Communications
À
À
linear CF3SO3 in 3a and planar MeC6H4SO3 in 4a are
obviously mismatched with the cage cavity. Instead, MeOH
molecules are encapsulated as guests (Figure S1c). In addi-
tion, the strong tendency to form Cu+ cage is evident from
observations that 1a, 3a, and 4a can be obtained directly from
Cu2+ salts under ambient conditions, regardless of the anion.
Our preliminary investigations into these reactions in MeOH
by monitoring in situ through ESIMS spectra revealed that
the cage assembly proceeded rapidly (within 10 minutes),
which is indicative of fast Cu2+ to Cu+ reduction,[8] probably
along with oxidization of MeOH to formaldehyde. The
noticeable tendency towards cage formation may significantly
affect the redox potential of the Cu2+/Cu+ couple.
the rigid cage structure does not allow the benzyl ring to
À
rotate freely along the N Cmethylene bond, the two H6 atoms
thus become diastereotopic, split into two resolved peaks, and
move significantly upfield. These results suggest that the
[CuI4L4]4+ cage structure is retained in solution; the spectra
show only one set of well resolved signals in accordance with
formation of a high symmetrical cuboctahedral structure.
Further evidence for the solution structure came from the
ESIMS spectra (Figures S4 and 5). The peaks related to the
M4L4 cage were observed and confirmed by comparisons
between their measured and simulated isotopic distributions.
These Cu+ cages display distinct redox behavior in air
depending on the nature of the anion. When crystals of the
Cu+ complexes were kept in the mother liquors, slow
conversion into Cu2+ complexes occurred, which can be
easily judged from observation that the yellow crystals
disappeared gradually and green crystals grew. Conversion
of 3a and 4a into 3b and 4b, respectively, is complete within
one week, but complexes 1a and 2a are stable in solution for
several months. The structural analyses revealed a dinuclear
structure for 3b in which two Cu2+ ions take octahedral
geometry (Figure 1c). Complex 4b shows a tetranuclear
structure containing two octahedral and two square-pyrami-
dal Cu2+ ions (Figure 1d). In both structures, the ligand L was
hydroxylated to LOH and acts as a bridging ligand to chelate
two Cu2+ ions with the remaining coordination sites occupied
1H NMR and ESIMS measurements were carried out to
elucidate the solution structures of the cage compounds. As
shown in Figures 2, S2, and S3, the proton signals of L in 1a
were drastically shifted relative to those in the free ligand.
Generally, the peaks of Bim H atoms (other than H5) are
moved downfield, whereas the peaks of benzene, benzyl, and
methylene H atoms are moved upfield. The assignments of
these peaks have been verified carefully by 1H-COSY spectra
with clear proton correlation (Figure S3). Analyses of these
proton shifts show good consistency with the solid-state cage
structure. Coordination of Bim groups to the Cu+ ions is
expected to cause downfield shift of Bim protons owing to
metal-induced effects.[5b] However, an abnormal upfield shift
of the H5 peak is observed, as a result of specific disposition
of the Bim rings upon formation of the cage structure. As
discussed above, four Cu+ ions fix four benzene rings into an
aromatic core with six Bim pairs arraying in an offset parallel
fashion at cage windows. This configuration makes H5 atom
point to an adjacent Bim ring, thus subjecting it to ring current
shielding. Similarly, the H1 atom on the benzene ring is also
directed towards a neighboring Bim ring, and is consequently
upfield shifted. The H7–H9 atoms on the benzyl groups are all
located above the central aromatic core, thereby displaying an
upfield shift owing to arene ring shielding. The most
informative change is observed for the H6 atoms on
methylene group, which acts as a juncture to link Bim and
benzyl groups. The singlet peak in free L is divided into two
separate peaks with an upfield shift of more than 1.3 ppm. On
the basis of the solid-state cage structure, two H6 atoms of
each methylene are anchored beside a Bim ring and a central
benzene ring in every six offset parallel Bim pairs. Because
À
À
by O atoms from CF3SO3 or MeC6H4SO3 anions. Further
identification of L hydroxylation was accomplished by ESIMS
measurements. As seen in Figure S6, all salient peaks of 3b
can be assigned to dimeric or monomeric species containing
the LOÀ ion, confirming the formation of LOH from L.
Although a detailed mechanism of the hydroxylation of L
is still waiting for thorough investigation, an O2-activated
À
arene C H bond oxidation process, which has been widely
accepted in various synthetic copper model complexes,[7] may
be expected. A lot of predesigned multinuclear Cu+ precur-
sors have been proven to be able to capture O2 molecules to
mediate ligand hydroxylation, and the trigonal Cu+ ion in
multinuclear enzymes is believed to be purposeful for O2
reactivity.[7] To investigate the role of the Cu+ cages in
hydroxylation of L, we carried out a series of comparative
experiments by treating L with different Cu+ and Cu2+ salts in
MeOH at room temperature. In situ monitoring of the
reaction medium with ESIMS spectra revealed that, regard-
less of whether Cu+ and Cu2+ salts were used, the [CuI4L4]4+
cage structures were formed quickly (within 10 minutes), but
the hydroxylated LOH ligand could not be detected within
24 h. This result probably means that the Cu+ cage is the most
favorable thermodynamic product in the reaction of L with
Cu+/Cu2+ salts, and hydroxylation of L is initiated later by O2
attack of the [CuI4L4]4+ cage. Once Cu+ ions capture O2 with
conversion into Cu2+, the cage could undergo structural
rearrangement to facilitate the final hydroxylation of L. Such
a process may also account for the formation of the final
dinuclear and tetranuclear Cu2+ complexes 3b and 4b.
Further investigations on the mechanical details of the
hydroxylation are currently in progress.
Figure 2. 1H NMR spectra of ligand L (bottom) and complex 1a (top)
measured in [D6]DMSO. Shifts of the proton peaks are shown by the
arrows.
On the basis of above discussions, host–guest dependent
redox activity for coordination cages 1–4a may be speculated,
6158
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 6156 –6159