Angewandte
Communications
Chemie
ing sufficient solubility. Dimerization of cavitands 1 and 2 was
investigated by a 1H NMR dilution study at 283 K (see
Section S3.2). A highly concentrated solution of 2 at 16.2 mm
was diluted stepwise to 0.1 mm. At high concentration,
dimerization forces both hemispheres into a more rigid
structure. Upon dilution, the capsular assembly dissociates
resulting in a more flexible vase conformation, indicated by
a clear upfield shift of the methine resonance frequency and
all aromatic signals. The association constant could be
determined by non-linear curve fitting of the dimerization-
induced chemical shift as a function of concentration. The
multitude of weak S···N interactions forming a dimeric
capsule 2···2 resulted in an association constant of Ka =
786mꢁ1 [21,22]
Dimerization of 2 was further confirmed by
.
1
determination of the diffusion coefficient through H diffu-
sion ordered spectroscopy (DOSY).[23] With increasing con-
centration the diffusion coefficient decreased, indicating
a doubling of volume and thus, corroborating dimerization
of S cavitand 2 to a chalcogen-bonded capsule 2···2 (see
Section S3.4.1).
Figure 4. Variable temperature (VT) study of 1C6 in [D8]toluene
(10 mm, 1H NMR, 500 MHz). The region of the methine and aromatic
signals is displayed. Signals corresponding to species (1C6···1C6)a are
indicated in red, to species (1C6···1C6)b in blue. The equilibrium is
shifted upon heating from 20 to 858C in less than one hour.
1
As expected, the H NMR spectra of Te cavitand 1 in
[D8]THF did not display a change in chemical shift upon
dilution (Figure S12). We concluded that strong chalcogen-
bonding interactions favor formation of a supramolecular
structure at all concentrations measured. Hence, a dimeric
assembly 1···1 could be detected exclusively within the NMR
detection limit, never the monomeric cavitand 1.
both species (Dð1ꢂꢂꢂ1Þ = 2.99 ꢁ 10ꢁ10 m2 sꢁ1, Dð1ꢂꢂꢂ1Þ = 2.83 ꢁ
a
b
10ꢁ10 m2 sꢁ1, see Section S3.4.2), suggesting a comparable
volume of (1···1)a and (1···1)b. Furthermore, mass spectrom-
etry showed signals of the singly charged homodimer and
monomer but not of oligomeric assemblies (Section S5).
Therefore, we propose for initial species (1···1)a a dimeric
assembly, similar to the crystal structure obtained for 2C6···2C6
(Figure 2a, right) displaying a shifted supramolecular capsule
based on chalcogen bonding.[25] In a shifted capsule, fewer
Te···N interactions than the maximum of 16 in a linear capsule
with 8 2Te–2N squares would make this structure more easily
accessible but less stable.
The Te capsule 1···1 (and analogously 1C6···1C6) displayed
a time-dependent transformation in solution from an initial
species (1···1)a to a second species (1···1)b, noticeable by
a doubling and shifting of NMR signals (see Section S3.3).
The pace of this transformation was dependent on solvent and
concentration ranging from full conversion within one day
(e.g. in CDCl3, 1 mm) to more than one week (e.g. in
[D8]toluene, 10 mm). The equilibrium could be shifted in
favor of the second species by heating to 858C in [D8]toluene,
as shown for 1C6···1C6 in Figure 4. The transformation proves
irreversible upon cooling as the second species (1C6···1C6)b
continues to prevail. Considering that crystals of the Te
capsule 1C6···1C6 were grown over several days, the second
species formed is more likely to crystallize. Indeed,
Finally, a native electrospray ionization mass spectrome-
try (ESI-MS)[26] titration was performed to determine the
association constant of the dimerization of Te cavitand 1 to
a supramolecular capsule (Section S5). To ensure the pres-
ence of the vase conformation of 1 at low concentration while
maintaining an environment suitable for ESI experiments,
a solvent mixture of DMSO/toluene/MeOH 10:6:1 (v/v/v)
was used. Comprising DMSO and protic MeOH, this solvent
system allows competitive interactions to chalcogen bonding,
possibly weakening the association. Furthermore, the opti-
mized conditions resulted in precipitation after one day,
which is why presumably only the association of species
(1···1)a could be determined. Figure 5 shows an increasing
intensity for the singly charged dimer at concentrations going
from 9.5 mm to 380 mm while the singly charged monomer
signal substantially decreases. This concentration-dependent
change in intensity allowed determination of a high associ-
1
a H NMR investigation of the redissolved crystals verified
the sole presence of the second species (1C6···1C6)b. The
tremendous strength of 16 Te···N interactions in 8 2Te–2N
squares leads to the kinetically inert, thermodynamically most
stable capsule (1C6···1C6)b. This perfect capsule with 16 Te···N
interactions forms very slowly from an initial less symmet-
rical, less stable capsular assembly with a lower number of
Te···N interactions. Fujita and co-workers reported a similar
observation for the self-assembly of PdII- or PtII-linked
macrocycles with 4,4’-bipyridine (bpy).[24] While the thermo-
dynamically most stable cyclic tetramer was immediately
formed with weak Pd···N interactions, strong Pt···N interac-
tions led first to formation of a mixture of oligomers which
eventually, after heating to 1008C for several weeks, resulted
in the thermodynamically stable cyclic tetramer after all.
However, in the case of our Te cavitand 1, oligomerization
could be excluded on taking into account that 1H DOSY
experiments in C6D6 gave an equal diffusion coefficient for
ation constant of Ka,ð1ꢂꢂꢂ1Þ = 2.9 ꢃ 0.4 ꢁ 107 mꢁ1, caused by the
a
cooperativity of several strong Te···N interactions. For the
thermodynamic species (1···1)b under optimal conditions,
without competitive solvent effects and a linear arrangement
with 16 Te···N interactions in 8 2Te–2N squares, an even much
larger association constant can be expected.
4
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2018, 57, 1 – 7
These are not the final page numbers!