme 1. The vast majority of known CTV analogues with
mixed upper rim substituents feature either a methoxy or hy-
droxyl group as one of the substituents,[6] examples of other
combinations of mixed upper rim groups are much rarer.[12,13]
Results and Discussion
Propylated-cyclotriguaiacylene 2 (p-CTG) was prepared from
propylated-cyclotriveratrylene 1 (p-CTV),[14] which was deme-
thylated using lithium diphenylphosphide generated in situ by
lithiation of diphenylphosphine with n-butyl lithium
(Scheme 2).[13] Lithium diphenylphosphide selectively demeth-
ylates aryl methyl ethers over other alkyls, and Collet had pre-
viously used the same approach for hetero-functionalisation of
[a] J. J. Henkelis, Dr. J. Fisher, Dr. S. L. Warriner, Prof. M. J. Hardie
School of Chemistry, University of Leeds
Woodhouse Lane, Leeds, LS2 9JT (UK)
E-mail: m.j.hardie@leeds.ac.uk
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201304437.
Chem. Eur. J. 2014, 20, 4117 – 4125
4117
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
the CTV framework.[13] The 1H NMR spectrum of p-CTG dis-
played the characteristic diastereotopic resonances of the
endo- and exo-protons of the tribenzo[a,d,g]cyclononatriene
core at 3.31 and 4.56 ppm ([D6]DMSO), respectively. p-CTG was
highly soluble in common organic solvents and was observed
to act as a gelator for the solvents dichloromethane, chloro-
form, nitromethane, acetonitrile and tetrahydrofuran. Gelator
behaviour of CTV analogues and derivatives has been previous-
ly observed.[15] p-CTG was converted to L1 and L2 using adapt-
ed versions of previously reported syntheses,[9a,16] with each
ligand being obtained as a racemic mixture in high yields ac-
cording to Scheme 2.
The crystal structure of L1 was determined from its clathrate
complex L1·0.5(MeNO2)·1.5(H2O). The asymmetric unit features
one molecule of L1, a MeNO2 disordered across an inversion
centre, and three poorly resolved regions of solvent, modelled
as partial water molecules. L1 deviates from strict molecular
C3-symmetry and all ester groups are oriented with the carbon-
yl groups away from the cavity of the cyclononatriene core
(Figure 1). The closest aromatic separation between the ligands
Scheme 1. Previously reported and target (L1, L2) pyramidal tripodal ligands
with 4-pyridyl donor groups and the crystal structure of [Pd6(L1a)8]12+ from
ref. [9b].
Figure 1. From the crystal structure of complex L1·0.5(MeNO2)·1.5(H2O).
is 4.4 ꢁ, which is too long to suggest any p–p interactions.
There are, however, p–H intermolecular interactions between
the terminal methyl and pyridyl groups of nearby ligands, with
CÀH···pyridine separations of 3.06 ꢁ. The overall crystal lattice
has a bilayer-like arrangement of sheets of L1 ligands separat-
ed by solvent (Figure S4 in the Supporting Information).
Self-assembly of [M6L8]12+ stella octangula cages
Self-assembly of the propylated stella octangula, [Pd6(L1)8]12+
,
was achieved through combination of eight equivalents of
racemic ligand L1 with six equivalents of PdX2 (where X=
À
NO3À, BF4 or CF3CO2À) in a variety of different solvents,
namely DMSO, DMF, MeCN, MeNO2 and, surprisingly, a 9:1 mix-
ture of water/MeCN (the last observed only by mass spectro-
metry). Electrospray mass spectrometry (ESI-MS) studies of so-
lution mixtures of [Pd(MeCN)4](BF4)2 and L1 shows rapid forma-
tion of the stella octangula cage with mass peaks of (m/z)
1949.6343, 1542.0858, 1270.3947 and 1076.7391 being attribut-
Scheme 2. Synthesis of the propylated CTG ligands L1 and L2.
Chem. Eur. J. 2014, 20, 4117 – 4125
www.chemeurj.org
4118
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 2. ESI-MS from mixture of L1 and [Pd(MeCN)4](BF4)2, showing various
{[Pd6(L1)8]·m(BF4)}12Àm+ ions, asterisk indicates a cage-DMSO adduct.
ed to {[Pd6(L1)8]·8BF4}4+, {[Pd6(L1)8]·7BF4}5+, {[Pd6(L1)8]·6BF4}6+
and {[Pd6(L1)8]·5BF4}7+, respectively (Figure 2). Cage-DMSO ad-
ducts were also observed, for instance in the 4+ charge state,
mass peaks of 1948.9994, 1968.5019 and 1988.5064 were at-
tributed to {[Pd6(L1)8]·8BF4}4+, {(DMSO)ꢀ[Pd6(L1)8]·8BF4}4+ and
{(DMSO)2ꢀ[Pd6(L1)8]·8BF4}4+, respectively. Similar mass spectra
were seen in all solvents utilised and were independent of the
counter anion used, and the spectra did not substantially
change when monitored over a period of weeks (Figures S8
and S11 in the Supporting Information).
Figure 3. 1H NMR spectra of ligand L1 (bottom trace) and the corresponding
stella octangula complex [Pd6(L1)8]·12BF4 monitored over a week of stand-
ing the solution: a) [D6]DMSO solution; b) [D3]MeCN (time increases on verti-
cal scale).
2D diffusion-ordered spectroscopic experiments (DOSY) in
DMSO indicated a single large species in solution with a diffu-
sion coefficient of 0.439ꢂ10À10 m2 sÀ1 (Figure S11 in the Sup-
porting Information). Based on the diffusion coefficient of free
ligand, 1.293ꢂ10À10 m2 sÀ1, a Dcomplex/Dligand ratio of 0.33:1 was
established which, through the Stokes–Einstein relationship,
was estimated to give a hydrodynamic radius (r) of 23.4 ꢁ. This
is larger than the hydrodynamic radius of 19.4 ꢁ that was mea-
mediately obtained in [D3]MeNO2, however, the cage formation
was not quantitative as there is also free ligand in solution
(Figure S14 in the Supporting Information). For all solvents,
ESI-MS studies are dominated by the [Pd6(L1)8]12+ species and
do not vary when monitored over days or weeks.
sured for [Pd6(L1a)8]12+ [9b]
consistent with the longer and
,
more conformationally flexible propyl chains.
1
The 1D H NMR spectra of the [Pd6(L1)8]12+ cage indicated
We propose that the broadened spectra are indicative of
a mixture of [Pd6(L1)8]12+ cages, which are not chirally re-
solved, while sharpened spectra indicate chiral self-sorting.
Hence, [Pd6(L1)8]12+ self-sorts rapidly in MeNO2, more slowly
and incompletely in MeCN, but not over a timescale of many
weeks in DMSO or DMF. In the original [Pd6(L1a)8]12+ stella oc-
tangula,[9] sharp 1H NMR spectra are obtained in [D6]DMSO
and the crystal structure of the nitrate salt showed chiral
self-sorting as each [Pd6(L1a)8]12+ cage was crystallographically
ordered and contained only one ligand enantiomer
(Scheme 1).[9b]
that aspects of the solution-phase self-assembly are solvent-
dependent. All 1H NMR spectra show strong coordination-in-
duced downfield shifting of the pyridyl ortho and meta pro-
tons, and upfield shifts of propyl group protons. Those ob-
tained in [D6]DMSO, or [D7]DMF, display broad peaks that do
not sharpen over several weeks of monitoring (Figure 3a and
Figure S9 in the Supporting Information). Heating the
[D6]DMSO solution to 608C for 18 h before monitoring does
not lead to any changes in the spectrum. Heating the PdII pre-
cursor in [D6]DMSO for 1 h prior to cooling to room tempera-
ture then addition of the ligand also results in no variation.
The spectrum obtained in [D3]MeCN, however, initially shows
broad resonances but this resolves to a sharper spectrum over
a period of three days (Figure 3b and Figure S12 in the Sup-
porting Information). The spectra do not sharpen up entirely
and the DOSY NMR in [D3]MeCN indicates the presence of
a second species in solution of very similar size (Figure S13 in
À
Single crystals of the BF4 salt of [Pd6(L1)8]12+ were obtained
from DMSO solution, and X-ray diffraction data were collected
using synchrotron radiation. The structures of [Pd6(L1a)8]12+
and [Pd6(L1)8]12+ are not isomorphic, however, both crystallise
with tetragonal unit cells.[17] The structure of [Pd6(L1)8]·12(BF4)
gives PdII positions in a near octahedral arrangement with
Pd···Pd separations 16.3 ꢁ, comparable with the symmetry and
Pd···Pd separation (16.6 ꢁ) observed for [Pd6(L1a)8]12+. The
1
the Supporting Information). A sharp H NMR spectrum is im-
Chem. Eur. J. 2014, 20, 4117 – 4125
www.chemeurj.org
4119
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
75%, despite heating to 708C, overnight, then standing for
a week (Figure S15 in the Supporting Information). Mass peaks
of (m/z) 1371.6856, 1665.9895 and 2107.4224 were observed
and corresponded to {[Pt6(L1)8]·6ClO4}6+, {[Pt6(L1)8]·7ClO4}5+
and {[Pt6(L1)8]·8ClO4}4+, respectively.
The treatment of eight equivalents of L2 with six equivalents
of [Pd(MeCN)4](BF4)2 in DMSO likewise results in the rapid and
quantitative formation of the Pd6L8 stella octangula (Fig-
ures S16–S18 in the Supporting Information). The 1D NMR
spectrum of the cage in [D6]DMSO is broad, with the pyridyl
ortho-resonances shifted strongly downfield due to the pyridyl-
palladium coordination, again consistent with cage formation
but not self-sorting. DOSY NMR spectroscopy showed the
diffusion of one large species in solution with a diffusion coc-
tjl;efficient of 0.348ꢂ10À10 m2 sÀ1, and ESI-MS gives mass
peaks of (m/z) 910.0709 and 1020.8581 corresponding to
{[Pd6(L2)8]·2BF4}10+ and {[Pd6(L2)8]·3BF4}9+, respectively.
Figure 4. From the crystal structure of [Pd6(L1)8]·12(BF4)·2(H2O). a) Detail
showing disordered ligand bridging between three PdII centres with the su-
perposition of both ligand enantiomers, one enantiomer is shown in ball-
and-stick to highlight; b) the disordered [Pd6(L1)8]12+ stella octangula cage.
bridging L1 ligand, however, is completely disordered within
the structure. The disorder has been modelled such that each
full ligand position is a superposition of both ligand enantio-
mers (Figure 4 and Figure S6 in the Supporting Information).
This is consistent with [Pd6(L1)8]12+ cages forming from a mix-
ture of ligand enantiomers, with the apparent superposition
an average ligand position across all unit cells. The 1H NMR
spectrum of isolated single crystals that were redissolved in
[D6]DMSO is identical to the original spectrum shown in Fig-
ure 3a, supporting the notion that the broadened spectra re-
flect a mixture of cage stereoisomers.
Guest binding and chemical disassembly–reassembly
Preliminary host–guest binding studies of [Pd6(L1)8]12+ moni-
tored by ESI-MS indicates that guest o-carborane association
occurs in DMF solution. Each [Pd6(L1)8]·n(BF4)12Àn charge state
showed multiple guest adducts. For instance, two o-carborane
adducts were observed as part of the 5+ mass-charge enve-
lope, with mass peaks of (m/z) 1541.8719, 1570.4850 and
1598.4942 corresponding to {[Pd6(L1)8]·7BF4}5+, {(carborane)ꢀ-
[Pd6(L1)8]·7BF4}5+ and {(o-carborane)2ꢀ[Pd6(L1)8]·7BF4}5+, re-
spectively. Similar host–guest phenomena were displayed for
the 8, 7 and 6+ charge states. Crystal structures of solid-state
o-carborane and halogenated carborane anion host–guest
complexes with CTVs have been previously shown, and usually
feature CÀH···p hydrogen bonding from the acidic carborane
CÀH group.[22]
Another interpretation is that broadened spectra are indica-
tive of incomplete self-assembly to the symmetrical
[Pd6(L1)8]12+ cages, and that we are observing lower symmetry
variants with the same stoichiometry. Yoneya and co-workers
have recently reported molecular dynamics simulations of the
self-assembly of [Pd6L8] spherical cages where L is the achiral
tripodal ligand 1,3,5-tris(methyl-4-pyridyl)-benzene.[18] In that
study, lower symmetry complexes were found leading up to
the formation of the symmetrical [Pd6L8]. Given the short life-
times of these species, and the high symmetry—albeit disor-
dered—cage found in the crystal structure, we find this to be
a less plausible explanation for this system.
The [Pd6(L1)8]12+ stella octangula cage can be chemically dis-
assembled and reassembled. According to NMR spectroscopy
observations (Figure S20 in the Supporting Information), treat-
ment of the preformed cage with 24 equivalents of 4,4’-dime-
thylaminopyridine (DMAP) resulted in the quantitative disas-
sembly of the cage and generated free L1, alongside [Pd-
(DMAP)4]·2BF4. This occurs as DMAP is a stronger Lewis base
than the pyridyl group due to the inductive effects of the
amine group. The cage was quantitatively reassembled by sub-
sequent addition of 24 equivalents of para-toluenesulfonic acid
(TsOH) to regenerate the [Pd6(L1)8]·12BF4 assembly and H+
-DMAP (Figure S20 in the Supporting Information). This dem-
onstrates scope for application in cargo delivery and the selec-
tive sequestration or release of guests upon initiation by a lo-
calised trigger. The DMAP-TsOH chemical trigger has been pre-
viously used with metallo-cages and in other supramolecular
systems, such as switchable molecular shuttles.[23]
Homochiral self-sorting in metallo-supramolecular assem-
blies, where ligand enantiomers recognise one another from
a racemic mixture, has been previously reported both for stella
octangula cages[9] and for other systems,[19] and includes exam-
ples in which sorting from a stereomeric mixture occurs over
several days.[19c] However, systems in which such self-sorting is
dependent on what solvent is used are a much rarer occur-
rence, and we are unaware of another example that parallels
this one. Solvent-effects have been reported for equilibria be-
tween diastereomers of the atrane structure of a hemicrypto-
phane–oxidovanadium complex,[20] and solvent may affect ste-
reoselectivity in organic synthesis.[21]
A platinum(II) congener, [Pt6(L1)8]·12ClO4, is formed but not
1
in quantitative yields. The H NMR spectrum in [D6]DMSO was
Ligand exchange and speciation control
symmetrical and displayed the characteristic downfield shifts
of the pyridyl resonances, similar to the palladium(II) analogue
described above; yet, due to the decreased lability of the
metal centre, conversion to the cage was measured to be only
The availability of the sterically and interactionally similar
ligand pairs, L1a/L1 and L2b/L2, all of which form a [Pd6L8]12+
cage, allows us to study the formation of [Pd6L8]12+ cages from
Chem. Eur. J. 2014, 20, 4117 – 4125
www.chemeurj.org
4120
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
a mixture of ligands, as well as any ligand exchange that
occurs between the cages.
This is contrasting behaviour to that reported by Dalcanale
and co-workers on similar types of cavitand-based metallocage
systems.[24] They mixed two distinct homoleptic Pt4L2 cages
where the L ligands were pyridyl-derived calix[4]resorcinarene
cavitands with different lower-rim groups, and observed for-
mation of heteroleptic cages in solution on heating and stand-
ing. Zheng and Stang have likewise shown that combinations
of homoleptic Pt-based supramolecular polygons undergo dy-
namic ligand exchange to form mixtures of heteroleptic poly-
gons.[25] Conversely, Fujita and co-workers have remarked on
the high kinetic inertness displayed by much larger Pd12L24
metallo-cages,[26] and slow ligand exchange in metallo-cages
has likewise been reported by both Raymond[27] and Ward.[28]
Ward has also investigated metal exchange in [M4L6]4+ metal-
lo-cages, where M=CoII or CdII. They observed that metal
scrambling began soon after mixing, and a near binomial dis-
tribution of [Co4ÀnCdnL2]4+ species was achieved after 150
days.[29] Our results initially seemed more in keeping with
those of Fujita, and attributable to the larger size of our cage
compared with Dalcanale’s, and the presence of more M-cavi-
tand bonds (24 compared with 8). Additional experiments,
however, revealed that the solution behaviour of the Pd6L8
cages is more complicated, and that under some conditions,
formation of the [Pd6(L1a)8]12+ cage is more favourable than
formation of the [Pd6(L1)8]12+ counterpart.
Heteroleptic [Pd6(L1)8Àn(L1a)n]12+ and [Pd6(L2)8Àn(L2a)n]12+
cages can be formed. The combination of six equivalents of
[Pd(MeCN)4](BF4)2 with four equivalents of each of L1 and L1a
was allowed to stand, overnight. Electrospray mass spectrome-
try indicated heteroleptic [Pd6(L1)8Àx(L1a)x]·12BF4 cage forma-
tion, where each mass-charge envelope for a given charge
state was identified to be a near statistical mixture of ligand
combinations. For example, the mass spectra for the 5+ mass-
charge envelope displayed mass peaks of (m/z) 1424.7916,
1440.9024, 1457.9135, 1474.3383, 1491.3380 and 1510.6050,
which corresponded to {[Pd6(L1)1(L1a)7]·7BF4}5+, {[Pd6(L1)2-
(L1a)6]·7BF4}5+
(L1a)4]·7BF4}5+
,
,
{[Pd6(L1)3(L1a)5]·7BF4}5+
{[Pd6(L1)5(L1a)3]·7BF4}5+
,
{[Pd6(L1)4-
and ([Pd6(L1)6-
(L1a)2]·7BF4}5+, respectively. Once formed, there was no evi-
dence of subsequent ligand exchange over several weeks of
monitoring, and the mass spectra procured were consistent
across DMSO, DMF and MeCN. The 1H NMR spectrum in
[D6]DMSO was broad and was not observed to sharpen. The
formation of a mixture of heteroleptic cages was also observed
from similar experiments with the L2a/L2 ligand pair, and the
ESI-MS obtained is shown in Figure 5.
The addition of eight equivalents of methylated ligand L1a
to the preformed [Pd6(L1)8]12+ assembly saw immediate con-
version to the methylated cage [Pd6(L1a)8]12+. The resultant
1H NMR spectrum is an overlay of the spectra of L1 and
[Pd6(L1a)8]12+ (Figure S28 in the Supporting Information). ESI-
MS shows a clear bias towards L1a-containing cages with the
observation of [Pd6(L1)8Àn(L1a)n]12+ species, where n=4 to 8.
Interestingly, ligand–cage adducts are also observed, which
may indicate that the ligand exchange proceeds through an
associative mechanism. The converse reaction, where pre-
formed methylated [Pd6(L1a)8]12+ is treated with propylated
L1 in DMSO, also shows a bias towards the methylated
[Pd6(L1a)8]12+ cage by ESI-MS and no ligand exchange was ob-
servable by NMR spectroscopy (Figure S29 in the Supporting
Information).
This difference in cage favourability in the presence of addi-
tional ligand according to ligand identity, along with DMAP-in-
duced chemical disassembly, allows us to control the predomi-
nant speciation in solution, summarised in Scheme 3. A mix-
ture of heteroleptic [Pd6(L1)8Àn(L1a)n]·12BF4 cages is generated
from the combination of six equivalents of [Pd(MeCN)4](BF4)2
with four equivalents of each of L1 and L1a, as there is insuffi-
cient L1a in solution to only form the favoured [Pd6(L1a)8]12+
cage. Addition of a further 4 equivalents of each of L1 and
L1a to the same solution results in rapid, more selective cage
formation with the methylated [Pd6(L1a)8]12+ and free propy-
lated ligand L1 observed as the major species in solution by
NMR spectroscopy. The addition of a further six equivalents of
[Pd(MeCN)4](BF4)2 returns the system to a cage mixture, al-
though one that is strongly biased towards the homoleptic
over heteroleptic species according to NMR spectroscopy and
ESI-MS (Figure 6 and Scheme 3, respectively). Disassembly of
Figure 5. ESI-MS showing formation of heteroleptic [Pd6(L2)8Àn(L2a)n] cages
from a 4:4:6 mixture of L2/L2a/[Pd(MeCN)4](BF4)2. Different charge states
show different {[Pd6(L2)8Àn(L2a)n]·m(BF4)}12Àm+ series.
A 1:1 mixture of preformed [Pd6(L1)8]·12(BF4) and [Pd6-
(L1a)8]·12(BF4) in [D6]DMSO was allowed to stand at room tem-
perature to determine whether ligand exchange would occur.
1
ESI-MS and H NMR spectroscopy showed only a 1:1 mixture of
homoleptic cages. There was no evidence of the formation of
any heteroleptic species. Neither heating to 508C, overnight,
nor standing for six months produced observable changes to
the ESI-MS. A mixture of preformed [Pd6(L2)8]·12BF4 and [Pd6-
(L2a)8]·12BF4 showed that, after overnight heating then stand-
ing for two months, limited ligand exchange does occur with
the exchange of up to three ligands, per cage, according to
ESI-MS.
Chem. Eur. J. 2014, 20, 4117 – 4125
www.chemeurj.org
4121
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[Pd6L8]12+ cages have improved
solubility over their previously
reported methyl counterparts.[9]
This is likely to be advantageous
for host–guest studies and pre-
liminary work showed that
[Pd6(L1)8]12+ associates with the
spherical guest o-carborane.
The crystal structure of
[Pd6(L1)8]·12(BF4)·2H2O
grown
from DMSO solution was very
disordered, with superposition
of both enantiomers of the L1
ligand, indicating no chiral self-
sorting of the ligands. The broad
1H NMR spectra for [Pd6(L1)12]12+
in DMSO are also indicative
of a mixture of cage isomers.
This was not the case for
[Pd6(L1)12]12+ in MeCN or MeNO2
in which sharpened spectra indi-
cate self-sorting.
Scheme 3. Ligand speciation control of [Pd6L8]12+ stella octangula cages through: a) stoichiometry; b) chemical
disassembly and reassembly.
Despite the lability of PdÀN
bonds, homoleptic mixtures of preformed [Pd6L8]12+ stella oc-
tangula cages show no, or only very minor, ligand exchange in
DMSO even after many months in solution. This is in keeping
with studies on other large cage systems, where ligand ex-
change was observed to be slow,[26–28] although it is unusual to
observe no exchange at all, as was the case for the
[Pd6(L1)8]12+ and [Pd6(L1a)8]12+ mixture. The cages can be dis-
rupted by addition of more ligand, and a significant degree of
ligand exchange was seen on addition of L1a to [Pd6(L1)8]12+
.
However, the degree of ligand exchange was considerably
smaller when L1 was added to [Pd6(L1a)8]12+. This suggests
that the [Pd6(L1a)8]12+ cage is the favoured product over
[Pd6(L1)8]12+ on addition of excess ligand. The observation of
ligand–cage adducts during ligand exchange experiments sup-
ports the notion that ligand exchange occurs through an asso-
ciative mechanism. The favouring of [Pd6(L1a)8]12+ in the pres-
ence of additional ligand may therefore reflect that L1a has
less sterically demanding methyl groups on its upper rim com-
pared with the more sterically demanding propyl groups of L1.
The formation of near binomial mixtures of heteroleptic
cages from 4:4:6 L1/L1a/PdII mixtures (and from L2/L2a coun-
terpart) is as expected from the solution stoichiometry and in-
dicates kinetic control. Addition of more ligand biases the
system to [Pd6(L1a)8]12+, which is consistent with ligand ex-
change experiments described above. This leaves more L1 in
solution than L1a, which favours a homoleptic cage distribu-
tion on addition of more equivalents of PdII. This allows us to
exercise a degree of control over the predominant species in
solution.
Figure 6. 1H NMR of process of Scheme 3a. Speciation is changed from
a mixture of heteroleptic cages to predominance of [Pd6(L1a)8]12+ on addi-
tion of additional equivalents of both ligands, then to a mixture of homolep-
tic cages on addition of more [Pd(MeCN)4](BF4)2 (marked “PdII added”). The
top spectrum is a mixture of preformed homoleptic cages for comparison.
the 1:1 mixture of homoleptic cages can be affected using
DMAP. Subsequent addition of TsOH reassembles the cages to
once again form a mixture of heteroleptic cages, as the L1/
L1a/PdII stoichiometric ratio in the disassembled solution is
now 4:4:6 (Scheme 3). ESI-MS taken after this disassembly–re-
assembly process also gives evidence of ligand–cage adduct
formation (Figure S34 in the Supporting Information).
Conclusion
The propylated ligands L1 and L2 form [Pd6L8]12+ stella octan-
gula metallo-cage species, and the first PtII stella octangula
[Pt6(L1)8]12+ was synthesised. As anticipated, the propyl
Chem. Eur. J. 2014, 20, 4117 – 4125
www.chemeurj.org
4122
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
808.3232; elemental analysis calcd (%) for C30H36O6·(H2O): C 69.80,
H 5.74, N 5.09; found: C 70.00, H 5.55, N 4.80.
Experimental Section
Synthesis
(Æ)-2,7,12-Tripropoxy-3,8,13-tris(4-pyridyl-4-phenylcarboxy)-
10,15-dihydro-5H-tribenzo[a,d,g]cyclononene
(tris(4-pyridyl-4-
(Æ)-2,7,12-Tripropoxy-3,8,13-trimethoxy-10,15-dihydro-5H-tribenzo-
[a,d,g]cyclononene (1),[14] lithium diphenylphosphide,[13] 4-(4-pyri-
dyl)benzoyl chloride hydrochloride[9a] tris(isonicotinoyl)cyclotri-
guaiacylene (1a)[16] and tris(4-pyridyl-4-benzoxy)cyclotriguaiacylene
(2a)[9b] were prepared according to literature methods. Reagents
were obtained from commercial sources and were used as re-
ceived.
benzoxy)-tris(propyl)-cyclotricatechylene) (L2): Anhydrous tri-
ethylamine (1.32 mL, 7.56 mmol) was added to a stirred solution of
2 (310 mg, 0.630 mmol) in anhydrous THF (50 mL), at À788C,
under an argon atmosphere. After 1 h, 4-(4-pyridyl)benzoylchloride
hydrochloride (960 mg, 3.78 mmol) was added to the reaction mix-
ture and stirred at À788C for a further 2 h before being left at RT
for 48 h. The solvent was removed in vacuo and the resultant resi-
due was triturated in ethanol to afford the target compound as
a white solid, which was isolated by filtration and dried in vacuo.
(Æ)-2,7,12-Tripropoxy-3,8,13-trihydroxy-10,15-dihydro-5H-
tribenzo[a,d,g]cyclononene (2): (Æ)-2,7,12-Tripropoxy-3,8,13-trime-
1
Yield 609 mg, 93%. M.p. decomposes >2708C; H NMR (300 MHz,
thoxy-10,15-dihydro-5H-tribenzo[a,d,g]cyclononene
(1;
2.01 g,
[D6]DMSO, 258C): d=8.70 (d, J=6.2 Hz, 6H; Py-H2), 8.20 (d, J=
8.5 Hz, 6H, Ph-H3), 8.03 (d, J=8.5 Hz, 6H; Ph-H2), 7.81 (d, J=6.2 Hz,
6H; Py-H3), 7.53 (s, 3H; aryl-H), 7.34 (s, 3H; aryl-H), 4.87 (d, J=
12.5 Hz, 3H; CTG exo-H), 3.94 (t, J=6.2 Hz, 6H; propyl S-H), 3.73
(d, J=12.5 Hz, 3H; CTG endo-H), 1.53 (q, J=6.6 Hz, 6H; propyl b-
H), 0.78 ppm (t, J=7.4 Hz, 9H; propyl Y-H); 13C{1H} NMR (75 MHz,
[D6]DMSO, 258C): d=184.2, 150.7, 148.6, 145.7, 142.2, 138.4, 136.6,
131.8, 130.3, 129.2, 127.2, 121.5, 69.7, 21.8, 9.9 ppm; IR (FT-IR): n˜ =
2960, 1734 (s), 1594, 1508, 1400, 1263 (s), 1181, 1093, 820,
762 cmÀ1; HRMS (ESI+): m/z calcd for [M+Na]+: 1058.3993; found:
1058.3941; elemental analysis calcd (%) for C66H57N3O9·0.5(CHCl3): C
72.88, H 5.29, N 3.83; found: C 73.15, H 5.40, N 3.80.
3.74 mmol) and anhydrous THF (10 mL) were added to a flame-
dried Schlenk tube and stirred vigorously. Lithium diphenylphos-
phide was added dropwise via cannulae transfer over 2 h, during
which time it decolourised. The reaction mixture was stirred, over-
night, and solidified. The resultant lithium phenoxide was hydro-
lysed with concentrated aq. HCl and volatiles were removed in
vacuo. Organics were extracted into dichloromethane (6ꢂ100 mL)
and then back-extracted with 6m aqueous sodium hydroxide (6ꢂ
100 mL). The sodium hydroxide layer was washed with dichlorome-
thane (4ꢂ100 mL) and acidified with 6m aqueous HCl to precipi-
tate the desired product as an off-white solid. The solid was al-
lowed to stand for 1 h before being filtered, washed with water
(2ꢂ50 mL) and dried. Subsequent dissolution of the solid in
chloroform, filtration through a silica pad and evaporation of the
solution afforded the title compound as a colourless glass. Yield
[Pd6(L1)8]·12(BF4)·n(MeCN) stella octangula: [Pd(MeCN)4](BF4)2
(5.0 mg, 0.0113 mmol) and L1 (12.10 mg, 0.0150 mmol) were dis-
solved in [D3]MeCN (~2 mL) and stirred for 1 h, resulting in a pale-
1
1
974 mg, 55%. M.p. decomposes >2708C; H NMR (300 MHz, CDCl3,
yellow solution, whereby H NMR spectroscopy displayed quantita-
258C, TMS): d=8.52 (s, 3H; phenol), 6.82 (s, 3H; aryl-H), 6.80 (s,
3H; aryl-H), 4.55 (d, J=13.4 Hz, 3H; CTG exo-H), 3.86 (t, J=6.6 Hz,
6H; propyl S-H), 3.31 (d, J=13.4 Hz, 3H; CTG endo-H), 1.69 (q, J=
7.2 Hz, 6H; propyl b-H), 0.96 ppm (t, J=7.4 Hz, 9H; propyl Y-H);
13C{1H} NMR (75 MHz, CDCl3, 258C, TMS): d=145.2, 145.0, 132.6,
130.4, 116.7, 115.3, 70.2, 35.0, 22.1, 10.4 ppm; IR (FT-IR): n˜ =3550–
3110 (broad), 2945, 2910, 1645, 1485, 1390 cmÀ1; HRMS (ESI+): m/z
calcd for [M+Na]+: 515.2410; found: 515.2410; elemental analysis
calcd (%) for C30H36O6·0.5(H2O): C 71.83, H 7.43; found: C 72.15, H
7.35.
tive cage formation. Diffusion of diethyl ether vapour into the solu-
tion afforded small, yellow prisms that were isolated, washed with
a portion of diethyl ether and dried in vacuo. Quantitative H NMR
(300 MHz, [D3]MeCN, 258C): d=9.31 (m, 1H; Py-H2, achiral cage),
9.16 (d, 5H; Py-H2, chiral cage), 8.13 (d, 6H; Py-H3), 7.27 (s, 3H;
aryl-H), 7.10 (s, 3H; aryl-H), 4.84 (d, 3H; CTG exo-H), 3.87 (m, 6H;
propyl S-H), 3.69 (d, 3H; CTG endo-H), 1.62 (m, 1H; propyl b-H),
1.40 (m, 5H; propyl b-H), 0.85 (m, 2H; propyl Y-H), 0.60 ppm (m,
7H; propyl Y-H); IR (FT-IR): n˜ =3494, 2968, 2901, 1751, 1619, 1508,
1
1270 cmÀ1 (s); HRMS (ESI+): m/z 1076.7391 {[Pd6L8]·5BF4}7+
,
1270.3947
{[Pd6L8]·6BF4}6+
,
1542.0858{[Pd6L8]·7BF4}5+
and
(Æ)-2,7,12-Tripropoxy-3,8,13-tris(4-pyridylcarboxy)-10,15-dihy-
1949.6343 {[Pd6L8]·8BF4}4+; satisfactory elemental analysis could
not be obtained owing to high levels of solvation.
dro-5H-tribenzo[a,d,g]cyclononene
(tris(isonicotinoyl)-tris-
(propyl)-cyclotricatechylene) (L1): Anhydrous triethylamine
(2.4 mL, 13.56 mmol) was added to a stirred solution of 2 (555 mg,
1.13 mmol) in anhydrous THF (150 mL), at À788C, under an argon
atmosphere. After 1 h, isonicotinoyl chloride hydrochloride
(800 mg, 4.50 mmol) was added to the reaction mixture and stirred
at À788C for a further 2 h before being left at RT for 48 h. A
second portion of isonicotinoyl chloride hydrochloride (800 mg,
4.50 mmol) was added, and left to stir for a further 48 h, during
which time the reaction mixture discoloured. The solvent was re-
moved in vacuo and the resultant residue triturated in ethanol to
afford the target compound as a white solid, which was isolated
by filtration and dried in vacuo. Yield 640 mg, 66%. M.p. decom-
[Pt6(L1)8]·12(ClO4)
0.00928 mmol) was added to
stella
octangula:
Pt(ClO4)2
(3.66 mg,
a
solution of L1 (10.12 mg,
0.0124 mmol) in [D6]DMSO (1 mL) and stirred at 708C, overnight.
1H NMR spectroscopy on the cooled solution displayed partial cage
formation (~75% based on relative integrals). 1H NMR (300 MHz,
[D6]DMSO, 258C): d=9.53–9.44 (bm, 3H; Py-H2), 9.16 (d, 3H; Py-
H2), 8.88 (d; free L1), 8.33–8.25 (bm, 6H; Py-H3), 7.96 (d; free L1),
7.55 (s, 3H; aryl-H), 7.32 (s, 3H; aryl-H), 4.89 (bd, 3H; CTG exo-H),
3.93 (bm, 6H; propyl S-H), 3.70 (bd, 3H; CTG endo-H), 1.53 (q; free
L1), 1.30 (bq, 6H; propyl b-H), 0.75 (t; free L1), 0.53 ppm (m, 9H;
propyl Y-H); HRMS (ESI+): m/z 1371.6856 {[Pt6L8]·6ClO4}6+
1665.9895 {[Pt6L8]·7ClO4}5+ and 2107.4224 {[Pt6L8]·8ClO4}4+
,
1
.
poses >2708C; H NMR (500 MHz, CDCl3, 258C, TMS): d=8.84 (d,
J=6.0 Hz, 6H; Py-H2), 7.97 (d, J=6.0 Hz, 6H; Py-H3), 7.16 (s, 3H;
aryl-H), 6.94 (s, 3H; aryl-H), 4.82 (d, J=13.6 Hz, 3H; CTG exo-H),
3.93 (t, J=6.2 Hz, 6H; propyl S-H), 3.67 (d, J=13.6 Hz, 3H; CTG
endo-H), 1.66 (q, J=7.4 Hz, 6H; propyl b-H), 0.86 ppm (t, J=7.2 Hz,
9H; propyl Y-H); 13C{1H} NMR (75 MHz, CDCl3, 258C, TMS): d=162.6,
149.6, 148.2, 138.9, 137.9, 137.5, 131.9, 123.8, 123.6, 115.5, 69.8,
34.9, 21.7, 10.1 ppm; IR (FT-IR): n˜ =3100, 2875, 1745 (strong), 1605,
1520 cmÀ1; HRMS (ESI+): m/z calcd for [M+H]+: 808.3234; found:
[Pd6(L2)8]·12(BF4) stella octangula: [Pd(MeCN)4](BF4)2(3.2 mg,
0.00725 mmol) and L2 (10.00 mg, 0.00966 mmol) were dissolved in
[D6]DMSO (~2 mL) and stirred for 1 h, resulting in a pale-yellow so-
lution, where both 1D and 2D 1H NMR spectroscopy displayed
cage formation. Diffusion of acetone vapour into the solution af-
forded a microcrystalline solid which was isolated, washed with
a portion of acetone and dried in vacuo. 1H NMR (300 MHz,
[D6]DMSO, 258C): d=9.34 (bm, 6H; Py-H2), 8.24–8.08 (bm, 18H; Py-
Chem. Eur. J. 2014, 20, 4117 – 4125
www.chemeurj.org
4123
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
H3, Ph-H2, Ph-H3), 7.46 (s, 3H; aryl-H), 7.30 (s, 3H; aryl-H), 4.82 (bd,
3H; CTG exo-H), 3.84 (bm, 6H; propyl S-H), 3.69 (bd, 3H; CTG
endo-H), 1.38 (bq, 6H; propyl b-H), 0.61 ppm (bt, 9H; propyl Y-H);
IR (FT-IR): n˜ =3384 (broad), 1742 (weak), 1622 (weak), 1024 cmÀ1
(weak); HRMS (ESI+): m/z 1020.9683 {[Pd6L8]·3BF4}9+, 1159.3551
M. W. Johnson, R. G. Bergman, K. N. Raymond, J. Am. Chem. Soc. 2011,
133, 11964–119466; c) M. Yoshizawa, J. K. Klosterman, M. Fujita, Angew.
Chem. 2009, 121, 3470–3490; Angew. Chem. Int. Ed. 2009, 48, 3418–
3438; d) Z. Yan, T. McCracken, S. Xia, V. Maslak, J. Galluci, C. M. Hadad,
J. D. Badjic, J. Org. Chem. 2008, 73, 355–363; e) M. Yoshizawa, M.
Tamura, M. Fujita, Science 2006, 312, 251–254.
{[Pd6L8]·4BF4}8+
,
1337.3494
{[Pd6L8]·5BF4}7+
,
1574.9554
[4] T. Ichijo, S. Sato, M. Fujita, J. Am. Chem. Soc. 2013, 135, 6786–6789.
[5] For examples and reviews, see: a) S. Pasquale, S. Sattin, E. C. Escudero-
Adꢃn, M. Martꢄnez-Belmonte, J. de Mendoza, Nat. Commun. 2012, 3,
785; b) T. Schrçder, S. N. Sahu, J. Mattay, Top. Curr. Chem. 2012, 319, 99–
124; c) R. Pinalli, F. Boccini, E. Dalcanale, Isr. J. Chem. 2011, 51, 781–797;
d) P. Jin, S. J. Dalgarno, J. L. Atwood, Coord. Chem. Rev. 2010, 254, 1760–
1768; e) T. Haino, M. Kobayashi, Y. Fukuzawa, Chem. Eur. J. 2006, 12,
3310–3319; f) H. Jude, D. J. Sinclair, N. Das, M. S. Sherburn, P. J. Stang, J.
Org. Chem. 2006, 71, 4155–4163; g) O. D. Fox, J. Cookson, E. J. S. Wilkin-
son, M. G. B. Drew, E. J. MacLean, S. J. Teat, P. D. Beer, J. Am. Chem. Soc.
2006, 128, 6990–7002; h) R. M. McKinlay, G. W. V. Cave, J. L. Atwood,
Proc. Natl. Acad. Sci. USA 2005, 102, 5944–5948; i) E. Menozzi, M. Busi,
R. Ramingo, M. Campagnolo, S. Geremia, E. Dalcanale, Chem. Eur. J.
2005, 11, 3136–3148; j) K. Kobayashi, Y. Yamada, M. Yamanaka, Y. Sei, K.
Yamaguchi, J. Am. Chem. Soc. 2004, 126, 13896–13897; k) L. Baldini, P.
Ballester, A. Casnati, R. M. Gomila, C. A. Hunter, F. Sansone, R. Ungaro, J.
Am. Chem. Soc. 2003, 125, 14181–14189; l) O. D. Fox, N. K. Dalley, R. G.
Harrison, J. Am. Chem. Soc. 1998, 120, 7111–7112.
[6] For review, see: M. J. Hardie, in Supramolecular Chemistry: from Mole-
cules to Nanomaterials, Vol. 3 (Eds.: P. A. Gale, J. W. Steed), 2012, Wiley,
pp. 895–916.
[7] a) J. J. Henkelis, T. K. Ronson, L. P. Harding, M. J. Hardie, Chem. Commun.
2011, 47, 6560–6562; b) T. K. Ronson, H. Nowell, A. Westcott, M. J.
Hardie, Chem. Commun. 2011, 47, 176–178; c) A. Westcott, J. Fisher,
L. P. Harding, P. Rizkallah, M. J. Hardie, J. Am. Chem. Soc. 2008, 130,
2950–2951; d) Z. Zhong, A. Ikeda, S. Shinkai, S. Sakamoto, K. Yamagu-
chi, Org. Lett. 2001, 3, 1085–1087.
{[Pd6L8]·6BF4}6+ and 1906.9543 {[Pd6L8]·7BF4}5+; satisfactory ele-
mental analysis could not be obtained due to high levels of solva-
tion.
X-ray crystallography
Crystals were mounted on MiTeGen loops under oil and flash
frozen under N2. Data were collected on a Bruker X8 diffractometer
with MoKa radiation (l=0.71073 ꢁ) or on a Rigaku Saturn diffrac-
tometer with synchrotron radiation (l=0.6889 ꢁ) at station I19 of
the Diamond Light Source. Data were corrected for absorption
using a multiscan method, and structures were solved by direct
methods and refined by full-matrix least squares on F2 using the
SHELX suite of programs,[30] interfaced through the program X-
Seed.[31] Summaries of refinements are given below, full details are
available in the Supporting Information.
L1·0.5(MeNO2)·1.5(H2O): C48.5H50.5N3.5O11.5
: Mr =866.42, triclinic,
¯
space group P1, a=13.506(2), b=15.49593), c=16.120(3) ꢁ, a=
62.596(8), b=65.374(8), g=64.841(8)8; V=2606.2(8) ꢁ3; Z=2;
qmax =22.988; data/restraints/parameters: 6798/4/572; R1(obs.
data)=0.1777.
[Pd6(L1)8]·12(BF4)·6(H2O): C228H228B12F48N24O54Pd6: Mr =6629.54, tet-
ragonal, space group I4/mmm, a=30.688(5), c=45.906(11) ꢁ; V=
43234(15) ꢁ3; Z=2; qmax =20.008; data/restraints/parameters: 6101/
0/90; R1(obs. data)=0.1577. Structure showed significant disorder
and most aromatic rings were refined with a rigid body constraint,
propyl groups and anions were not located in the difference map
(in the Supporting Information). There were large solvent-accessi-
ble voids hence the data were treated with the SQUEEZE routine
of PLATON.[32]
[8] a) B. F. Abrahams, N. J. FitzGerald, R. Robson, Angew. Chem. 2010, 122,
2958–2961; Angew. Chem. Int. Ed. 2010, 49, 2896–2899; b) C. J. Sumby,
M. J. Carr, A. Franken, J. D. Kennedy, C. A. Kilner, M. J. Hardie, New J.
Chem. 2006, 30, 1390–1396; c) C. J. Sumby, J. Fisher, T. J. Prior, M. J.
Hardie, Chem. Eur. J. 2006, 12, 2945–2959.
[9] a) T. K. Ronson, C. Carruthers, J. Fisher, T. Brotin, L. P. Harding, P. J. Rizkal-
lah, M. J. Hardie, Inorg. Chem. 2010, 49, 675–685; b) T. K. Ronson, J.
Fisher, L. P. Harding, M. J. Hardie, Angew. Chem. 2007, 119, 9244–9246;
Angew. Chem. Int. Ed. 2007, 46, 9086–9088.
CCDC-955885
(L1·0.5(MeNO2)·1.5(H2O))
and
-971406
([Pd6(L1)8]·12(BF4)·6(H2O)) contain the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[10] T. K. Ronson, J. Fisher, L. P. Harding, P. J. Rizkallah, J. E. Warren, M. J.
Hardie, Nat. Chem. 2009, 1, 212–216.
[11] X. Xu, R. Warmuth, J. Am. Chem. Soc. 2008, 130, 7520–7521.
[12] For example a) A. Kraszewska, P. Rivera-Fuentes, G. Rapenne, J. Cras-
sous, A. G. Petrovic, J. L. Alonso-Gomez, E. Huerta, F. Diederich, C. Thil-
gen, Eur. J. Org. Chem. 2010, 4402–4411; b) T. Brotin, V. Roy, J.-P. Dutas-
ta, J. Org. Chem. 2005, 70, 6187–6195; c) A. Arduini, F. Calzavacca, D.
Demuru, A. Pochini, A. Secchi, J. Org. Chem. 2004, 69, 1386–1388;
d) R. M. Thomas, D. S. Iyengar, Synth. Commun. 1999, 29, 2507–2509.
[13] a) J. Canceill, L. Lacombe, A. Collet, J. Chem. Soc. Chem. Commun. 1987,
219–221; b) J. Canceill, A. Collet, J. Gabard, G. Gottarelli, G. P. Spada, J.
Am. Chem. Soc. 1985, 107, 1299–1308.
[14] R. Ahmad, M. J. Hardie, Supramol. Chem. 2006, 18, 29–38.
[15] a) A. Westcott, C. J. Sumby, R. D. Walshaw, M. J. Hardie, New J. Chem.
2009, 33, 902–912; b) D. Bardelang, F. Camerel, R. Ziessel, M. Schmutz,
M. J. Hannon, J. Mater. Chem. 2008, 18, 489–494; c) Y. Kubo, W. Yoshizu-
mi, T. Minami, Chem. Lett. 2008, 37, 1238–1239.
Acknowledgements
We thank the EPSRC and the Brotherton Trust (JJH) for fund-
ing, I. Blakeley for microanalysis and N. Cookson for helpful dis-
cussions. This work was carried out with the support of the Di-
amond Light Source for time in station I19 under proposal
mt7847.
Keywords: coordination cage
· ligand exchange · self-
[16] M. J. Hardie, C. J. Sumby, Inorg. Chem. 2004, 43, 6872–6874.
[17] Unit cell of [Pd6(L1a)8]·12(NO3) tetragonal, space group P4/n, a=
32.5130(6), c=31.5740(13) ꢁ.[9b]
assembly · self-sorting · supramolecular chemistry
[1] For reviews, see: a) S. R. Seidel, P. J. Stang, Acc. Chem. Res. 2002, 35,
972–983; b) M. Fujita, K. Umemoto, M. Yoshizawa, N. Fujita, T. Kusuka-
wa, K. Biradha, Chem. Commun. 2001, 509–518; c) D. L. Caulder, K. N.
Raymond, Acc. Chem. Res. 1999, 32, 975–982.
[2] For example: a) P. Mal, B. Breiner, K. Rissanen, J. R. Nitschke, Science
2009, 324, 1697; b) M. Yoshizawa, T. Kusukawa, M. Fujita, K. Yamaguchi,
J. Am. Chem. Soc. 2000, 122, 6311.
[18] M. Yoneya, T. Yamaguchi, S. Sato, M. Fujita, J. Am. Chem. Soc. 2012, 134,
14401–14407.
[19] For example: a) C. Gꢅtz, R. Hovorka, G. Schnakenburg, A. Lꢅtzen, Chem.
Eur. J. 2013, 19, 10890–10894; b) M. M. Safont-Sempere, G. Fernꢃndez,
F. Wꢅrthner, Chem. Rev. 2011, 111, 5784–5814; c) C. G. Claessens,
I. Sꢃnchez-Molina, T. Torres, Supramol. Chem. 2009, 21, 44–47;
d) U. Kiehne, T. Weilandt, A. Lꢅtzen, Eur. J. Org. Chem. 2008, 2056–
2064; e) I.-W. Hwang, T. Kamada, D. M. Ko, T. Nakamura, A. Tsuda, A.
Osuka, D. Kim, J. Am. Chem. Soc. 2004, 126, 16187–16198; f) A. Lꢅtzen,
[3] For example: a) Z. J. Wang, K. N. Clary, R. G. Bergman, K. N. Raymond,
F. D. Toste, Nat. Chem. 2013, 5, 100–103; b) C. J. Brown, G. M. Miller,
Chem. Eur. J. 2014, 20, 4117 – 4125
www.chemeurj.org
4124
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
M. Hapke, J. Griep-Raming, D. Haase, W. Saak, Angew. Chem. 2002, 114,
2190–2194; Angew. Chem. Int. Ed. 2002, 41, 2086–2089; g) J.-M. Vin-
cent, C. Philouze, I. Pianet, J.-B. Verlhac, Chem. Eur. J. 2000, 6, 3595–
3599; h) M. A. Masood, E. J. Enemark, T. D. P. Stack, Angew. Chem. 1998,
110, 973–977; Angew. Chem. Int. Ed. 1998, 37, 928–932.
[25] Y.-R. Zheng, P. J. Stang, J. Am. Chem. Soc. 2009, 131, 3487–3489.
[26] S. Sato, Y. Ishido, M. Fujita, J. Am. Chem. Soc. 2009, 131, 6064–6065.
[27] A. J. Terpin, M. Ziegler, D. W. Johnson, K. N. Raymond, Angew. Chem.
2001, 113, 161–164; Angew. Chem. Int. Ed. 2001, 40, 157–160.
[28] M. D. Ward, Chem. Commun. 2009, 4487–4499.
[20] A. Martinez, V. Robert, H. Gornitzka, J.-P. Dutasta, Chem. Eur. J. 2010, 16,
520–527.
[21] G. Cainelli, P. Galletti, D. Giacomini, Chem. Soc. Rev. 2009, 38, 990–1001.
[22] R. J. Blanch, M. Williams, G. D. Fallon, M. G. Gardiner, R. Kaddour, C. L.
Raston, Angew. Chem. 1997, 109, 520–522; Angew. Chem. Int. Ed. Engl.
1997, 36, 504–506.
[29] B. R. Hall, L. E. Manck, I. S. Tidmarsh, A. Stephenson, B. F. Taylor, E. J.
Blaikie, D. A. van der Griend, M. D. Ward, Dalton Trans. 2011, 40, 12132–
12145.
[30] G. M. Sheldrick, Acta Crystallogr. Sect. A 2008, 64, 112–122.
[31] L. J. Barbour, Supramol. Chem. 2001, 1, 189–191.
[32] A. L. Spek, Acta Crystallogr. Sect. A 1990, 46, C34.
[23] a) J. E. M. Lewis, E. L. Gavey, S. A. Cameron, J. D. Crowley, Chem. Sci.
2012, 3, 778–784; b) J. D. Crowley, D. A. Leigh, P. J. Lusby, R. T. McBur-
ney, L.-E. Perret-Aebi, C. Petzold, A. M. Z. Slawin, M. D. Symes, J. Am.
Chem. Soc. 2007, 129, 15085–15090.
Received: November 12, 2013
[24] L. Pirondini, F. Bertolini, B. Cantadori, F. Ugozzoli, C. Massera, E. Dalca-
nale, Proc. Natl. Acad. Sci. USA 2002, 99, 4911–4915.
Published online on March 3, 2014
Chem. Eur. J. 2014, 20, 4117 – 4125
www.chemeurj.org
4125
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim