One-step versus multistep equilibrium of carbazole-bridged dinuclear zinc(II) complex formation: Metal-assisted π-association and-dissociation processes

Article abstract of DOI:10.1002/chem.201403036

This work demonstrates a selection criteria that determines whether molecular assembly occurs through a one-step or stepwise manner in ligand-bridged dinuclear zinc(II) (Zn²⁺) complex formation, which is associated with the π stacking of building blocks. The building blocks of carbazole ligands (L¹ and L⁴) that contain two imidazole moieties at the 3,6-positions form 4:2 complexes (i.e., [L]₄-(Zn²⁺)₂) at a molar ratio of 0.50 ([Zn²⁺]/[L]₀=0.50), thereby providing π stacking between the carbazole ligands. At the molar ratio of 0.67 ([Zn²⁺]/[L]₀=0.67), the 4:2 complexes change to 3:2 complexes (i.e., [L]₃-(Zn²⁺)₂) with no p-stacked carbazole unit. In contrast, when the imidazole groups in L¹ are replaced with benzoimidazole groups (L³), L³ also yields the 4:2 complex [(L³)₄-(Zn²⁺)₂] at a molar ratio of 0.50. However, there is no structural transition from (L³)₄-(Zn²⁺)₂ to other complex species above a molar ratio of 0.50. Similarly, when two imidazole groups are introduced into the carbazole ring at 2,7-positions (L⁵), L⁵ also gives the 4:2 complex [(L⁵)₄-(Zn²⁺)₂] that shows no structural transition to other complex species at a higher molar ratio.

Full text of DOI:10.1002/chem.201403036

DOI: 10.1002/chem.201403036
Full Paper
&
Supramolecular Systems
One-Step Versus Multistep Equilibrium of Carbazole-Bridged
Dinuclear Zinc(II) Complex Formation: Metal-Assisted
p-Association and -Dissociation Processes
Norie Inukai, Tsuyoshi Kawai,* and Junpei Yuasa*[a]
Abstract: This work demonstrates a selection criteria that
determines whether molecular assembly occurs through
a one-step or stepwise manner in ligand-bridged dinuclear
zinc(II) (Zn2+) complex formation, which is associated with
the p stacking of building blocks. The building blocks of car-
bazole ligands (L1 and L4) that contain two imidazole moiet-
ies at the 3,6-positions form 4:2 complexes (i.e., [L]4ꢀ(Zn2+)2)
at a molar ratio of 0.50 ([Zn2+]/[L]0 =0.50), thereby providing
p stacking between the carbazole ligands. At the molar ratio
of 0.67 ([Zn2+]/[L]0 =0.67), the 4:2 complexes change to 3:2
complexes (i.e., [L]3ꢀ(Zn2+)2) with no p-stacked carbazole
unit. In contrast, when the imidazole groups in L1 are re-
placed with benzoimidazole groups (L3), L3 also yields the
4:2 complex [(L3)4ꢀ(Zn2+)2] at a molar ratio of 0.50. However,
there is no structural transition from (L3)4ꢀ(Zn2+)2 to other
complex species above a molar ratio of 0.50. Similarly, when
two imidazole groups are introduced into the carbazole ring
at 2,7-positions (L5), L5 also gives the 4:2 complex [(L5)4ꢀ
(Zn2+)2] that shows no structural transition to other complex
species at a higher molar ratio.
Introduction
portant questions: What are the selection criteria that deter-
mine whether molecular assembly occurs in one step or a step-
wise manner? And how do building molecules lead to a single
thermodynamic product at each respective molar ratio in the
multistep mechanism under thermodynamic control? Once we
answer these critical questions, it will open up the opportunity
to rationally plan unique supramolecular systems for achieving
high diversity in structures and organization from a limited
number of building blocks.
Natural principles are based on the creation of vast libraries of
structures from a limited number of building blocks in a simul-
taneous multistep assembly process, thereby achieving a high
level of complexity.[1–3] The multistep assembly strategy is
a powerful method for achieving structural diversity from
simple covalent building blocks under thermodynamic control,
in which multiple species can interconvert between two or
more self-assembled structures through a relatively flat poten-
tial-energy surface (Scheme 1a).[4–11] The delicate thermody-
namic balance results in a multistep equilibrium system,[12–21] in
which the concentration of each of the molecular constituents
is determined by the concentrations and molar ratio of the pri-
mary building blocks. This leads to a single complex species at
each respective molar ratio, which can undergo structural tran-
sition to the other assembled structures at different molar
ratios (Scheme 1a).[12,13] However, most molecular assembly
strategies involve one-step equilibrium systems, which lead to
a single thermodynamic product as a result of its much higher
stability relative to the other members (Scheme 1b).[22,23] In
contrast, an almost flat energy landscape generates a complex
mixture of products (Scheme 1c). These classifications raise im-
In this work, these questions are addressed by investigating
metal-assisted p-association and -dissociation processes in one
step versus multistep formation of carbazole-bridged dinuclear
complexes.[24–30] The carbazole ligand (L) forms a 4:2 complex
[(L)4ꢀ(Zn2+)2] at a molar ratio of zinc ions (Zn2+) to the ligand
of 0.50,[31–33] which provides the opportunity for p stacking be-
tween the carbazole ligands. In the multistep mechanism, the
4:2 complex undergoes a structural transition to a 3:2 complex
[(L)3ꢀ(Zn2+)2] at the higher molar ratio, thereby resulting in dis-
sociation of p stacking between the carbazole ligands. Con-
versely, in the one-step mechanism, no clear structural transi-
tion into other complex species occurs, thus leading to [(L)4ꢀ
(Zn2+)2] as a single thermodynamic product. The p stacking be-
tween the carbazole ligands in the molecular assembly acts as
a suitable noncovalent interaction motif to achieve a delicate
thermodynamic balance between the multiple complex spe-
cies.[34] Systematic comparison between the two mechanisms
provides a valuable insight into how building molecules such
as carbazole ligands are capable of providing a single thermo-
dynamic product at each respective molar ratio in the stepwise
mechanism.
[a] Dr. N. Inukai, Prof. Dr. T. Kawai, Dr. J. Yuasa
Graduate School of Materials Science
Nara Institute of Science and Technology
8916-5 Takayama, Ikoma, Nara 630-0192 (Japan)
Fax: (+81)743-72-6179
E-mail: tkawai@ms.naist.jp
yuasaj@ms.naist.jp
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201403036.
Chem. Eur. J. 2014, 20, 15159 – 15168
15159
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
spectrum showed a subsequent change (Figure 1a, B
to C).[35] Such biphasic UV/Vis spectral changes are as-
cribed to multistep complex formation between L1
and Zn2+, whereby L1 forms two types of complex
species in which structures interconvert depending
on the Zn2+ concentration. The biphasic spectral
change is also found during titration of L4 with Zn2+
(Figure 1d), thus indicating a multistep equilibrium
between them.[37] However, in the titration of the
monosubstituted ligand (L2) that contained one bind-
ing site and was capable of forming only a mononu-
clear Zn2+ complex, a successive spectral change was
observed while maintaining an isosbestic point at
332 nm (Figure 1b). Similarly, L3 and L5 show one-
step UV/Vis spectral changes with strict isosbestic
points (Figure 1c and e, respectively). These succes-
sive spectral changes are a clear indication that these
ligands (L2, L3, and L5) form a single complex species
(thermodynamic product) with Zn2+
.
On the basis of the UV/Vis titration experiments
(Figure 1), the binding stoichiometry of the carbazole
ligands (L1–5) with Zn2+ was estimated. The results
are shown in Figure 1 f–j, in which each absorbance
is plotted against the ratio of the Zn2+ concentration
to the initial concentration of the carbazole ligands
([Zn2+]/[L]0). In the titration of L1 and L4 with Zn2+
(Figure 1 f and i, respectively), the titration curve re-
veals two breaks at around [Zn2+]/[L]0 =0.50 and
0.67, thus suggesting a binding stoichiometry of
[L]/[Zn2+]=2:1 and 3:2, respectively. The existence of
two breaks in the titration curves is also a clear indi-
cation of two types of complex species. Conversely,
in the titration curves of L3 and L5, only a single satu-
ration point exists at [Zn2+]/[L]0 =0.50 (Figure 1h and
j, respectively), thereby suggesting that these ligands
(L3 and L5) form a single complex species with Zn2+
in [L]/[Zn2+]=2:1 stoichiometry. As expected from
the chemical structure of L2 with one imidazole unit
that is capable of forming only a mononuclear spe-
cies, the titration plot (Figure 1g) of L2 with Zn2+ re-
veals a simple 4:1 stoichiometry ([Zn2+]/[L]0 =0.25),
Scheme 1. Complex formation of building blocks and its free-energy landscape.
Results and Discussion
One-step versus multistep complex formation between
carbazole ligands and Zn2+ ions
The chemical structures of all of the carbazole ligands (L1–5
)
used in this work are shown in Scheme 2. Initially, the complex
formation of carbazole ligands with Zn(OTf)2 (OTf=OSO2CF3)
was investigated by UV/Vis titration (see below). Upon addition
of 0–0.5 equivalents of Zn2+ to a solution of L1 ([L1]0 =1.5ꢁ
10ꢀ4 m) in acetonitrile (MeCN), L1 shows UV/Vis absorption
spectral changes with a clear isosbestic point observed at
375 nm (Figure 1a, A to B).[35,36] When more than 0.5 equiva-
lents of Zn2+ were introduced into the solution, the absorption
Scheme 2. Chemical structures of the carbazole ligands.
Chem. Eur. J. 2014, 20, 15159 – 15168
www.chemeurj.org
15160
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
reference compound L2 with one
imidazole unit, L2 gives rise to
a simple NMR spectroscopic pat-
tern in the presence of Zn2+ (see
Figure S2 in the Supporting In-
formation), which is in sharp
contrast to the bridging ligand
L1. Hence, further splitting of the
N-Me signal of L1 (Figure 2) indi-
cates that two sets of L1 ligands
are located in different chemical
environments and lost their C2
symmetry through formation of
a ligand-bridged complex. In ad-
dition to this, it should be noted
that the doublet peak due to
the aromatic proton of the car-
bazole ring (C1ꢀH) shows a con-
siderable upfield shift from
d=7.75
to
5.73 ppm
at
[Zn2+]/[L1]0 =0.50 (Figure 2). This
large upfield shift clearly shows
the shielding effects of the car-
bazole rings, a clear indication of
the p-stacked carbazole units.
This upfield signal disappears
with increasing concentration of
Zn2+ (Figure 2), whereby the
four splitting signals due to N-
Me protons of the imidazole
Figure 1. Observed UV/Vis absorption spectral changes upon addition of a) Zn2+ (A. 0m, B. ꢀ7.8ꢁ10ꢀ5 m,
C. ꢀ1.6ꢁ10ꢀ4 m) to a solution of L1 in MeCN (1.5ꢁ10ꢀ4 m); b) Zn2+ (A. 0m, B. ꢀ1.7ꢁ10ꢀ4 m) to a solution of L2 in
MeCN (3.1ꢁ10ꢀ4 m); c) Zn2+ (A. 0m, B. ꢀ9.8ꢁ10ꢀ5 m) to a solution of L3 in MeCN (8.5ꢁ10ꢀ5 m); d) Zn2+ (A. 0m,
B. ꢀ6.2ꢁ10ꢀ5 m, C. ꢀ5.4ꢁ10ꢀ4 m) to a solution of L4 in MeCN (1.2ꢁ10ꢀ4 m); and e) Zn2+ (A. 0m, B. ꢀ9.0ꢁ10ꢀ5 m)
to a solution of L5 in MeCN (9.0ꢁ10ꢀ5 m) at 298 K (1 mm path length). Plots of absorbance versus [Zn2+]/[L]0 for
the UV/Vis absorption titration of the carbazole ligands by Zn2+, in which [L]0 denotes the initial concentrations
of L1–L5. The absorbance is monitored at f) l=365 nm for L1, g) l=296 nm for L2, h) l=360 nm for L3,
i) l=360 nm for L4, and j) l=400 nm for L5.
thereby suggesting the formation of a simple mononuclear
groups become one peak at d=2.93 ppm (Figure 2 top, closed
square). These results indicate that above the molar ratio of
0.50 ([Zn2+]/[L1]0 >0.50), the complex with p-stacked carbazole
units undergoes a structural transition to another complex that
has no p-stacking carbazole ring, in which all L1 ligands bound
to Zn2+ have the same chemical environments with C2
symmetry.
complex (i.e., (L2)4–Zn2+).[38,39]
1H NMR spectroscopic characterization of complexes formed
between L1 and Zn2+
With these results in hand, the complex formation of the car-
bazole ligands (L1–5) with Zn2+ was carefully characterized by
systematic NMR spectroscopic titration analysis (see below). As
expected from the UV/Vis titration experiments, L1 also shows
two-stage NMR spectral responses during the NMR spec-
To address a full characterization of these complex species,
we have tried to prepare a single crystal of the complex, but
the multistep equilibrium between L1 and Zn2+ makes the for-
mation of the single crystal difficult. Finally, we obtained
a small, fiberlike crystal after careful screening of suitable crys-
tallization conditions. Single crystals were grown by the slow
evaporation method at a constant temperature of 258C from
a solution in MeCN. Unfortunately, the resolution of the X-ray
crystal structure is insufficient for unambiguously deducing the
1
troscopic titration (Figure 2).[35] The H NMR spectroscopic sig-
nals of L1 (2.2ꢁ10ꢀ3 m) show broadening owing to rapid ex-
change between free L1 and that bound to Zn2+ in a molar
ratio that ranged from [Zn2+]/[L1]0 =0.20 to 0.45 (Figure 2). The
1H NMR spectroscopic signal becomes sharp at molar ratio of
around [Zn2+]/[L1]0 =0.50, thus giving rise to a complex NMR
spectrum (Figure 2).[35] The 2:1 binding stoichiometry
([Zn2+]/[L1]0 =0.50) is consistent with that obtained from the
UV/Vis titration analysis (see above, Figure 1 f). The single peak
at d=3.82 ppm (Figure 2 bottom, closed square) owing to the
N-Me protons of the imidazole rings splits into four singlet
peaks at d=4.07, 3.85, 3.64, and 3.30 ppm at a molar ratio
([Zn2+]/[L1]0) of 0.50 (Figure 2). If L1 had formed a simple 2:1
complex with Zn2+ (i.e., (L1)2–Zn2+), only two distinct singlet
signals for the free and bound imidazole groups would have
ꢀ
position of the counteranion (OSO2CF3 ) and the solvent mole-
cule (MeCN); however, X-ray crystallography revealed p-
stacked carbazole units in the complex (not shown). In the
crystal structure, L1 forms a three-dimensional metal–organic
network architecture ([ꢀL1ꢀZn2+ꢀ]n) with Zn2+ [32]
,
in which
each Zn2+ binds to four imidazole rings. Coordination metal–
organic networks should be a result of the successive assembly
process of the complex species during the crystallization (con-
densation) process (Scheme 3). Hence, a structure of the carba-
zole-bridged dinuclear complex [(L1)4ꢀ(Zn2+)2] was extracted
from the crystal structure and optimized by molecular mechan-
1
appeared. In fact, in a H NMR spectroscopic titration of the
Chem. Eur. J. 2014, 20, 15159 – 15168
www.chemeurj.org
15161
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ments (Figure 3a, A–D), thereby providing a quantita-
tively consistent explanation of the 1H NMR spec-
troscopic data (Figure 2).
These systematic investigations have allowed us to
postulate a multistep equilibrium between L1 and
Zn2+ (Scheme 3). According to Scheme 3, L1 forms
the 4:2 complex [(L1)4ꢀ(Zn2+)2] with p-stacked carba-
zole units at the molar ratio of [Zn2+]/[L1]0 =0.50
(first step in Scheme 3), which changes to a 3:2 com-
plex [(L1)3ꢀ(Zn2+)2] above [Zn2+]/[L1]0 =0.50 (second
step in Scheme 3). The (L1)3ꢀ(Zn2+)2 complex has no
p-stacking carbazole ring, in which three L1 ligands
are bridged by two Zn2+ ions. Hence, reversible for-
mation and dissociation of p stacking between the
carbazole rings take place through a change in Zn2+
concentration. A model structure of (L1)3ꢀ(Zn2+)2 was
also created on the basis of the crystal structure.[41]
The model structure has no p-stacked L1 ligand, and
three equivalent L1 ligands bridge to two Zn2+ ions
with C2 symmetry (Figure 4), which is consistent with
the 1H NMR spectroscopic assignment (see above).
The (L1)3ꢀ(Zn2+)2 complex was also identified by ESI-
MS, whereby the positive-ion ESI mass spectrum of
a signal at m/z 818.1 corresponds to {Zn2[L1]3-
(OSO2CF3)2}2+ (see Figure S4 in the Supporting Infor-
mation). The characteristic distribution of the isoto-
pomer in its signals agrees closely with its calculated
isotopic distributions. However, the (L1)3ꢀ(Zn2+)2 com-
plex changes to (L1)4ꢀ(Zn2+)2 through dissociation
and reassociation at equilibrium. The exchange rate
is slow enough on the NMR spectroscopic timescale
to allow two distinct signals for (L1)4ꢀ(Zn2+)2 and
(L1)3ꢀ(Zn2+)2 (Figure 2). However, the chemical ex-
change is rapid on the nuclear Overhauser effect
(NOE) timescale (see below). When irradiated on the
doublet signals of the upfield-shifted carbazole
proton of (L1)4ꢀ(Zn2+)2 at d=5.73 ppm (Figure 3b),
unusually strong NOE signals were detected for the
protons of (L1)4ꢀ(Zn2+)2 at d=7.64 and 7.52 ppm
(4H) with the C1 proton of (L1)3ꢀ(Zn2+)2 at d=
7.15 ppm (Figure 3d). These intense NOE signals can
be assigned to rapid exchange between the carba-
zole ligands of the complex species on the NOE time-
scale through the ligand association and dissociation
processes.[42] These exchange NOE peaks enable us to
identify the doublet signals of (L1)4ꢀ(Zn2+)2 at d=
5.73, 7.14, and 7.52 ppm (4H) as the C1 proton in the
carbazole units.[43] The complex 1H NMR spectrum of (L1)4ꢀ
(Zn2+)2 was successfully identified by means of the above-men-
Figure 2. Stacked 1H NMR spectra of L1 (2.2ꢁ10ꢀ3 m) in the presence of Zn2+
(0–2.0ꢁ10ꢀ3 m) in CD3CN at 298 K. Numbers and symbols correspond to those in the
chemical structure of L1.
Scheme 3. Complex formation between L1 and Zn2+
.
ics (MM) (Figure 3a).[40] The model structure consists of two
types of L1, free and p-stacked L1 (Figure 3a; L1a and L1b, re-
spectively), in which the C1 proton of L1a are shielded by the
carbazole rings, which is consistent with the 1H NMR spec-
troscopic assignment (see above). The partial shielding be-
tween the carbazole rings causes the loss of C2 symmetry in
the bridged L1 ligands (L1a). In addition, the free and bound
imidazole groups in L1b result in the loss of C2 symmetry of L1b
(Figure 3a). As a consequence, the imidazole group is differen-
tiated into four parts, each with different chemical environ-
1
tioned procedures combined with H correlation spectroscopy
(COSY) and 2D rotating-frame nuclear Overhauser effect spec-
troscopy (ROESY) NMR spectroscopy (see details in the Experi-
mental Section and Figure S5 and S6 in the Supporting Infor-
mation), whereby each signal of L1 (C1ꢀH, C2ꢀH, C4ꢀH, and two
imidazole protons) is differentiated into four peaks (Fig-
ure 3b).[44]
Chem. Eur. J. 2014, 20, 15159 – 15168
www.chemeurj.org
15162
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 3. a) The structures of (L1)4ꢀ(Zn2+)2 modeled by molecular mechanics
(MM). A–D distinguishes imidazole moieties in different chemical environ-
1
ments. H NMR spectra of b) (L1)4ꢀ(Zn2+)2 and c) (L1)3ꢀ(Zn2+)2 in CD3CN at
298 K. d) NOE spectrum of a solution of L1 in CD3CN (4.9ꢁ10ꢀ2 m) in the
presence of Zn2+ (3.0ꢁ10ꢀ2 m) at irradiation of the carbazole C1 proton at
d=5.73 ppm. Numbers and symbols correspond to those in the chemical
structure of L1 and the structure of (L1)4ꢀ(Zn2+)2.
Figure 4. The structures of (L1)3ꢀ(Zn2+)2 modeled by molecular mechanics
(MM).
Thermodynamic parameters between (L1)4ꢀ(Zn2+
)
2
and (L1)3ꢀ(Zn2+
)
2
Evaluation of thermodynamic parameters between
(L1)4ꢀ(Zn2+)2 and (L1)3ꢀ(Zn2+)2 should yield important
insights into the multistep equilibrium between L1
and Zn2+ (Scheme 3). Hence, thermodynamic param-
eters were extracted from NMR spectra of a solution
in CD3CN that contained constant concentrations of
L1 (2.1ꢁ10ꢀ3 m) and Zn2+ (1.8ꢁ10ꢀ3 m) as a function
of temperature (see below). Under these conditions
with the molar ratio above 0.67 ([Zn2+]/[L1]0 =0.86),
almost all L1 ligands form (L1)3ꢀ(Zn2+)2 at 295 K, at
which there is no (L1)4ꢀ(Zn2+)2 peak in the NMR spec-
trum (Figure 5, bottom). However, with decreasing
temperature the (L1)4ꢀ(Zn2+)2 peak appears and its
intensity increases gradually (Figure 5 from bottom
to top), thereby suggesting that (L1)4ꢀ(Zn2+)2 is en-
thalpically favored over (L1)3ꢀ(Zn2+)2. The concentra-
tions of (L1)4ꢀ(Zn2+)2 and (L1)3ꢀ(Zn2+)2 at each tem-
perature were determined by the integration ratio of
Figure 5. 1H NMR spectra of L1 (2.1ꢁ10ꢀ3 m) in the presence of Zn2+ (1.8ꢁ10ꢀ3 m) in
CD3CN at 233–295 K.
Chem. Eur. J. 2014, 20, 15159 – 15168
www.chemeurj.org
15163
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
N-Me protons of (L1)4ꢀ(Zn2+)2 and (L1)3ꢀ(Zn2+)2 (Figure 5, closed
circles and squares, respectively), then the equilibrium constant
(K2) was calculated from the concentrations of (L1)4ꢀ(Zn2+)2,
(L1)3ꢀ(Zn2+)2 and free Zn2+ for each temperature (see Table S7
in the Supporting Information). Enthalpy (DH) and entropy
(DS) for the conversion of (L1)4ꢀ(Zn2+)2 into (L1)3ꢀ(Zn2+)2 were
obtained on the basis of temperature dependences of the K2
value as DH=(12.7ꢁ1.9) kcalmolꢀ1 and DS=(82.3ꢁ
7.5) calmolꢀ1 Kꢀ1 (see Figure S8 in the Supporting Information).
The large DH with positive DS should be a result of p stacking
between the carbazole rings, thereby inducing a higher order
of organization of (L1)4ꢀ(Zn2+)2. In addition to this, the affinity
of the free (unbound) ligand for the solvent also contributes to
the large enthalpy change (see below). Apparently the conver-
sion of (L1)4ꢀ(Zn2+)2 into (L1)3ꢀ(Zn2+)2 is an entropy-driven pro-
cess, in which the structural transition breaks the p stacking
between the carbazole rings. The free-energy change (DG2) for
the conversion (L1)4ꢀ(Zn2+)2 to (L1)3ꢀ(Zn2+)2 becomes negative
(a spontaneous process) at 298 K (DG2 =ꢀ11.6 kcalmolꢀ1) ac-
companied by a contribution of positive entropy (TDS=
(24.5ꢁ2.2) kcalmolꢀ1).
The formation constant (K1) of (L1)4ꢀ(Zn2+)2 is too large to be
determined accurately, whereby formation of (L1)4ꢀ(Zn2+)2
from free L1 and Zn2+ (first step in Figure 1 f) is almost stoi-
chiometric. Such an experimental result fits the calculated
curve when assuming a K1 value larger than 1.0ꢁ1026 mꢀ5 (see
Figure S10 in the Supporting Information), which corresponds
to a free-energy change DG1 <ꢀ35.5 kcalmolꢀ1. In contrast to
the large energy difference (3DG1 <ꢀ106.6 kcalmolꢀ1) be-
tween A and B states, B and D lie on a practically flat poten-
tial-energy surface (DG2 =ꢀ11.6 kcalmolꢀ1), which corresponds
to the structural transition of (L1)4ꢀ(Zn2+)2 into (L1)3ꢀ(Zn2+)2
with two Zn2+ ions. The flat potential-energy surface makes it
possible to show multistep equilibrium between L1 and Zn2+
.
However, conversion of state B to C (B!C) represents the
structural transition of (L1)4ꢀ(Zn2+)2 to (L1)3ꢀ(Zn2+)2 to release
free L1 ligands from (L1)4ꢀ(Zn2+)2. The free-energy change (DG3)
was estimated as DG3 >6.0 kcalmolꢀ1 from DG1 and DG2
values using Equation (5) (see Figure S11 in the Supporting
Information).
DG2 ¼ 4DG3 þ DG1
ð5Þ
In light of these results, we can elucidate a multistep equilib-
rium based on a process with four states as shown in
Scheme 4 (A–D).[12c,d] The equilibrium constants K1, K2, and K3
are defined in Equations (1), (2), and (3), respectively.
The free-energy change for conversion of state B to C, 3DG3 >
18.0 kcalmolꢀ1, is positive (exothermic process). Hence there is
no structural transition from (L1)4ꢀ(Zn2+)2 to (L1)3ꢀ(Zn2+)2 at the
molar ratio of [Zn2+]/[L1]0 =0.50. This point is important for the
carbazole ligand to be capable of forming the single thermo-
dynamic product at each respective molar ratio ([Zn2+]/[L1]0 =
0.50 and 0.67). If the DG3 value were negative (DG3 <0), L1
would directly form (L1)3ꢀ(Zn2+)2 in a one-step manner over
the whole range of molar ratios (Scheme 5a). Conversely, with
DG3 larger than DG1/4 (DG3 >DG1/4), the DG2 value would be
positive according to Equation (5), whereby no structural tran-
sition of (L1)4ꢀ(Zn2+)2 to (L1)3ꢀ(Zn2+)2 would occur (Scheme 5b).
Thus, the large formation energy (DG1) of (L)4ꢀ(Zn2+)2 with
small positive DG3 (0<DG3 <DG1/4) is an aspect of primary im-
portance in the multistep equilibrium system (Scheme 5c), in
which simple adjustments to the concentration of
DG1, K1
4 L þ 2 Zn2þ
ðLÞ ꢀðZn2þÞ2
ƒƒƒ!
ð1Þ
ð2Þ
ð3Þ
ƒƒƒ
4
DG2, K2
3ðLÞ4ꢀðZn2þÞ2 þ 2 Zn2þ
4ðLÞ ꢀðZn2þÞ2
ƒƒƒ!
ƒƒƒ
3
DG3, K3
ðLÞ4ꢀðZn2þÞ
ðLÞ ꢀðZn2þÞ2 þ L
ƒƒƒ!
ƒƒƒ
2
3
The ratio of complex formation (a) of (L1)4ꢀ(Zn2+)2 from free L1
and Zn2+ can be expressed by Equation (4) (see Figure S9 in
the Supporting Information).
ꢀ3
4 K1 ¼ a½Lꢂ0 ð1ꢀaÞꢀ4ð½Zn2þꢂꢀ0:5a½Lꢂ0Þꢀ2
ð4Þ
their constituent subunits provide the single thermo-
dynamic product at each respective molar ratio.
1H NMR spectroscopic observation for complex
formation of L3 and L4 with Zn2+
According to Scheme 5, the thermodynamic stability
of the free ligand in state C should be a crucial deter-
minant of multistep or one-step equilibrium systems.
Hence, the imidazole groups of L1 were replaced
with benzoimidazole groups (L3) with lower affinity
(solubility) for the polar solvent (MeCN). This modifi-
cation should result in a larger positive DG3, since
the benzoimidazole groups reduce the thermody-
namic stability of the unbound ligand in state C
(Scheme 5b). As expected from the UV/Vis titration
of L3 with Zn2+ (Figures 1c and 1h), L3 shows only
a one-step NMR spectroscopic spectral change as
Scheme 4. Energy diagram for complex formation between L1 and Zn2+ at 298 K.
Chem. Eur. J. 2014, 20, 15159 – 15168 www.chemeurj.org
15164
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1
Figure 6. a) Stacked H NMR spectra of L3 ([L3]0 =1.0ꢁ10ꢀ3 m) in the pres-
ence of Zn2+ (0–1.3ꢁ10ꢀ3 m) in CD3CN at 298 K. b) 1H NMR spectrum of L3
(1.0ꢁ10ꢀ3 m) in the presence of Zn2+ (7.2ꢁ10ꢀ4 m) in CD3CN at 298 K. Num-
bers, characters, and symbols correspond to those in the chemical structure
of L3.
Scheme 5. Energy diagram for complex formation between the carbazole
ligand (L) and Zn2+ for a) DG3 <0, b) DG3 >DG1/4, and c) 0<DG3 <DG1/4.
a function of Zn2+ concentration (Figure 6a), which is assign-
able to one-step complex formation between L3 and Zn2+
1
(Scheme 6a). In this case, too, the H NMR spectroscopic sig-
nals of L3 show exchange broadening that arises from rapid ex-
change between free L3 and the bound ligand below a molar
ratio of 0.50 [Zn2+]/[L3]0 <0.50 (Figure 6a). A series of sharp
NMR spectroscopic peaks are then observed at [Zn2+]/[L3]0 =
0.50, when almost all L3 ligands convert to (L3)4ꢀ(Zn2+)2 (Fig-
ure 6a). In this case, no further significant spectral change is
observed above the molar ratio of 0.50 [Zn2+]/[L3]0 >0.50,
thereby suggesting that no structural transition of (L3)4ꢀ(Zn2+)2
into other complex species occurs, even at the higher molar
ratio. These observations are consistent with the assumption
that complex formation between L3 and Zn2+ can be catego-
rized into Scheme 5b.
Scheme 6. Complex formation of Zn2+ with a) L3 and b) L4.
Supporting Information). The result is shown in Figure 6b, in
which each signal of L1 (C1ꢀH, C2ꢀH, C4ꢀH, and benzoimidazole
protons) is differentiated into four peaks. Additionally, aromatic
protons of the carbazole ring (C1 and C4) show considerable
upfield shifts to d=4.82, 5.21, and 5.23 ppm (Figure 6b), thus
suggesting p stacking between the carbazole rings. Such an
1
The H NMR spectrum of (L3)4ꢀ(Zn2+)2 was identified by prac-
tically the same procedures as employed in 1H NMR spec-
troscopic characterization of (L1)4ꢀ(Zn2+)2 (see Figure S12 in the
Chem. Eur. J. 2014, 20, 15159 – 15168
www.chemeurj.org
15165
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
NMR spectroscopic pattern is quite similar to that of (L1)4ꢀ comitant appearance of new signals that arise from the forma-
(Zn2+)2, which suggests that L1 forms the 4:2 complex [(L3)4ꢀ tion of (L4)3ꢀ(Zn2+)2, in which no clear upfield-shifted aromatic
(Zn2+)2] with Zn2+ as a single thermodynamic product.
proton is detected (Figure 7a). Thus, complex formation be-
tween L4 and Zn2+ can be categorized into Scheme 5c, in
which modification of the alkyl chain at the carbazole nitrogen
seems to have no significant influence on the DG3 value.
In the next step, the alkyl chain at the carbazole nitrogen of
L1 was modified to replace the ethyl group (L4). In this case, L4
exhibits the biphasic NMR spectral changes that accompany
complex formation with Zn2+ (Figure 7a), thus suggesting
a multistep equilibrium between L4 and Zn2+ (Scheme 6b).
The biphasic NMR spectral response is consistent with the
UV/Vis titration of L4 with Zn2+ (Figures 1d and 1i). Similarly to
L1 and L3 with Zn2+ (Figures 2 and 6a, respectively), L4 gives
a complex NMR spectrum at the molar ratio of [Zn2+]/[L4]0 =
0.50 (Figure 7a), thereby suggesting the formation of (L4)4ꢀ
1H NMR spectroscopic observation for complex formation of
L5 with Zn2+
Finally, we investigated complex formation between L5 (regio-
isomer of L1) and Zn2+ by using 1H NMR spectroscopy
(Figure 8). Since the L5 ligand has two imidazole moieties at
the 2,7-positions (in an almost linear arrangement), L5 should
1
(Zn2+)2. The H NMR spectrum of (L4)4ꢀ(Zn2+)2 was also identi-
1
fied by virtually the same procedures as employed in H NMR
spectroscopic identification of (L1)4ꢀ(Zn2+)2 and (L3)4ꢀ(Zn2+)2
(see Figures S13 and S14 in the Supporting Information). The
NMR spectroscopic data shows a common feature of the 4:2
complexes; each signal of L4 (C1ꢀH, C2ꢀH, C4ꢀH, and benzo-
imidazole protons) is differentiated into four peaks with signifi-
cantly upfield-shifted carbazole proton (C1ꢀH), as shown in Fig-
1
ure 7b. The H NMR spectroscopic signals of (L4)4ꢀ(Zn2+)2 de-
crease above the molar ratio of [Zn2+]/[L4]0 >0.50 with a con-
Figure 8. Stacked 1H NMR spectra of L5 (5.0ꢁ10ꢀ3 m) in the presence of Zn2+
(0–8.5ꢁ10ꢀ3 m) in CD3CN at 298 K. Numbers, characters, and symbols corre-
spond to those in the chemical structure of L5.
act as a less polar ligand and have lower thermodynamic sta-
1
bility in the polar solvent (MeCN). H NMR spectroscopic peaks
of L5 gradually shift with increasing concentration of Zn2+
,
then reaching saturation at the molar ratio of [Zn2+]/[L5]0 =
0.50 (Figure 8). The derived binding stoichiometry (two L5 li-
gands bound per Zn2+) agrees with that obtained with the
UV/Vis titration (see above, Figure 1i). In contrast to the
1H NMR spectroscopic structures of the other 4:2 complexes
(Figures 2b, 6b, and 7b), the complex between L5 and Zn2+
gives rise to a deceptively simple NMR spectroscopic pattern
(Figure 8). In such a case, there are two possible ways to ex-
plain the simple NMR spectroscopic pattern: 1) All L5 ligands
are located in the same chemical environments and maintain
C2 symmetry in the complex, or 2) the complex undergoes
rapid dissociation and regeneration on the NMR spectroscopic
timescale. For verification of this point, the complex stoichiom-
etry of L5 with Zn2+ was also analyzed by ESI-MS, a convenient
method of probing the stoichiometry and distribution of metal
complexes in solutions. Intense mass signals were detected at
1
Figure 7. a) Stacked H NMR spectra of L4 ([L4]0 =6.0ꢁ10ꢀ3 m) in the pres-
ence of Zn2+ (0–9.6ꢁ10ꢀ3 m) in CD3CN at 298 K. b) 1H NMR spectrum of L4
(8.0ꢁ10ꢀ3 m) in the presence of Zn2+ (4.3ꢁ10ꢀ3 m) in CD3CN at 298 K. Num-
bers, characters, and symbols correspond to those in the chemical structure
of L4.
m/z 817.7 {Zn2[L5]3(OSO2CF3)2}2+, 1019.3 {Zn2[L5]4(OSO2CF3)2}2+
,
Chem. Eur. J. 2014, 20, 15159 – 15168
www.chemeurj.org
15166 ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1200.2 {Zn4[L5]3(OSO2CF3)4}2+, 1422.4 {Zn[L5]3(OSO2CF3)}+, and
1784.3 for {Zn2[L5]3(OSO2CF3)3}+ (Figure 9). The mass signal at
m/z 1019.3 corresponds to the 4:2 complex [(L5)4ꢀ(Zn2+)2], and
the characteristic distribution of isotopomers in the signal
assembling processes are considered a primitive case of metal-
assisted p-association and -dissociation processes by means of
simple adjustments in the concentration of their building
units. Conversely, when the imidazole groups of L1 are re-
placed with benzoimidazole groups (L3) with lower affinity for
the polar solvent, L3 forms the 4:2 complex [(L3)4ꢀ(Zn2+)2] in
a one-step manner, in which there is no structural transition to
other complex species at the higher molar ratio. In addition,
when two imidazole groups are introduced into the carbazole
ring at the 2,7-positions (L5) to reduce the thermodynamic sta-
bility in the polar solvent, L5 also gives the 4:2 complex [(L5)4ꢀ
(Zn2+)2] and shows no structural transition to other complex
species. Thus, the simple modification of L1 changes the multi-
step to a one-step mechanism. This fact indicates that the ther-
modynamic stability of free building blocks is a key determi-
nant in whether molecular assembly occurs in a one-step or
multistep mechanism. Additionally, moderate thermodynamic
stability of building blocks should be important for achieving
a single thermodynamic product at each respective ratio of
their constituent units in the multistep mechanism. This insight
will open up new opportunities for creating novel supramolec-
ular systems and their applied materials with high complexity
under thermodynamic control.
Figure 9. Positive-ion ESI-MS of a solution of L5 (2.6ꢁ10ꢀ4 m) in MeCN in the
presence of Zn2+ (1.3ꢁ10ꢀ4 m). Inset: Isotopically resolved signals at m/z
1019.3 and the calculated isotopic distributions for {[Zn2(L5)4](OSO2CF3)2}2+
Assignment: m/z 817.7 for {Zn2[L5]3(OSO2CF3)2}2+, 1019.3 for {Zn2[L5]4-
(OSO2CF3)2}2+, 1200.2 for {Zn4[L5]3(OSO2CF3)4}2+, 1422.4 for {Zn[L5]3-
(OSO2CF3)}+, and 1784.3 for {Zn2[L5]3(OSO2CF3)3}+.
.
Experimental Section
agrees closely with their calculated isotopic distribution
(Figure 9, inset). The other intense mass signals correspond to
3:2, 4:3, and 3:1 complexes ((L5)3ꢀ(Zn2+)2, (L5)4ꢀ(Zn2+)3, and
(L5)3ꢀZn2+, respectively) (see Figure S15 in the Supporting In-
formation). Judging from the titration analysis (Figures 1i and
Figure 8) in comparison with the ESI-MS spectroscopy
(Figure 9), L5 seems to form (L5)4ꢀ(Zn2+)2 as a major compo-
nent with other minor complex species ((L5)3ꢀ(Zn2+)2, (L5)4ꢀ
(Zn2+)3, and (L5)3ꢀZn2+), in which ligand dissociation and reas-
sociation are rapid on the NMR spectroscopic timescale (pat-
tern 2).
General
Chemicals were purchased from Wako Pure Chemical Industries
Ltd. and used as received without further purification. Acetonitrile
(MeCN) used as a solvent was obtained from Nacalai Tesque. De-
tailed synthetic procedures and characterization of L1–5 are given in
the Supporting Information.
Spectral measurements
Complex formation between L1–5 and Zn2+ was examined from the
UV/Vis spectral change of L1–5 in the presence of various concen-
trations of Zn2+ by using a Jasco V-660 spectrophotometer. The
complex formation was also examined from the 1H NMR spectral
change of L1–5 in the presence of various concentrations of Zn2+
using a JEOL AL-300N FT NMR system (300 MHz). 1H NMR spec-
troscopic assignment of the 4:2 complexes is given in the Support-
ing Information. The complexes formed between L5 and Zn2+ were
detected by ESI-MS. Mass spectra (ESI-MS and FAB-MS) were mea-
sured using mass spectrometers (JEOL JMS-700 MStation).
Thus, complex formation between L5 and Zn2+ can be cate-
gorized under Scheme 5b, in which the substitution position
of metal binding sites (2,7- or 3,6-position) is sensitive to the
DG3 value.
Conclusion
In this work, we have successfully demonstrated the selection
criteria that determine whether molecular assembly occurs Acknowledgements
through a one-step or stepwise manner in carbazole-bridged
dinuclear Zn2+ complex formation. The building blocks of the
This work was partly supported by a Grant-in-Aid for Young
Scientists (A), Scientific Research (C) (no. 21750147), Scientific
carbazole ligand (L1) with two imidazole groups at the 3,6-po-
sitions forms two types of ligand-bridged complexes, (L1)4ꢀ Research on Innovative Areas: “Science of Super-molecular
(Zn2+)2 and (L1)3ꢀ(Zn2+)2, at molar ratios ([Zn2+]/[L1]0) of 0.50
and 0.67, respectively. On the one hand, formation of (L1)4ꢀ
(Zn2+)2 provides the opportunity for p stacking between the
carbazole ligands; but on the other hand, structural transition
into (L1)3ꢀ(Zn2+)2 results in dissociation of the p stacking. Multi-
step equilibrium is also observed for L4, which contains a 2-
methylbutyl group at the carbazole nitrogen. Such multistep
Structures and Creation of Chemical Elements”, and Scientific
Research (A) (no. 21107520) from the Ministry of Education,
Culture, Sports, Science and Technology, Japan.
Keywords: carbazoles
·
ligand effects
· supramolecular
chemistry · thermodynamics · zinc
Chem. Eur. J. 2014, 20, 15159 – 15168
www.chemeurj.org
15167
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[1] P. Ball, The Self-Made Tapestry: Pattern Formation in Nature, Oxford Uni-
versity Press, Oxford, 1999.
[2] D. Philip, J. F. Stoddart, Angew. Chem. Int. Ed. Engl. 1996, 35, 1155.
[3] C. R. South, C. Burd, M. Weck, Acc. Chem. Res. 2007, 40, 63.
[4] J.-M. Lehn, A. V. Eliseev, Science 2001, 291, 2331.
[34] The p-stacking interaction is one of the most common noncovalent in-
teraction motifs. It determines the function and structure of the molec-
ular assemblies. See: a) W. Saenger, Principles of Nucleic Acid Structure,
Springer, New York, 1984; b) D. B. Amabilino, J. F. Stoddart, Chem. Rev.
1995, 95, 2725; c) H. W. Roesky, M. Andruh, Coord. Chem. Rev. 2003, 236,
91.
[5] a) S. J. Rowan, S. J. Cantrill, G. R. L. Cousins, J. K. M. Sanders, J. F. Stod-
dart, Angew. Chem. 2002, 114, 938; Angew. Chem. Int. Ed. 2002, 41, 898;
b) F. Cougnon, J. Sanders, Acc. Chem. Res. 2012, 45, 2211.
[6] a) S. Otto, R. L. E. Furlan, J. K. M. Sanders, Science 2002, 297, 590; b) P. T.
Corbett, J. Leclaire, L. Vial, K. R. West, J.-L. Wietor, J. K. M. Sanders, S.
Otto, Chem. Rev. 2006, 106, 3652; c) R. A. R. Hunta, S. Otto, Chem.
Commun. 2011, 47, 847; d) S. Otto, Acc. Chem. Res. 2012, 45, 2200.
[7] A. V. Davis, R. M. Yeh, K. N. Raymond, Proc. Natl. Acad. Sci. USA 2002, 99,
4793.
[8] a) R. S. K. Kishore, V. Kalsani, M. Schmittel, Chem. Commun. 2006, 3690;
b) K. Mahata, M. Schmittel, J. Am. Chem. Soc. 2009, 131, 16544.
[9] a) S. Hiraoka, T. Yi, M. Shiro, M. Shionoya, J. Am. Chem. Soc. 2002, 124,
14510; b) K. Harano, S. Hiraoka, M. Shionoya, J. Am. Chem. Soc. 2007,
129, 5300.
[10] T. Haino, T. Fujii, A. Watanabe, U. Takayanagi, Proc. Natl. Acad. Sci. USA
2009, 106, 10477.
[11] Structural diversity can also be achieved by kinetic trapping of different
molecular assemblies based on a single building block. See: Y. Tidhar, H.
Weissman, S. G. Wolf, A. Gulino, B. Rybtchinski, Chem. Eur. J. 2011, 17,
6068.
[35] Some preliminary results on stepwise complex formation between L1
and Zn2+ have appeared. See: N. Inukai, J. Yuasa, T. Kawai, Chem.
Commun. 2010, 46, 3929.
[36] The clear isosbestic point indicates that all L1 ligands convert the single
complex, that is, (L1)4ꢀ(Zn2+)2 at a molar ratio of [Zn2+]/[L1]0 =0.50.
[37] Because L4 has a chiral alkyl group at the carbazole nitrogen, we also
examined the circular dichroism (CD) spectral analysis of L4 with Zn2+
.
Although L4 shows no detectable CD signal in the absence of Zn2+ (see
Figure S1 in the Supporting Information), successive addition induces
intense CD bands in the spectral range of the p–p* transition of the
carbazole ligand. The intense CD signal should arise from the exciton
coupling of the bridging carbazole units in the dinuclear complex.
[38] The 4:1 complex formation between L2 and Zn2+ was also confirmed
by 1H NMR spectroscopic titration (see Figure S2 in the Supporting In-
formation).
[39] The binding stoichiometries were also determined by Job plots for L1–
L5 with Zn2+ (see Figure S3 in the Supporting Information). The deter-
mined stoichiometries are same as those determined by UV/Vis titration
1
(Figure 1) and H NMR spectroscopic titration (Figures 2, 6–8).
[40] In the preliminary study, we tentatively concluded that L1 forms
a dimer complex with Zn2+. As part of a continuing investigation, we
obtained the crystal structure of the Zn2+ complex with L1 (not shown)
and reevaluated the binding stoichiometry between L1 and Zn2+. This
systematic investigation has allowed us to conclude that L1 forms
(L1)4ꢀ(Zn2+)2. It should be noted that we have provided the additional
supplementary materials in the preliminary study; see ref. [35].
[41] Although the species at high Zn2+ concentration is difficult to assign
with certainty from the simple NMR spectroscopic pattern (Figure 2)
and there might be other possible assignments, we tentatively assigned
the complex at high Zn2+ concentration as (L)3ꢀM2-type bridging com-
plexes from the titration analysis and ESI-MS detection. For recent ex-
amples of (L)3ꢀM2-type bridging complexes, see: a) N. Kundu, M. Maity,
P. B. Chatterjee, S. J. Teat, A. Endo, M. Chaudhury, J. Am. Chem. Soc.
2011, 133, 20104; b) T. Haino, H. Shio, R. Takano, Y. Fukazawa, Chem.
Commun. 2009, 2481; c) B. Birkmann, A. W. Ehlers, R. Frçhlich, K. Lam-
mertsma, F. E. Hahn, Chem. Eur. J. 2009, 15, 4301; d) E. Terazzi, L.
Guꢃnꢃe, B. Bocquet, J.-F. Lemonnier, N. D. Favera, D. Piguet, Chem. Eur.
J. 2009, 15, 12719.
[12] a) T. Ogawa, J. Yuasa, T. Kawai, Angew. Chem. 2010, 122, 5236; Angew.
Chem. Int. Ed. 2010, 49, 5110; b) J. Yuasa, A. Mitsui, T. Kawai, Chem.
Commun. 2011, 47, 5807; c) N. Inukai, T. Kawai, J. Yuasa, Chem.
Commun. 2011, 47, 9128; d) N. Inukai, T. Kawai, J. Yuasa, Chem. Eur. J.
2013, 19, 5938.
[13] a) J. Yuasa, S. Fukuzumi, J. Am. Chem. Soc. 2006, 128, 15976; b) J. Yuasa,
S. Fukuzumi, J. Am. Chem. Soc. 2008, 130, 566.
[14] a) P. N. W. Baxter, Chem. Eur. J. 2002, 8, 5250; b) P. N. W. Baxter, Chem.
Eur. J. 2003, 9, 2531.
[15] S. Goswami, D. Sen, N. K. Das, Org. Lett. 2010, 12, 856.
[16] a) S. Akine, S. Hotate, T. Nabeshima, J. Am. Chem. Soc. 2011, 133, 13868;
b) S. Akine, S. Sairenji, T. Taniguchi, T. Nabeshima, J. Am. Chem. Soc.
2013, 135, 12948.
[17] a) J. Yuasa, T. Suenobu, S. Fukuzumi, J. Am. Chem. Soc. 2003, 125,
12090; b) J. Yuasa, S. Fukuzumi, Org. Biomol. Chem. 2004, 2, 642; c) J.
Yuasa, T. Suenobu, S. Fukuzumi, J. Phys. Chem. A 2005, 109, 9356.
[18] I. C. Pintre, S. Pierrefixe, A. Hamilton, V. Valderrey, C. Bo, P. Ballester,
Inorg. Chem. 2012, 51, 4620.
[19] J. Etxebarria, A. Vidal-Ferran, P. Ballester, Chem. Commun. 2008, 5939.
[20] V. V. Borovkov, J. M. Lintuluoto, Y. Inoue, Org. Lett. 2002, 4, 169.
[21] S. E. Boiadjiev, D. A. Lightner, J. Am. Chem. Soc. 2000, 122, 378.
[22] J.-M. Lehn, Supramolecular Chemistry, VCH, Weinheim, 1995.
[23] G. M. Whitesides, B. Grzybowski, Science 2002, 295, 2418.
[24] C. Piguet, G. Bernardinelli, G. Hopfgartner, Chem. Rev. 1997, 97, 2005.
[25] M. Albrecht, Chem. Rev. 2001, 101, 3457.
[26] M. J. Hannon, L. J. Childs, Supramol. Chem. 2004, 16, 7.
[27] M. Albrecht, R. Frçhlich, Bull. Chem. Soc. Jpn. 2007, 80, 797.
[28] B. H. Northrop, Y. Zheng, K.-W. Chi, P. J. Stang, Acc. Chem. Res. 2009, 42,
1554.
[42] E. Alessio, L. Hansen, M. Iwamoto, L. Marzilli, J. Am. Chem. Soc. 1996,
118, 7593.
[43] These exchange NOE signals were further confirmed by 2D ROESY NMR
spectrum (red dashed lines in Figure S5 in the Supporting Information),
in which red spots in the 2D ROESY show negative chemical exchange
peaks. No ROE peak was observed under these conditions (L1 (2.0ꢁ
10ꢀ3 m) and Zn2+ (1.0ꢁ10ꢀ3 m)). 2D ROESY of free L1 (2.0ꢁ10ꢀ3 m) is also
given Figure S5 in the Supporting Information.
[44] 2D DOSY spectra of (L1)4ꢀ(Zn2+
)
and (L1)3ꢀ(Zn2+
) are given in Fig-
2
2
ures S16 and S17 of the Supporting Information. All 1H signals of (L1)4ꢀ
(Zn2+
) (green squares in Figure S16 in the Supporting Information)
2
[29] R. Chakrabarty, P. S. Mukherjee, P. J. Stang, Chem. Rev. 2011, 111, 6810.
[30] J. Vicente, J. Gil-Rubio, N. Barquero, V. Cꢂmara, N. Masciocchi, Chem.
Commun. 2010, 46, 1053.
have the same diffusion coefficient (D=1.5ꢁ10ꢀ9 m2 sꢀ1, s=0.01750).
This result supports that the complicated 1H signals do not originate
from the mixture of other complexes in the ratio of 2:1 (i.e., (L)2ꢀZn2+
,
[31] Carbazole derivatives have been extensively utilized as building blocks
for metal–organic network architectures. See: a) J.-R. Li, H.-C. Zhou,
Angew. Chem. 2009, 121, 8617; Angew. Chem. Int. Ed. 2009, 48, 8465;
b) J.-R. Li, H.-C. Zhou, Nat. Chem. 2010, 2, 893; c) J.-R. Li, A. Yakovenko,
W. Lu, D. Timmons, W. Zhuang, D. Yuan, H.-C. Zhou, J. Am. Chem. Soc.
2010, 132, 17599; d) W. Lu, D. Yuan, T. A. Makal, J.-R. Li, H.-C. Zhou,
Angew. Chem. 2012, 124, 1612; Angew. Chem. Int. Ed. 2012, 51, 1580.
[32] H.-J. Liu, X.-T. Tao, J.-X. Yang, Y.-X. Yan, Y. Ren, H.-P. Zhao, Q. Xin, W.-T.
Yu, M.-H. Jiang, Cryst. Growth Des. 2008, 8, 259.
(L)6ꢀ(Zn2+)3, (L)8ꢀ(Zn2+)4, and so on). As expected, the diffusion coeffi-
cient (D=1.3ꢁ10ꢀ9 m2 sꢀ1
,
s=0.01756) of (L1)3ꢀ(Zn2+
)
determined
2
from the 2D DOSY spectra is close to that of (L1)4ꢀ(Zn2+
10ꢀ9 m2 sꢀ1), thus indicating that the size of the (L1)4ꢀ(Zn2+
different from that of (L1)3ꢀ(Zn2+)2.
)
(D=1.5ꢁ
2
)
is not so
2
Received: April 11, 2014
Revised: July 4, 2014
[33] E. T. Spielberg, W. Plass, Eur. J. Inorg. Chem. 2011, 826.
Published online on September 18, 2014
Chem. Eur. J. 2014, 20, 15159 – 15168
www.chemeurj.org
15168
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim