Ligand Aspect Ratio as a Decisive Factor for the Self-Assembly of Coordination Cages

Article abstract of DOI:10.1021/jacs.5b13190

It is possible to control the geometry and the composition of metallasupramolecular assemblies via the aspect ratio of their ligands. This point is demonstrated for a series of iron- and palladium-based coordination cages. Functionalized clathrochelate co

Full text of DOI:10.1021/jacs.5b13190

Article
pubs.acs.org/JACS
Ligand Aspect Ratio as a Decisive Factor for the Self-Assembly of
Coordination Cages
†
,¶
†,¶
†
†
§
Suzanne M. Jansze, Giacomo Cecot, Matthew D. Wise, Konstantin O. Zhurov, Tanya K. Ronson,
§
†
#,⊥
†
†
‡
Ana M. Castilla, Alba Finelli, Philip Pattison, Euro Solari, Rosario Scopelliti, Genrikh E. Zelinskii,
‡
‡
,§
,†
Anna V. Vologzhanina, Yan Z. Voloshin, Jonathan R. Nitschke,* and Kay Severin*
†
́
́ ́ ́
Institut des Sciences et Ingenierie Chimiques, Ecole Polytechnique Federale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland
§
‡
#
Department of Chemistry, University of Cambridge, Cambridge CB2 1EW, United Kingdom
Nesmeyanov Institute of Organoelement Compounds of the Russian Academy of Sciences, 119991 Moscow, Russia
Laboratory of Crystallography, Ecole Polytechnique Fed
́
er
Swiss-Norwegian Beamline, ESRF, 38000 Grenoble, France
S Supporting Information
́
ale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland
⊥
*
ABSTRACT: It is possible to control the geometry and the composition
of metallasupramolecular assemblies via the aspect ratio of their ligands.
This point is demonstrated for a series of iron- and palladium-based
coordination cages. Functionalized clathrochelate complexes with variable
II
aspect ratios were used as rod-like metalloligands. A cubic Fe L cage
8
12
was obtained from a metalloligand with an intermediate aspect ratio. By
increasing the length or by decreasing the width of the ligand, the self-
II
assembly process resulted in the clean formation of tetrahedral Fe L
4
6
cages instead of cubic cages. In a related fashion, it was possible to control
II
the geometry of Pd -based coordination cages. A metalloligand with a
II
large aspect ratio gave an entropically favored tetrahedral Pd L
4
8
II
assembly, whereas an octahedral Pd L cage was formed with a ligand of the same length but with an increased width. The
6
12
II
aspect ratio can also be used to control the composition of dynamic mixtures of Pd cages. Out of two metalloligands with only
II
II
marginally different aspect ratios, one gave rise to a self-sorted collection of Pd L and Pd L cages, whereas the other did not.
4
8
6 12
INTRODUCTION
make heteroleptic assemblies by mixing cis-blocked L MX
2 2
■
complexes (M = Pd or Pt; X = weakly coordinating anion) with
The structural outcomes of metallasupramolecular self-
assembly processes are largely controlled by the nature of the
building blocks, i.e., metal ions, ligands, and possibly
1
a,3
neutral N-donor ligands and anionic carboxylate ligands.
2
+
Taken together, the homoleptic [L M(N-donor) ] and
2
2
1
L M(O CR) complexes are thermodynamically less stable
2 2 2
templates. From a design perspective, the ligand is arguably
+
than the mixed [L M(N-donor)(O CR)] complex, resulting in
2
2
the most important building block, because chemists can use
the repertoire of synthetic chemistry to make a vast collection
of ligands with diverse structures and properties.
the clean formation of heteroleptic assemblies.
Strategic functionalization of ligands with bulky groups can
also be used to control the self-assembly behavior. For example,
it is possible to favor heteroleptic complexes over homoleptic
complexes by using two ligands, one of which has bulky groups
By choosing an appropriate ligand, it is possible to influence
the size, the geometry, and the functionality of the resulting
metallasupramolecular assembly. Several factors are known to
be of importance in this context. Structural rigidity of the ligand
is a prerequisite for the assembly of polynuclear assemblies,
because flexible ligands tend to favor complexes of low
nuclearity. For rigid ligands, the distance and the relative
orientation of the donor atoms are key parameters, allowing
control of the size and geometry of the final assembly. This
point is nicely illustrated by work from the Fujita group, who
have explored the assembly of spherical coordination cages
4
in the vicinity of the donor atom(s). In this case, the
coordination of two bulky ligands to the same metal center is
thermodynamically disfavored, pushing the system toward the
formation of heteroleptic complexes. Steric interactions that are
remote from the metal binding site have also been employed as
5
an element of control, but this strategy is less explored.
The investigation described herein deals with an under-
appreciated parameter in ligand design: the aspect ratio of a
rigid rod-type ligand. We show that the length-to-width ratio of
a ligand can be used to control the geometries of coordination
II
2
from bent dipyridyl ligands and Pd ions. By slight variation of
the bend angle, they were able to make Pd 12L24 or Pd 24L48
cages in a controlled fashion. The nature of the donor atom
II
II
(
e.g., oxygen vs nitrogen) and the charge of the ligand are
Received: December 17, 2015
likewise important parameters. For example, it is possible to
©
XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b13190
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
cages. In particular, we demonstrate that entropically disfavored
cubic and octahedral cages are formed instead of tetrahedra if
the aspect ratio of the ligand is reduced (Scheme 1).
via the boronate ester cap, and the lateral size of the ligand can
be varied by choosing an appropriate dioxime ligand. The iron
center in clathrochelate complexes shows a distorted trigonal
prismatic coordination geometry. Therefore, clathrochelates
Scheme 1. Rigid Rod-Type Ligands with Different Aspect
Ratios Give Rise to Different Coordination Cages
display approximate C symmetry when viewed along the B···B
3
axis. Functionalized clathrochelate complexes are thus intrinsi-
cally 3-dimensional metalloligands with rigid rod geometry.
For our investigation, we have focused on three types of
coordination cages which are known to form with rigid rod
II
ligands. Fe 2nL3n cages can be obtained from ligands with two
terminal formylpyridine groups upon reaction with amino-
II
12,13
benzene derivatives and an Fe salt (Scheme 3a).
In similar
II
II
Scheme 3. Fe - and Pd -Based Coordination Cages with
Rigid Rod-Type Ligands
Furthermore, we have investigated the self-sorting behavior of
ligands with variable aspect ratios. Minor differences in the
length-to-width ratio were found to have a pronounced effect
on the compositions of dynamic mixtures of cages.
RESULTS AND DISCUSSION
■
Clathrochelate Complexes as Metalloligands with a
Variable Aspect Ratio. Clathrochelate complexes capped
with boronate ester groups are formed by a metal-templated6
condensation reaction of a dioxime ligand with a boronic acid.
II
When Fe is employed as the metal template, the resulting
complexes are diamagnetic, robust, and inert. Such clathroche7-
late complexes can easily be converted into metalloligands.
Different functional groups can be introduced by choosing an
8
appropriate boronic acid (in particular, arylboronic acids), and
the lateral size and the solubility can be modulated by variation
of the dioxime building block (Scheme 2). We note that
numerous boronic acids and some dioximes are commercially
available, and that the yields of clathrochelate syntheses are
typically good.
Scheme 2. Clathrochelate Complexes as Robust, Flexible,
and Easy-to-Access Metalloligands
fashion, FeII2nL3n cages are formed when ligands bearing
terminal aniline groups react with 2-formylpyridine in the
II
12,14
presence of Fe salts (Scheme 3b).
The assembly of
II
Pd 2nL4n cages can also be achieved by the combination of
linear ligands with terminal 3-pyridyl groups (Scheme
c).
In order to synthesize a clathrochelate ligand with apical
formylpyridine groups, we used an acetal-protected boronic
10,15,16
3
17
acid, the synthesis of which has been reported previously.
Reactions with dimethylglyoxime, nioxime, or diethylglyoxime
We and others have recently shown that clathrochelate
complexes with apical pyridyl groups can be used as N-donor
ligands in metallasupramolecular chemistry. Different molecula9r
architectures have been synthesized, including macrocycles,
cages, and metal−organic frameworks. Furthermore,
pyridyl-capped clathrochelates have been used to bridge
and FeCl in methanol produced protected clathrochelates. The
2
yields for these reactions are modest (21, 33, and 29%,
respectively) because side products are formed by a
protodeboronation reaction. Clean deprotection of the acetal
was achieved by a reaction with p-toluenesulfonic acid using
microwave heating to give the target clathrochelates 1, 2, and 3
(Scheme 4).
1
0
9b
heterometallic Cr Ni rings. These complexes display interesting
7
1
1
magnetic behavior.
For the syntheses of 4-aminophenyl-terminated clathroche-
lates, we also employed a protecting group strategy. Starting
Clathrochelate-based metalloligands appear to be ideally
suited to investigate the influence of the aspect ratio on self-
assembly behavior. The length of the ligands can be modulated
1
8
from (4-(dibenzylamino)phenyl)boronic acid, we first
II
synthesized the Fe clathrochelate complexes, which were
B
DOI: 10.1021/jacs.5b13190
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 4. Synthesis of the Clathrochelates 1, 2, and 3
Figure 1. Molecular structures of the clathrochelates 1 (top), 4·HCl
(middle), and 5·HCl (bottom) in the crystal. A stick representation
(side view, left) and a space-filling representation (view along the B···B
axis, right) are given for each complex. Chloride anions are omitted for
clarity. Gray: C; white: H; dark blue: N; green: B; red: O; and orange:
Fe.
a
Reagents and conditions: (i) MeOH, reflux, 4 h; (ii) TsOH (0.1
equiv), CHCl /t-BuOH (4:1), 140 °C, 4 h.
3
subsequently deprotected through hydrogenation to give
complexes 4 and 5 with overall yields of 53% and 73%,
respectively (Scheme 5).
pseudo C symmetry of the complexes. Furthermore, it is
3
The new clathrochelates 1−5 were characterized in solution
by NMR spectroscopy (Figures S11−S20) and high-resolution
mass spectrometry. In addition, we have analyzed the molecular
structure of 1 and of the HCl adducts of 4 and 5 in the solid
state by single-crystal X-ray crystallography. The structures are
shown in Figure 1. The view along the B···B axis reveals the
evident that the lateral size of the nioxime-based complex 5 is
significantly larger than that of the dimethylglyoxime-based
complexes 1 and 4.
The 3-pyridyl-capped clathrochelates 6−8 can be obtained
by reaction of commercial 3-pyridylboronic acid with FeCl and
2
the respective dioxime ligand (Scheme 6). We have recently
a
a
Scheme 5. Synthesis of the Clathrochelates 4 and 5
Scheme 6. Synthesis of the Clathrochelates 6−8
a
Conditions: (i) MeOH, reflux, 4 h (6 and 7); TFA, reflux, 6 h (8).
II
shown that Pd L octahedral coordination cages are formed
6
12
II 10
upon combination of metalloligand 6 with Pd . In order to
probe if the assembly process can be influenced by the aspect
ratio of clathrochelate-based ligands, we have synthesized the
new ligands 7 and 8. Clathrochelate 7 is slightly larger than the
nioxime-based 6 because the six ethyl groups of 7 require more
space than the three conformationally fixed -(CH ) - rings of 6.
2
4
a
Reagents and conditions: (i) MeOH, reflux, 4 h; (ii) H , Pd/C,
Clathrochelate 8, in contrast, is the thinnest (largest aspect
ratio) of the three metalloligands 6−8. The difference in
2
MeOH/CHCl (1.5:1), 50 °C, 4 h.
3
C
DOI: 10.1021/jacs.5b13190
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
equatorial steric bulk of 7 and 8 is evident when comparing the
Scheme 7. Synthesis of Coordination Cages 9−11
structures in the solid state (Figure 2).
Figure 2. Molecular structures of the clathrochelates 7 (top) and 8
(
bottom) in the crystal. A stick representation (side view, left) and a
space-filling representation (view along the B···B axis, right) are given
for each complex. Gray: C; white: H; dark blue: N; green: B; red: O;
and orange: Fe.
II
Fe -Based Coordination Cages. The coordination cages
9
−11 were formed by heating an acetonitrile solution
containing clathrochelate 1, 2, or 3 (3 equiv); p-toluidine (6
equiv); and Fe(OTf) or Fe(NTf ) (2 equiv) at 50 °C for 18 h
2
2 2
(
Scheme 7). Workup allowed isolation of the cages in high
1
yield (>90%). The H NMR spectra of the products were very
complex, in particular those of cages 10 and 11 (Figures S23,
S25, S27, and S29), which showed broad and ill-resolved peaks.
Such behavior is not unexpected because this type of cage is
often observed as a mixture of diastereoisomers with different
12,19
relative stereochemistries of the iron centers at the corners.
Furthermore, the cages may not provide sufficient space for free
rotation of the clathrochelate core. Restricted ligand rotation on
the NMR time scale would lead to further reduction of the
apparent symmetry. The DOSY NMR spectra of 9 (Figure
S22), 10 (Figures S24 and S26), and 11 (Figures S28 and S30)
revealed the presence of large assemblies with a uniform
diffusion constant. However, the diffusion constant of 9 (4.43 ×
−6
2
−6
−6
a
10
cm /s) was larger than those of 10 (10a: 2.89 × 10
Conditions: 1, 2, or 3 (3 equiv), Fe(OTf) or Fe(NTf ) (2 equiv), p-
2 2 2
2
−6
2
cm /s; 10b: 2.85 × 10 cm /s) and 11 (11a: 2.36 × 10
toluidine (6 equiv), CH CN, 50 °C, 18 h.
3
2
−6
2
cm /s; 11b: 2.62 × 10 cm /s), even though ligands of the
same lengths were employed (1−3). This result provides
evidence that complexes 10 and 11 are not typical tetrahedral
coordination cages, but larger assemblies.
II
II
II
II
other metal ions in the vertices (M = Cu , Zn , Ni , or Co )
have also been studied.
template effects
freedom, resulting in variable coordinate vectors. In our case,
22,23
These cages were obtained using
2
3a
High-resolution ESI mass spectrometry revealed that cage 9
or ligands with some conformational
8
+
has the composition [Fe (L1) ] (Figure 3). The tetrahedral
4
6
geometry is in line with what has been observed for other
assemblies based on linear, rigid rod-type ligands with terminal
the coordinate vectors of L2 and L3 should favor the formation
II
of a tetrahedral Fe L cage. The fact that we observe an
4
6
12,13
II
formylpyridine groups.
hand, have the composition [Fe (L2/3) ] (Figure 3 for 11b,
The cages 10 and 11, on the other
entropically disfavored Fe L cage instead of a tetrahedron
8 12
1
6+
can be attributed to the small aspect ratio of ligands L2 and L3.
Their lateral size is significantly larger than that of L1, and a
tetrahedral geometry would lead to unfavorable steric
interactions between the lateral -(CH ) - or Et groups.
8
12
and S56−S66 for both).
We infer 10 and 11 to possess an approximately cubic
structure, with eight iron(II) centers at its vertices. Cubic
2
4
II
coordination cages with Fe (N,N′-chelate) complexes as
Next, we investigated the formation of coordination cages
starting from the aniline-terminated clathrochelates 4 and 5.
The lateral size of these complexes is the same as for 1 and 2
3
20,21
vertices have already been described.
These complexes
are often based on four-fold-symmetric, face-capping ligands,
II
II
and have an Fe L stoichiometry. A cubic Fe L cage has
(Me and -(CH ) - side chains), but condensation with 2-
8
6
8
12
2 4
also been reported, but the coordinate vectors of its ligands are
oriented at 120° with respect to each other, in contrast with
the parallel coordinate vectors of L2 and L3. M8L12 cages with
formylpyridine results in ligands with an increased distance
between the two N,N′-chelating sites (cf. Scheme 3a and 3b).
The aspect ratio of the metalloligands is thus increased.
21
D
DOI: 10.1021/jacs.5b13190
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 8. Synthesis of the Coordination Cages 12 and 13
a
Conditions: 4 or 5 (3 equiv), Fe(OTf) or Fe(NTf ) (2 equiv), p-
2
2 2
toluidine (6 equiv), CH CN, 50 °C, 18 h.
3
Figure 3. High-resolution ESI MS spectra of the tetrahedral cage 9
(
top) and of the cubic cage 11b (middle), along with zoom-ins on the
peaks at 1677 and 1957 m/z (bottom). The simulated spectra are
placed above the experimental data.
When 4 or 5 was combined with 2-formylpyridine and
Fe(OTf) or Fe(NTf ) , DOSY spectroscopy (Figures S32,
2
2 2
S34, and S36) indicated all products diffused at indistinguish-
1
able rates. The H NMR spectra of cages 12 and 13 showed
several sets of signals, suggesting the presence of diastereo-
12,19
isomers (Figures S31, S33, and S35).
Clear evidence for the
stoichiometry of the cages was again obtained by high-
resolution mass spectrometry (Figures S67−S69): both cages
II
have the composition Fe L (Scheme 8).
4
6
The decanuclear cages 12b and 13 were analyzed by single-
crystal X-ray diffraction (Figure 4). Both cages crystallize as a
racemic mixture of ΔΔΔΛ and ΛΛΛΔ isomers, with cage 12b
showing a pseudo C symmetry axis while cage 13 displays full
3
II
crystallographic C symmetry. The four Fe (N,N′-chelate)
3
3
complexes in the vertices have an average distance of about
Figure 4. Molecular structures of the cages 12b and 13 in the crystal.
Hydrogen atoms, counteranions, and solvent molecules are omitted
for clarity. Gray: C; dark blue: N; green: B; red: O; and orange: Fe.
3
2
0 Å. The cavity size of cage 12b is approximately 480 Å as
determined by VOIDOO calculations. For cage 13, the cavity is
3
considerably smaller (377 Å ) because the lateral -(CH ) -
2
4
groups block part of the space inside the cavity (Figure S83).
II
The formation of Fe L cages with both aniline-terminated
controlling the geometry. In fact, ligand L5 (the condensation
product of 5 and 2-formylpyridine) and ligand L2 have the
same lateral size, but L5 has an increased aspect ratio because
4
6
clathrochelates is consistent with the inference that aspect ratio,
and not the absolute width of the ligand, is the decisive factor in
E
DOI: 10.1021/jacs.5b13190
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
the metal binding sites are farther apart. As a result of this
II
II
difference, we observe an Fe L cage with L2 but an Fe L
8
12
4 6
cage with L5.
II
Pd -Based Coordination Cages. The results obtained
II
with Fe -based coordination cages prompted us to investigate
whether the aspect ratio of a ligand can also influence the
assembly of other types of cages. To address this question, we
II
have focused on cages that can be obtained from Pd salts and
dipyridyl ligands. As mentioned above, we have previously
shown that the combination of 3-pyridyl-capped clathrochelate
II
12+
10
6
with Pd gives an octahedral [Pd (6) ] cage. Metal-
6 12
loligand 6 has nearly the same aspect ratio as ligand 7. We were
thus not surprised that an octahedral cage (14) was also formed
when 7 was mixed with [Pd(CH CN) ](BF ) in acetonitrile
3
4
4 2
(Scheme 9).
Scheme 9. Synthesis of the Coordination Cages 14, 15, and
a
1
7
Figure 5. Molecular structure of cage 14 in the crystal. Hydrogen
atoms, counteranions, and solvent molecules are omitted for clarity.
Gray: C; dark blue: N; green: B; red: O; pink: Pd; and orange: Fe.
is not due to restricted rotational freedom of the bridging
clathrochelate ligands. Instead, the data indicate the presence of
two kinds of ligands, which are magnetically distinct. The
DOSY NMR spectrum of 15 in deuterated DMSO (Figure
S52) confirmed the presence of only one type of assembly with
−7
2
a diffusion constant of 7.64 × 10 cm /s. This value is larger
than what was observed for 14 (6.14 × 10− cm /s; DMSO-d )
7
2
6
(
Figure S50), suggesting that 15 is a smaller assembly. This
point was confirmed by high-resolution mass spectrometry,
8
+
which showed that cage 15 has the formula [Pd (8) ] . The
4
8
presence of a tetrahedral cage is in line with the apparent
reduction of symmetry, because a tetrahedral cage should
feature two doubly bridged Pd(8) Pd links in addition to four
2
singly bridged Pd(8)Pd connections. This geometry was indeed
observed when cage 15 was analyzed by single-crystal X-ray
diffraction (Figure 6). The four Pd atoms form a slightly
distorted tetrahedron, with the Pd···Pd distances for the doubly
a
Conditions: CH CN, 70 °C, 3 h (12); DMF, 70 °C, 4 d (13);
3
DMSO, 90 °C, 10 min (15).
Cage 14 was characterized by NMR spectroscopy (Figures
S47−S50), high-resolution mass spectrometry (Figure S70),
and single-crystal X-ray crystallography (Figure 5). The average
Pd···Pd distance within a triangular face was found to be 16.0 Å,
a value which is very similar to what was observed for
1
2+ 10
[
Pd (6) ]
.
These first results suggest that the metal-
6
12
loligands 6 and 7 behave very similarly in self-assembly
reactions. However, we show in the next section that the small
difference in the aspect ratios of 6 and 7 can have a rather
II
pronounced effect on the dynamic behavior of Pd cages.
The reaction of metalloligand 8 with [Pd(CH CN) ](BF )
3
4
4 2
resulted in the clean formation of coordination cage 15.
1
Analysis by H NMR spectroscopy revealed a reduced
symmetry: instead of one set of signals for the bridging
clathrochelate ligands as found for 14, we observed two sets of
signals for 15 (Figure S51). Since the two sets of signals have
equal intensity, we infer that the apparent symmetry reduction
Figure 6. Molecular structure of cage 15 in the crystal. Hydrogen
atoms, counteranions, and solvent molecules are omitted for clarity.
Gray: C; dark blue: N; green: B; red: O; pink: Pd; and orange: Fe.
F
DOI: 10.1021/jacs.5b13190
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
bridged Pd(8) Pd units being slightly smaller than the distances
2
for the singly bridged Pd(8)Pd connections (14.7 vs 15.5 Å).
II
The distance between the two encapsulated Fe ions in the
doubly bridged Pd(8) Pd is only 5.7 Å, resulting in a tight
2
packing of the two clathrochelate frameworks. For the more
bulky ligand 7, the formation of doubly bridged Pd(7) Pd links
2
is expected to be unfavorable, thereby pushing the system
toward an octahedral geometry which shows exclusively singly
bridged Pd(7)Pd connections.
Overall, the structure of cage 15 resembles that of cage 17,
II
which is obtained by combination of “naked” Pd and the
simple phenylene-bridged ligand 16 (Scheme 9). This type of
cage had previously been described by the group of Fujita,
15
albeit with different counteranions. As in the case of 15, a
tetrahedral geometry is observed, with two doubly bridged
1
edges, resulting in a double set of H NMR signals for the
ligand protons (Figure S53).
The different structures that we have observed for octahedral
cage 14 on the one hand, and for the tetrahedral cages 15 and
1
7 on the other hand, provide further evidence that it is
possible to control the geometry of coordination cages via the
aspect ratio of the ligand. The difference between 14 and 15 is
particularly remarkable because the cages are both based on
clathrochelate ligands having exactly the same length, but just a
different width.
1
Figure 7. H NMR spectra in CD CN of (a) [Pd (6) ](BF ) , (b)
3
6
12
4 12
[Pd (16) ](BF ) , and (c) a mixture of [Pd(MeCN) ](BF ) and both
ligands 6 and 16 after heating for 8 days at 50 °C.
4
8
4
8
4
4 2
Self-Sorting of Pd-Based Coordination Cages. The
observations described in the last section suggest that there
exists a fine thermodynamic balance between potential products
II
within a simple supramolecular system consisting of Pd ions
and bis(3-pyridyl) ligands, which can be decisively tipped in
one direction or another by modulating a single property of the
donor building block: equatorial steric bulk. In order to better
understand the thermodynamic factors at play, scrambling
experiments were performed, in which presynthesized samples
of [Pd (6) ](BF ) or [Pd (7) ](BF ) (14) and
6
12
4
12
6
12
4 12
[
Pd (16) ](BF ) (17) were combined and heated to 70 °C
4 8 4 8
for 4 days. Given that the self-assembly behaviors of 6 and 7 are
essentially identical (both form Pd L octahedra), we were
II
6
12
struck by the dramatic differences in the outcomes of these two
1
simple experiments. The H NMR spectrum of a mixture of
[
Pd (6) ](BF ) and [Pd (16) ](BF ) did not significantly
6 12 4 12 4 8 4 8
change during heating, confirming that no heteroleptic
products were formed by ligand scrambling (Figure S72).
1
Conversely, the H NMR spectrum of a mixture of [Pd (7) ]-
6
12
(
BF ) and [Pd (16) ](BF ) became complex, suggesting that
4 12 4 8 4 8
a library of structures had formed (Figure S73).
To corroborate these results, we performed mixed-ligand
self-sorting experiments. Equimolar amounts of ligands 6 and
1
Figure 8. H NMR spectra in CD CN of (a) [Pd (7) ](BF ) , (b)
3
6
12
4 12
[Pd (16) ](BF ) , and (c) a mixture of [Pd(MeCN) ](BF ) and both
4
8
4
8
4
4 2
1
6, or 7 and 16, along with a corresponding amount of
ligands 7 and 16 after heating for 8 days at 50 °C.
[
Pd(MeCN) ](BF ) , were combined in CD CN and allowed
4
4 2
3
to reach thermodynamic equilibrium by heating for 8 days at 50
C. The results obtained from these experiments were in
agreement with those of the scrambling experiments described
[Pd (7) (16)] , [Pd (7) (16) ] , [Pd (7) (16) ]12+,
1
2+
12+
°
6 11 6 10 2 6 9 3
1
2+
8+
[Pd (7) (16) ] , and [Pd (7) (16) ] were all observed
6
8
4
4
5
7
II
1
above. In the presence of Pd ions, the H NMR spectrum of a
mixture of ligands 6 and 16 exclusively showed two distinct sets
of peaks that clearly corresponded to the assemblies
(Figures S79 and S80).
The observation that comprehensive narcissistic self-sorting
occurs following mixture of ligand 6, ligand 16, and Pd ,
whereas a multicomponent library of products is observed in
II
1
2+
8+
[
Pd (6) ] and [Pd (16) ] (Figure 7). Such behavior can
6
12
4
8
24
II
be described as “narcissistic self-sorting”.
In contrast, the H NMR spectrum of a mixture of ligands 7
and 16 in the presence of Pd ions showed multiple sets of
the case of ligand 7, ligand 16, and Pd , is remarkable for two
1
reasons. First, the difference between the aspect ratios of
metalloligands 6 and 7 is small, essentially arising due to the
greater rotational freedom of the ethyl side chains of 7 when
compared to the -(CH ) - groups of 6. The molecular masses
II
peaks, indicating that a library of products had formed (Figure
8). The presence of heteroleptic assemblies was confirmed by
2
4
−1
−1
ESI MS analysis, wherein peaks corresponding to
of 6 (654.09 g mol ) and 7 (660.13 g mol ) differ by only
G
DOI: 10.1021/jacs.5b13190
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
%. Second, the structural differences between the ligands 6
Article
1
same coordinate vector orientation and length, and differ only
very slightly in their equatorial steric bulk (-(CH ) - vs 2 ×
-C H ). The pyridine-induced disassembly was more pro-
and 7 are found far from the pyridine donor groups.
2
4
We hypothesize that the origin of the different self-sorting
behavior lay in the relative thermodynamic stability of the
2 5
nounced for the cage featuring a slightly thicker ligand (smaller
1
2+
12+
assemblies [Pd (6) ] and [Pd (7) ] . If the former were
aspect ratio). This difference in stability was found to manifest
6
12
6
12
II
less stable than the latter, this would facilitate the formation of a
itself in the self-sorting behavior of dynamic mixtures of Pd
4
L
8
8
+
II
II
library of products in the presence of [Pd (16) ] .
and Pd
6
L12 cages. Using the less stable Pd L12 cage gave rise
to a complex mixture of products, whereas narcissistic self-
sorting was observed for the more stable Pd L cage.
6
4
8
Disassembly experiments were undertaken as follows in order
II
1
2+
to test this hypothesis. A solution of [Pd (6) ]
or
6 12
6
12
1
2+
Overall, our results provide clear evidence that the aspect
ratio of a rigid rod-type ligand can be an important parameter
for metal-based self-assembly reactions. In order to use this
parameter as an element of control, it is desirable to employ a
ligand which allows modulation of its aspect ratio without too
much synthetic effort. Clathrochelate-based metalloligands
appear ideally suited for this purpose because their lengths,
widths, and functionalities can be varied easily. Further studies
with this versatile class of ligands will be reported in due course.
[
Pd (7) ]
in CD CN was treated with 24 equiv of
6 12
3
pyridine-d (Scheme 10).
5
Scheme 10. Differences in Thermodynamic Stability Are
Revealed in Destruction Experiments with Pyridine-d5
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.5b13190.
■
*
S
1
After equilibration for 24 h, the H NMR spectrum of the
1
2+
solution containing [Pd (6) ] showed the signals of the
6
12
assembly itself, along with small signals corresponding to free
Experimental procedures, analytical data of the ligands
and the cages ( H, C, DOSY, HRMS), and further
experimental details (PDF)
X-ray crystallography data for 1 (CIF; CCDC 1442394)
X-ray crystallography data for 4 (CIF; CCDC 1442248)
X-ray crystallography data for 5 (CIF; CCDC 1442090)
X-ray crystallography data for 7 (CIF; CCDC 1442226)
X-ray crystallography data for 8 (CIF; CCDC 1442089)
X-ray crystallography data for 12b (CIF; CCDC
ligand 6 (Figure S81). Integration of the signals revealed that
1
13
1
only 20% of the cage had disassembled. In contrast, the H
NMR spectrum of the solution initially containing
12+
[
Pd (7) ] indicated the near-complete destruction of the
6 12
cage. The main signals in the spectrum are attributed to the free
ligand 7, and several minor signals are likely due to Pd
complexes containing both pyridine-d and ligand 7 (Figure
5
S82). These results confirm that there is indeed a pronounced
difference in thermodynamic stability between the cages
1
442396)
12+
12+
[
Pd (6) ]
and [Pd (7) ] , even though the ligands 6
6
12
6 12
X-ray crystallography data for 13 (CIF; CCDC 1442397)
X-ray crystallography data for 14 (CIF; CCDC1415027)
X-ray crystallography data for 15 (CIF; CCDC 1442395)
and 7 are structurally very similar. During self-sorting
1
2+
experiments, the steric strain of [Pd (7) ] is mitigated by
6
12
the incorporation of ligand 16 into heteroleptic octahedral
complexes, as evidenced by the ESI MS data (Figures S78 and
S80).
AUTHOR INFORMATION
Corresponding Authors
*jrn34@cam.ac.uk
kay.severin@epfl.ch
Author Contributions
S.J. and G.C. contributed equally.
■
CONCLUSION
■
*
The size and the geometry of a ligand are of pivotal importance
for metal-based self-assembly reactions. However, the ligand
aspect ratio is a parameter which is rarely discussed in this
context. We have shown that the aspect ratio can be used as an
element of control during the formation of coordination cages.
For two different types of cages, we have observed a switch
toward a higher nuclearity structure when the length-to-width
¶
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
The work was supported by the Swiss National Science
II
■
ratio of the ligand was reduced. In the case of Fe -based cages,
II
we have obtained an unusual Fe L structure instead of the
8
12
Foundation, by the Ecole Polytechnique Fed
EPFL), and by the Marie Curie Initial Training Networks
́
er
́
ale de Lausanne
II
4
common Fe L tetrahedron when a ligand with a small aspect
6
(
“
II
ratio was employed. In the case of Pd -based cages, we have
found that we can form Pd L or Pd L cages, depending on
ReAd” and “Dynamol”. We are grateful to the Swiss-
Norwegian Beamline Consortium for providing access to
synchrotron radiation.
II
II
4
8
6 12
the aspect ratio of the ligand. Intramolecular steric interactions
between the ligands are likely responsible for these changes in
geometry. The reduced intraligand interactions in the larger
cages are enthalpically favorable, thereby compensating for the
entropic penalty associated with the formation of assemblies of
higher nuclearity.
REFERENCES
■
(
1) (a) Cook, T. R.; Stang, P. J. Chem. Rev. 2015, 115, 7001−7045.
b) Lifschitz, A. M.; Rosen, M. S.; McGuirk, C. M.; Mirkin, C. A. J.
(
Am. Chem. Soc. 2015, 137, 7252−7261. (c) Han, Y.-F.; Jin, G.-X. Acc.
Chem. Res. 2014, 47, 3571−3579. (d) Young, N. J.; Hay, B. P. Chem.
Commun. 2013, 49, 1354−1379. (e) Amouri, H.; Desmarets, C.;
Moussa, J. Chem. Rev. 2012, 112, 2015−2041. (f) Chakrabarty, R.;
Mukherjee, P. S.; Stang, P. J. Chem. Rev. 2011, 111, 6810−6918.
Even minor differences in ligand aspect ratio can have a
distinct effect on the thermodynamic stability of a coordination
cage. This fact was demonstrated by disassembly experiments
II
with two Pd L cages. The ligands of these cages have the
6
12
H
DOI: 10.1021/jacs.5b13190
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(
2
g) Wiester, M. J.; Ulmann, P. A.; Mirkin, C. A. Angew. Chem., Int. Ed.
011, 50, 114−137. (h) Jin, P.; Dalgarno, S. J.; Atwood, J. L. Coord.
Chem. Rev. 2010, 254, 1760−1768. (i) Therrien, B. Eur. J. Inorg. Chem.
(17) Sørensen, A.; Castilla, A. M.; Ronson, T. K.; Pittelkow, M.;
Nitschke, J. Angew. Chem., Int. Ed. 2013, 52, 11273−11277.
(18) Kalinin, A.; Lehnemann, B.; Meudt, A.; Scherer, S.; Sniekus, V.
European Patent EP 1479686 A1.
2
009, 2445−2453. (j) Saalfrank, R. W.; Maid, H.; Scheurer, A. Angew.
Chem., Int. Ed. 2008, 47, 8794−8824. (k) Tranchemontagne, D. J.; Ni,
Z.; O’Keeffe, M.; Yaghi, O. M. Angew. Chem., Int. Ed. 2008, 47, 5136−
5
(19) (a) Castilla, A. M.; Ramsay, W. J.; Nitschke, J. R. Acc. Chem. Res.
2014, 47, 2063−2073. (b) Castilla, A. M.; Ramsay, W. J.; Nitschke, J.
R. Chem. Lett. 2014, 43, 256−263.
147. (l) Dalgarno, S. J.; Power, N. P.; Atwood, J. L. Coord. Chem. Rev.
008, 252, 825−841.
2
́
(20) (a) Ramsay, W. J.; Szczypinski, F. T.; Weissman, H.; Ronson, T.
(
2) Harris, K.; Fujita, D.; Fujita, M. Chem. Commun. 2013, 49, 6703−
K.; Smulders, M. M. J.; Rybtchinski, B.; Nitschke, J. R. Angew. Chem.,
Int. Ed. 2015, 54, 5636−5640. (b) Ramsay, W. J.; Ronson, T. K.;
Clegg, J. K.; Nitschke, J. R. Angew. Chem., Int. Ed. 2013, 52, 13439−
6
712.
3) Mukherjee, S.; Mukherjee, P. S. Chem. Commun. 2014, 50, 2239−
248.
4) (a) Saha, M. L.; Neogi, S.; Schmittel, M. Dalton Trans. 2014, 43,
815−3843. (b) Zhao, L.; Northrop, B. H.; Zheng, Y.-R.; Yang, H.-B.;
Lee, H. J.; Lee, Y. M.; Park, J. Y.; Chi, K.-W.; Stang, P. J. J. Org. Chem.
008, 73, 6580−6586. (c) Yoshizawa, M.; Nakagawa, J.; Kumazawa,
(
2
13443. (c) Otte, M.; Kuijpers, P. F.; Troeppner, O.; Ivanovic-
Burmazovic, I.; Reek, J. N. H.; de Bruin, B. Chem. - Eur. J. 2013, 19,
́
(
́
3
10170−10178. (d) Meng, W.; Breiner, B.; Rissanen, K.; Thoburn, J.
D.; Clegg, J. K.; Nitschke, J. R. Angew. Chem., Int. Ed. 2011, 50, 3479−
3483.
2
K.; Nagao, M.; Kawano, M.; Ozeki, T.; Fujita, M. Angew. Chem., Int.
Ed. 2005, 44, 1810−1813. (d) Yoshizawa, M.; Nagao, M.; Kumazawa,
K.; Fujita, M. J. Organomet. Chem. 2005, 690, 5383−5388. (e) Grote,
Z.; Scopelliti, R.; Severin, K. Angew. Chem., Int. Ed. 2003, 42, 3821−
(21) Browne, C.; Brenet, S.; Clegg, J. K.; Nitschke, J. R. Angew.
Chem., Int. Ed. 2013, 52, 1944−1948.
(
(
22) Ward, M. D. Chem. Commun. 2009, 4487−4499.
23) (a) Riddell, I. A.; Hristova, Y. R.; Clegg, J. K.; Wood, C. S.;
Breiner, B.; Nitschke, J. R. J. Am. Chem. Soc. 2013, 135, 2723−2733.
b) Stephenson, A.; Ward, M. D. Dalton Trans. 2011, 40, 10360−
0369. (c) Najar, A. M.; Tidmarsh, I. S.; Adams, H.; Ward, M. D.
3
825.
5) Johnson, A. M.; Wiley, C. A.; Young, M. C.; Zhang, X.; Lyon, Y.;
Julian, R. R.; Hooley, R. J. Angew. Chem., Int. Ed. 2015, 54, 5641−5645.
6) Voloshin, Y. Z.; Kostromina, N. A.; Kramer, R. K. Clathrochelates.
Synthesis, Structure and Properties; Elsevier: Amsterdam, 2002.
7) (a) Li, L.; Fanna, D. J.; Shepherd, N. D.; Lindoy, L. F.; Li, F. J.
(
1
(
Inorg. Chem. 2009, 48, 11871−11881. (d) Argent, S. P.; Adams, H.;
Harding, L. P.; Ward, M. D. Dalton Trans. 2006, 542−544. (e) Bell, Z.
R.; Harding, L. P.; Ward, M. D. Chem. Commun. 2003, 2432−2433.
(
̈
(
(24) (a) He, Z.; Jiang, W.; Schalley, C. A. Chem. Soc. Rev. 2015, 44,
Inclusion Phenom. Macrocyclic Chem. 2015, 82, 3−12. (b) Wise, M. D.;
Severin, K. Chimia 2015, 69, 191−195.
7
79−789. (b) Isaacs, L. Acc. Chem. Res. 2014, 47, 2052−2062. (c) Lal
Saha, M.; Schmittel, M. Org. Biomol. Chem. 2012, 10, 4651−4684.
(
8) (a) Zelinskii, G. E.; Belov, A. S.; Lebed, E. G.; Vologzhanina, A.
(d) Safont-Sempere, M. M.; Fernan
́
dez, G.; Wu
̈
rthner, F. Chem. Rev.
V.; Novikov, V. V.; Voloshin. Inorg. Chim. Acta 2016, 440, 154−164.
2
011, 111, 5784−5814.
(
b) Lebed, E. G.; Belov, A. S.; Dolganov, A. V.; Vologzhanina, A. V.;
Szebesczyk, A.; Gumienna-Kontecka, E.; Kozlowski, H.; Bubnov, Y.
N.; Dubey, I. Y.; Voloshin, Y. Z. Inorg. Chem. Commun. 2013, 30, 53−
5
7. (c) Lebed, E. G.; Belov, A. S.; Dolganov, A. V.; Vologzhanina, A.
V.; Novikov, V. V.; Kuznetsov, E. V.; Voloshin, Y. Z. Inorg. Chem.
Commun. 2013, 33, 57−62.
(
9) (a) Zhang, Y.-Y.; Lin, Y.-J.; Jin, G.-X. Chem. Commun. 2014, 50,
2
327−2329. (b) Wise, M. D.; Ruggi, A.; Pascu, M.; Scopelliti, R.;
Severin, K. Chem. Sci. 2013, 4, 1658−1662.
10) Wise, M. D.; Holstein, J. J.; Pattison, P.; Besnard, C.; Solari, E.;
Scopelliti, R.; Bricogne, G.; Severin, K. Chem. Sci. 2015, 6, 1004−1010.
11) Ardavan, A.; Bowen, A. M.; Fernandez, A.; Fielding, A. J.;
(
(
Kaminski, D.; Moro, F.; Muryn, C. A.; Wise, M. D.; Ruggi, A.;
McInnes, E. J. L.; Severin, K.; Timco, G. A.; Timmel, C. R.; Tuna, F.;
Whitehead, G. F. S.; Winpenny, R. E. P. npj Quantum Inf. 2015, 1,
1
(
5012.
12) (a) Zarra, S.; Wood, D. M.; Roberts, D. A.; Nitschke, J. R. Chem.
Soc. Rev. 2015, 44, 419−432. (b) Smulders, M. M. J.; Riddell, I. A.;
Browne, C.; Nitschke, J. R. Chem. Soc. Rev. 2013, 42, 1728−1754.
(
c) Ronson, T. K.; Zarra, S.; Black, S. P.; Nitschke, J. R. Chem.
Commun. 2013, 49, 2476−2490.
(
13) For selected examples see: (a) Roberts, D. A.; Pilgrim, B. S.;
Cooper, J. D.; Ronson, T. K.; Zarra, S.; Nitschke, J. R. J. Am. Chem.
Soc. 2015, 137, 10068−10071. (b) Riddell, I. A.; Ronson, T. K.;
Nitschke, J. R. Chem. Sci. 2015, 6, 3533−3537. (c) Wood, C. S.;
Ronson, T. K.; Belenguer, A. M.; Holstein, J. J.; Nitschke, J. R. Nat.
Chem. 2015, 7, 354−358.
(
14) For selected examples see: (a) Ronson, T. K.; Roberts, D. A.;
Black, S. P.; Nitschke, J. R. J. Am. Chem. Soc. 2015, 137, 14502−14512.
b) Foster, J. A.; Parker, R. M.; Belenguer, A. M.; Kishi, N.; Sutton, S.;
Abell, C.; Nitschke, J. R. J. Am. Chem. Soc. 2015, 137, 9722−9729.
c) Wood, D. M.; Meng, W.; Ronson, T. K.; Stefankiewicz, A. R.;
Sanders, J. K. M.; Nitschke, J. R. Angew. Chem., Int. Ed. 2015, 54,
988−3992. (d) Ronson, T. K.; League, A. B.; Gagliardi, L.; Cramer,
C. J.; Nitschke, J. R. J. Am. Chem. Soc. 2014, 136, 15615−15624.
15) Chand, D. K.; Biradha, K.; Kawano, M.; Sakamoto, S.;
Yamaguchi, K.; Fujita, M. Chem. - Asian J. 2006, 1, 82−90.
16) For cages based on bent (instead of linear) ligands see: Han, M.;
Engelhard, D. M.; Clever, G. H. Chem. Soc. Rev. 2014, 43, 1848−1860.
(
(
3
(
(
I
DOI: 10.1021/jacs.5b13190
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX