From 2-fold completive to integrative self-sorting: A five-component supramolecular trapezoid

Article abstract of DOI:10.1021/ja907185k

The amalgamation of two incomplete self-sorting processes into a process that makes quantitative use of all members of the library is described by 2-fold completive self-sorting. Toward this goal, individual metal-ligand binding scenarios were optimized f

Full text of DOI:10.1021/ja907185k

Published on Web 10/28/2009
From 2-Fold Completive to Integrative Self-Sorting: A
Five-Component Supramolecular Trapezoid
Kingsuk Mahata and Michael Schmittel*
Center of Micro- and Nanochemistry and Engineering, Organische Chemie I, UniVersita¨t Siegen,
Adolf-Reichwein-Strasse 2, D-57068 Siegen, Germany
Received August 25, 2009; E-mail: schmittel@chemie.uni-siegen.de
Abstract: The amalgamation of two incomplete self-sorting processes into a process that makes quantitative
use of all members of the library is described by 2-fold completive self-sorting. Toward this goal, individual
metal-ligand binding scenarios were optimized for high thermodynamic stability and best selectivity, by
screening a variety of factors, such as steric and electronic effects, π-π interactions, and metal-ion specifics.
Using optimized, heteroleptic metal-ligand binding motifs, a library of four different ligands (1, 2, 3, 4) and
two different metal ions (Zn2+, Cu+) was set up to assess 2-fold completive self-sorting. Out of 20 different
combinations, the self-sorting library ended up with only two metal-ligand complexes in basically quantitative
yield. To demonstrate the value of 2-fold completive self-sorting for the formation of nanostructures, the
optimized, highly selective binding motifs were implemented into three polyfunctional ligands. Their integrative
self-sorting in the presence of Zn2+ and Cu+ led to the clean formation of the supramolecular trapezoid T,
a simple but still unknown supramolecular architecture. The dynamic trapezoid T consists of three different
ligands with four different donor-acceptor interactions. Its structure was established by 1H NMR
spectroscopy, electrospray ionization mass spectroscopy, and differential pulse voltammetry (DPV) and
by exclusion of alternative structures.
1. Introduction
energy change. The thermodynamically controlled outcome,
however, will depend upon the differences in Gibbs free energy
(∆∆Grxn) between all possible pathways. When ∆∆Grxn is
sufficiently large, a tangible difference in the products’ popula-
tion may be observed, with the process now being described as
self-sorting. Seminal self-sorting systems based upon metal-ligand
coordination,4-9 hydrogen bonding,10 solvophobic effects,11
electrostatic interactions,12 and dynamic covalent chemistry13
have been developed during the past decade. However, the
synthetic use of self-sorting is still in its infancy, and the
enlargement of ∆∆Grxn remains a challenge for chemists.
The generation of complex, functional architectures from a
diversity of different building blocks using noncovalent self-
assembly along a self-sorting algorithm is quite common in
nature. According to Darwin’s “Survival of the Fittest” prin-
ciple,1 biological systems built on diversity and complexity can
profit most from evolution and “Natural Selection” by the
instatement of new emergent properties. Thus, the enhancement
of complexity and multifariousness should be important for
artificial supramolecular assemblies as well. Despite a notable
wealth of self-assembled architectures published to date, the
majority is comprised of a single unit (ligand) by using only
one type of noncovalent interaction. As such, most of them lack
diversity and will not be able to “excel” in our highly functional
and sophisticated world.2 Unsurprisingly, most of the known
supramolecular architectures have been conceived to represent
regular shapes, such as squares, rectangles, or other highly
symmetric aggregates. To generate novel emergent properties,
future artificial supramolecular species need to be composed
of multiple components as well as multiple interactions.2 Proper
use of self-sorting would contribute to solve the problem.
A superb mastery of self-sorting is required3 to concatenate
different subunits with precise constitutional and/or positional
control. Diverse self-sorting algorithms have to compete in
multicomponent assemblies with their multitude of noncovalent
interactions, with each product giving rise to a different free
(4) Kra¨mer, R.; Lehn, J.-M.; Marquis-Rigault, A. Proc. Natl. Acad. Sci.
U.S.A. 1993, 90, 5394.
(5) Caulder, D. L.; Raymond, K. N. Angew. Chem., Int. Ed. 1997, 36,
1440.
(6) (a) Enemark, E. J.; Stack, T. D. P. Angew. Chem., Int. Ed. 1998, 37,
932. (b) Albrecht, M.; Schneider, M.; Ro¨ttele, H. Angew. Chem., Int.
Ed. 1999, 38, 557. (c) Hwang, I.-W.; Kamada, T.; Ahn, T. K.; Ko,
D. M.; Nakamura, T.; Tsuda, A.; Osuka, A.; Kim, D. J. Am. Chem.
Soc. 2004, 126, 16187. (d) Schultz, D.; Nitschke, J. R. Angew. Chem.,
Int. Ed. 2006, 45, 2453. (e) Kamada, T.; Aratani, N.; Ikeda, T.; Shibata,
N.; Higuchi, Y.; Wakamiya, A.; Yamaguchi, S.; Kim, K. S.; Yoon,
Z. S.; Kim, D.; Osuka, A. J. Am. Chem. Soc. 2006, 128, 7670. (f)
Saur, I.; Scopelliti, R.; Severin, K. Chem.sEur. J. 2006, 12, 1058.
(g) Barboiu, M.; Petit, E.; van der Lee, A.; Vaughan, G. Inorg. Chem.
2006, 45, 484. (h) Legrand, Y.- M.; van der Lee, A.; Barboiu, M.
Inorg. Chem. 2007, 46, 9540. (i) Ghosh, S.; Turner, D. R.; Batten,
S. R.; Mukherjee, P. S. Dalton Trans. 2007, 1869. (j) Barboiu, M.;
Dumitru, F.; Legrand, Y.-M.; Petit, E.; van der Lee, A. Chem.
Commun. (Cambridge) 2009, 2192.
(7) (a) Chi, K.-W.; Addicott, C.; Arif, A. M.; Stang, P. J. J. Am. Chem.
Soc. 2004, 126, 16569. (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. 2008, 73, 6580. (c) Rang, A.; Nieger, M.; Engeser,
M.; Lu¨tzen, A.; Schalley, C. A. Chem. Commun. (Combridge) 2008,
4789.
(1) Paul, D. B. J. Hist. Biol. 1988, 21, 411, and references cited therein.
(2) Schmittel, M.; Mahata, K. Angew. Chem., Int. Ed. 2008, 47, 5284.
(3) (a) Lehn, J.-M. Science 2002, 295, 2400. (b) Nitschke, J. R. Acc. Chem.
Res. 2007, 40, 103. (c) Northrop, B. H.; Zheng, Y.-R.; Chi, K.-W.;
Stang, P. J. Acc. Chem. Res. 2009, 42, 1554.
9
16544 J. AM. CHEM. SOC. 2009, 131, 16544–16554
10.1021/ja907185k CCC: $40.75  2009 American Chemical Society

From 2-Fold Completive to Intergrated Self-Sorting
A R T I C L E S
In the area of supramolecular assemblies driven by self-
sorting, Lehn et al.4 described the spontaneous formation of
helicates controlled by the number of binding sites in the ligands
as well as by the preferred coordination geometry of the metal
ions. Raymond et al.5 reported on self-sorting of supramolecular
triple helicates governed by the length of the ligands, whereas
Stang et al. elaborated the simultaneous formation of different
triangles, triangular prisms and squares, the latter being gener-
ated without formation of multiple diastereomers.7,8 Very
recently, Lehn et al. described an exhaustive study on controlling
factors of self-sorting and reported about formation of metallo-
macrocycles.9 In all cases self-sorting led either to the formation
of multiple assemblies or formation of a single species along
with free ligand. As such, the above processes were not making
quantitative use of all members of the library, and such selection
may be defined as incomplete self-sorting.
Chemists are interested in pure species, i.e., formation of a
single nanoassembly,14 and thus it is useful to exploit self-sorting
in a manner that no members of the library are left over,
generating a completiVe self-sorting library. A clever solution
to this problem was reported by Schalley et al.,12d who used
integrative12c,15 self-sorting for the formation of a cascade-
stoppered hetero[3]rotaxane. The same principle was applied
later to fabricate multicomponent rectangles using three different
components and two different interactions.12e
Among the various noncovalent binding tools to build
intricate assemblies, metal coordination has turned out to be
one of the most successful protocols. Thus, a wide range of
self-assembled architectures has been established over the past
2 decades, utilizing metal-ligand coordination protocols.16 We
have contributed to this area by developing the HETPHEN17
and HETTAP18 concepts, both of which open an easy access
to heteroleptic mononuclear complexes and equally to diverse
supramolecular architectures.18,19 In both concepts a quantitative
self-sorting involving two donors and one acceptor (AD1D2 type)
Figure 1. Self-sorting toward a five-component supramolecular isosceles
trapezoid (cartoon).
is involved, as the formation of the heteroleptic complex is
favored over that of the two competing homoleptic complexes.20
It appeared to us that the HETPHEN and HETTAP concepts
should open a venue to 2-fold completiVe self-sorting, i.e., the
event in which the combination of two incomplete self-sorting
processes develop into a completive process by making quan-
titative use of all members of the library. Herein, we present
the conceptual development of a library with four ligands and
two metal ions that gives rise to the clean formation of two
heteroleptic complexes out of 20 different possibilities; i.e.,
A1A2D1D2D3D4 f A1D1D2 + A2D3D4. As a proof of concept
and utility of 2-fold completive self-sorting, we fabricated a
sophisticated nanoassembly, a dynamic five-component su-
pramolecular isosceles trapezoid (Figure 1). We use here the
definition of a trapezoid as a quadrilateral having exactly one
pair of parallel sides, while the special case of an isosceles
trapezoid requires additionally the other pair to be of equal
length. Although the trapezoid is well present in nature, religion,
and science, it is unknown, to the best of our knowledge, as an
(8) (a) Yang, H.-B.; Ghosh, K.; Northrop, B. H.; Stang, P. J. Org. Lett.
2007, 9, 1561. (b) Zheng, Y.-R.; Yang, H.-B.; Northrop, B. H.; Ghosh,
K.; Stang, P. J. Inorg. Chem. 2008, 47, 4706. (c) Northrop, B. H.;
Yang, H.-B.; Stang, P. J. Inorg. Chem. 2008, 47, 11257. (d) Zheng,
Y.-R.; Yang, H.-B.; Ghosh, K.; Zhao, L.; Stang, P. J. Chem.sEur. J.
2009, 15, 7203.
(9) Ulrich, S.; Lehn, J.-M. J. Am. Chem. Soc. 2009, 131, 5546.
(10) (a) Jolliffe, K. A.; Timmerman, P.; Reinhoudt, D. N. Angew. Chem.,
Int. Ed. 1999, 38, 933. (b) Xu, S.; Giuseppone, N. J. Am. Chem. Soc.
2008, 130, 1826. (c) Ghosh, S.; Wu, A.; Fettinger, J. C.; Zavalij, P. Y.;
Isaacs, L. J. Org. Chem. 2008, 73, 5915. (d) Braekers, D.; Peters, C.;
Bogdan, A.; Rudzevich, Y.; Bo¨hmer, V.; Desreux, J. F. J. Org. Chem.
2008, 73, 701. (e) Ajami, D.; Hou, J.-L.; Dale, T. J.; Barrett, E.; Rebek,
J, Jr. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 10430.
(16) (a) Leininger, S.; Olenyuk, B.; Stang, P. J. Chem. ReV. 2000, 100,
853. (b) Schmittel, M.; Kalsani, V. Top. Curr. Chem. 2005, 245, 1.
(c) Severin, K. Chem. Commun. (Cambridge) 2006, 3859. (d)
Saalfrank, R. W.; Maid, H.; Scheurer, A. Angew. Chem., Int. Ed. 2008,
47, 8794. (e) Northrop, B. H.; Yang, H.-B.; Stang, P. J. Chem.
Commun. (Cambridge) 2008, 5896. (f) Zangrando, E.; Casanova, M.;
Alessio, E. Chem. ReV. 2008, 108, 4979.
(11) (a) Bilgier, B.; Xing, X.; Kumar, K. J. Am. Chem. Soc. 2001, 123,
11815. (b) Schnarr, N. A.; Kennan, A. J. J. Am. Chem. Soc. 2003,
125, 667.
(17) (a) Schmittel, M.; Ganz, A. Chem. Commun. (Cambridge) 1997, 999.
(b) Schmittel, M.; Lu¨ning, U.; Meder, M.; Ganz, A.; Michel, C.;
Herderich, M. Heterocycl. Commun. 1997, 3, 493.
(12) (a) Liu, S.; Ruspic, C.; Mukhopadhyay, P.; Chakrabarti, S.; Zavalij,
P. Y.; Isaacs, L. J. Am. Chem. Soc. 2005, 127, 15959. (b) Rekharsky,
M. V.; Yamamura, H.; Ko, Y. H.; Selvapalam, N.; Kim, K.; Inoue,
Y. Chem. Commun. (Cambridge) 2008, 2236. (c) Wang, F.; Han, C.;
He, C.; Zhou, Q.; Zhang, J.; Wang, C.; Li, N.; Huang, F. J. Am. Chem.
Soc. 2008, 130, 11254. (d) Jiang, W.; Winkler, H. D. F.; Schalley,
C. A. J. Am. Chem. Soc. 2008, 130, 13852. (e) Jiang, W.; Schalley,
C. A. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 10425. (f) Rudzevich,
Y.; Rudzevich, V.; Klautzsch, F.; Schalley, C. A.; Bo¨hmer, V. Angew.
Chem., Int. Ed. 2009, 48, 3867.
(18) (a) Schmittel, M.; Kalsani, V.; Kishore, R. S. K.; Co¨lfen, H.; Bats,
J. W. J. Am. Chem. Soc. 2005, 127, 11544. (b) Schmittel, M.; He, B.;
Mal, P. Org. Lett. 2008, 10, 2513. (c) Schmittel, M.; He, B. Chem.
Commun. (Cambridge) 2008, 4723. (d) Schmittel, M.; Mahata, K.
Inorg. Chem. 2009, 48, 822.
(19) (a) Schmittel, M.; Ganz, A.; Fenske, D. Org. Lett. 2002, 4, 2289. (b)
Schmittel, M.; Ammon, H.; Kalsani, V.; Wiegrefe, A.; Michel, C.
Chem. Commun. (Cambridge) 2002, 2566. (c) Schmittel, M.; Kalsani,
V.; Fenske, D.; Wiegrefe, A. Chem. Commun. (Cambridge) 2004, 490.
(d) Schmittel, M.; Kishore, R. S. K. Org. Lett. 2004, 6, 1923. (e)
Kishore, R. S. K.; Paulat, T.; Schmittel, M. Chem.sEur. J. 2006, 12,
8136. (f) Schmittel, M.; Kalsani, V.; Michel, C.; Mal, P.; Ammon,
H.; Ja¨ckel, F.; Rabe, J. P. Chem.sEur. J. 2007, 13, 6223.
(20) D ) donor, A ) acceptor; out of three possibilities (AD1D1, AD1D2,
AD2D2) only one complex (AD1D2) formation was observed due to
self-sorting.
(13) Rowan, S. J.; Hamilton, D. G.; Brady, P. A.; Sanders, J. K. M. J. Am.
Chem. Soc. 1997, 119, 2578.
(14) (a) Miyauchi, M.; Harada, A. J. Am. Chem. Soc. 2004, 126, 11418.
(b) Rudzevich, Y.; Rudzevich, V.; Moon, C.; Schnell, I.; Fischer, K.;
Bo¨hmer, V. J. Am. Chem. Soc. 2005, 127, 14168.
(15) An integratiVe self-sorting system is characterized by the formation
of one complex, in which more than two different subunits are bound
in two or more recognition events with positional control; see ref 12d.
9
J. AM. CHEM. SOC. VOL. 131, NO. 45, 2009 16545

A R T I C L E S
Mahata and Schmittel
Scheme 2. AD1D2D3 Type Self-Sorting Using Zn2+ Ion
Chart 1. Ligands Used for Self-Sorting
additional coordination with a suitable metal ion. Indeed, it is
known9 that O-donor atoms may provide extra coordination
when the metal ion needs donating substituents in the coordina-
tion shell.
To investigate initially AD1D2D3 self-sorting, a dynamic
library was generated from the three different ligands 1-3 in
the presence of Zn2+ (Scheme 2) to evaluate the controlling
factors (sterics, maximum site occupancy, and π-π stacking).
Ligands 1-3 and Zn2+ were taken up in equimolar amounts
(1:1:1:1) in CD3CN/CDCl3 (3:1) and were sonicated at 50 °C
for 2 h. The resultant yellow solution was characterized by 1H
NMR and electrospray ionization mass spectroscopy. Three
different ligands in the presence of a metal ion may form six
different complexes (Scheme 2). However, the number is
reduced as a combination of analytical 1H NMR and mass data
clearly suggested. Electrospray ionization mass spectra (ESI-
MS) of the library evidenced the presence of [Zn(2)(3)]2+ along
with free ligand 1 (Figure 2). A singly charged species at m/z
) 897.8 and a doubly charged species at m/z ) 374.6 were
observed after the loss of one and two OTf- counteranions,
respectively. Characteristic peaks of the other heteroleptic
complex, i.e., of [Zn(1)(3)]2+, were hardly visible.22 1H NMR
of the library also unambiguously supported the self-sorting
behavior (Figure 3). Diagnostic signals of the heteroleptic
complex [Zn(2)(3)]2+ showed up at 6.01 and 6.87 ppm,
corresponding to the aromatic substituents of 2 (doublet for a-H
at 6.01 ppm and triplet for b-H at 6.87 ppm; Chart 1). A sharp
singlet for the mesityl ArH protons of 1 at 6.17 ppm would
have been characteristic of the heteroleptic complex
[Zn(1)(3)]2+, but no such signal was detectable in that region.
Rather, a singlet at 6.95 ppm, diagnostic of the uncomplexed
ligand 1, was observed. Thus, a combination of both analytical
data suggested self-sorting behavior.
The self-sorting may be rationalized as follows: Formation
of [Zn(1)2]2+ is impossible due to steric reasons as outlined
above, while, in contrast, both other homoleptic complexes
[Zn(2)2]2+ and [Zn(3)2]2+ are feasible as well as all the
heteroleptic combinations. Our previous study,18a however,
suggests that in such case the output always favors heteroleptic
combination. The 1:1:1:1 stoichiometry along with maximum
site occupancy and full consumption of the zinc salt incites either
the formation of a heteroleptic complex along with a free ligand
or a mixture of heteroleptic complexes with some amount of
the two shielded ligands left unused. The heteroleptic combina-
tion [Zn(1)(2)]2+ is less likely due to high steric crowding at
the metal ion. The most probable outcomes are, thus,
[Zn(1)(3)]2+ or [Zn(2)(3)]2+. The higher thermodynamic stability
of the [Zn(2)(3)]2+ may arise from the additional coordination
of one methoxy group to yield a hexacoordinated zinc ion and/
or the higher σ basicity of the methoxy groups substituted ligand
2. In contrast, the other heteroleptic complex [Zn(1)(3)]2+ can
only realize pentacoordination of the zinc center.
Scheme 1. Self-Sorting Leading to Heteroleptic Complex
Formation
artificial supramolecular assembly and can be considered as of
higher order than other literature-known quadrilaterals, such as
the square, rectangle, etc.21
2. Results and Discussion
2.1. Self-Sorting in a AD1D2D3 Library: Using the Zn2+
Ion. To stimulate self-sorting among multiple (n g 2)
donor-acceptor pairs, we have evaluated and refined the
HETPHEN and HETTAP approaches.17,18 Both concepts are
based on the use of 1, a bulky 2,9-diaryl-substituted phenan-
throline (Chart 1), whose front-side shielding prevents the
formation of the homoleptic [M(1)2]n+ with Mn+ ) Cu+, Ag+,
Zn2+. However, in combination with a sterically unassuming
counterpart (e.g., 3 or 4) ligand 1 will furnish quantitatively
the heteroleptic metal complex [M(1)(3 or 4)]n+, given the
correct stoichiometry. Similarly, for 2 the homoleptic complex
[Cu(2)2]+ is precluded in MeCN, whereas [Zn(2)2]2+ is
possible.18a Hence, for copper(I) the heteroleptic complexes
[Cu(2)(3 or 4)]+ are accessible. All heteroleptic complexes [M(1
or 2)(3 or 4)]n+ profit from the extra thermodynamic stability
arising by the π-π interaction of the aryl groups of 1,2 and
ligand 3,4, rendering them more stable than the homoleptic
complexes [M(3 or 4)2]n+ (Scheme 1).17,18
How can we arrange for 2-fold completive self-sorting in a
putative A1A2D1D2D3D4 library composed of four ligands 1-4
as donors and two metal ion Mn+ as acceptors? On the basis of
the HETPHEN and HETTAP concepts, we can only exclude
the formation of the homoleptic complexes [M(1)2]n+ and
[M(2)2]n+ as well as of the heteroleptic complex [M(1)(2)]n+
.
All other combinations should show up more or less. Thus, to
differentiate the energetic stability of all possible heteroleptic
complexes involving the shielded ligands 1 and 2, we needed
to instill some beneficiary factor already on the level of a smaller
AD1D2D3 library. As our previous study with two shielded
ligands did not show the required selectivity,18a we turned our
attention to ligand 2, which has four methoxy groups for
To get further insight into the role of the metal-ligand
interaction, we turned our attention to Cu+ complexes, since
(21) For the construction of squares or rectangles only one or two ligands
are necessary, whereas for the trapezoid at least three ligands are
required.
1
(22) For comparison we prepared pure [Zn(1)(3)]2+ and [Zn(2)(3)]2+: H
NMR spectra of the complexes are shown in Figure 3.
9
16546 J. AM. CHEM. SOC. VOL. 131, NO. 45, 2009

From 2-Fold Completive to Intergrated Self-Sorting
A R T I C L E S
Figure 2. ESI-MS of an equimolar mixture of 1-3 in the presence of Zn2+ in a mixture of MeCN/CHCl3 (3:1).
2.2. Self-Sorting in a AD1D2D3 Library: Using Cu+ Ion. To
set up a complementary sorting process to the above complex-
ation scenario of two shielded ligands along with one unshielded
ligand, we decided to interrogate the competition of two
unshielded ligands with a shielded ligand (Scheme 3). The
library was generated by mixing ligands 2-4 along with Cu+
in a 1:1:1:1 molar ratio at ambient atmosphere. After refluxing
the mixture in acetonitrile for 2 h, the red solution was
characterized by 1H NMR, ESI-MS, and differential pulse
voltammetry (DPV). All analytical methods attested to the
exclusive formation of [Cu(2)(4)]+ along with free 3. In
particular, the ESI-MS of the mixture showed only a peak
corresponding to [Cu(2)(4)]+ at 695.1 (Figure 4). No peak was
observed that would correspond to other complexes.
Figure 3. Partial 1H NMR (400 MHz, 298 K) spectra of (a)
[Zn(2)(3)](OTf)2, (b) [Zn(1)(3)](OTf)2, and (c) self-sorting library among
1-3 in the presence of Zn2+ in CD3CN/CD3Cl (3:1).
Scheme 3. AD1D2D3 Type Self-Sorting Using Cu+ Ion
1
Another proof of evidence arises from diagnostic H NMR
signals of the aromatic substituents of 2 (doublet for a-H and
triplet for b-H). In the independently prepared complex
[Cu(2)(4)]+ they appear at 5.72 (for a-H) and at 6.41 ppm (for
b-H), whereas those protons appear at 6.09 and at 6.89 ppm in
[Cu(2)(3)]+. In the library mixture only signals appear at 5.72
ppm and at 6.41 ppm, as shown in Figure 5, allowing to
conclude that only [Cu(2)(4)]+ is present.
Further proof of self-sorting was observed from DPV through
a comparison of the oxidation potentials of [Cu(2)(4)]+ and
[Cu(2)(3)]+. In the pure complex [Cu(2)(4)]+ the copper(I)
oxidation occurs at 0.29 VSCE, whereas in [Cu(2)(3)]+ it appears
at -0.21 VSCE. A single peak at 0.29 VSCE after self-sorting
furthermore demonstrated exclusive formation of [Cu(2)(4)]+
(Supporting Information).
their tetrahedral coordination environment is limited to four
metal-donor interactions, even if the surrounding ligands
provide more than four donor atoms.23 As a consequence, we
studied the reaction of an equimolar mixture of 1, 2, and 4 (Chart
1) in the presence of Cu+. In the clear red solution, formation
of both heteroleptic complexes [Cu(1)(4)]+ and [Cu(2)(4)]+ was
observed in almost similar quantity, as evidenced from 1H NMR
and ESI-MS (Supporting Information). The finding that the
heteroleptic complex containing the methoxy-substituted ligand
2 is not dominating precludes its higher σ basicity to be a
decisive factor. Applying this insight to the self-sorting with
zinc suggests that the dominating [Zn(2)(3)]2+ is profiting mostly
from a direct coordination of one methoxy group driving the
sorting process and that such scenario is specific to Zn2+ due
to its tendency to build octahedral complexes.
Among the six different possibilities, π-π stacking is possible
in all complexes involving 2. However, as the homoleptic
complex [Cu(2)2]+ cannot form in acetonitrile,16b the most
favorable combinations are either [Cu(2)(3)]+ or [Cu(2)(4)]+.
In Cu+ complexes, the metal ion is generally tetrahedrally
Figure 4. ESI-MS of an equimolar mixture of 2-4 in the presence of Cu+ in a mixture of MeCN/CHCl3 (3:1).
J. AM. CHEM. SOC. VOL. 131, NO. 45, 2009 16547
9

A R T I C L E S
Mahata and Schmittel
Further unambiguous confirmation of the self-sorting process
was accomplished through the electroanalytical study of the
copper(I) complexes by DPV. In DPV, the independently
prepared complexes [Cu(1)(4)]+ and [Cu(2)(4)]+ exhibited well-
separated oxidation waves with E1/2 ) 0.44 VSCE and E1/2
)
0.29 VSCE, respectively, that remained distinguishable after
mixing (Figure 7). A single oxidation peak at E1/2 ) 0.44 VSCE
from the library mixture thus confirmed the presence of only
one copper(I) complex, supporting the formation of [Cu(1)(4)]+.
The exclusive formation of the heteroleptic complexes
[Zn(2)(3)]2+ and [Cu(1)(4)]+ in the above self-sorting highlights
one more time the extra stabilization gained from π-π stacking,
sterics, and maximum site occupancy. As noticed earlier in
section 2.2, the parent phenanthroline 4 prefers to combine with
the tetracoordinate Cu+ ion (Scheme 3), leaving two choices:
the heteroleptic copper(I) complexes [Cu(1)(4)]+ and
[Cu(2)(4)]+. On the other hand, Zn2+ favors a higher coordina-
tion number than four and thus will combine preferentially with
3 in a heteroleptic zinc(II) complex. The most probable zinc(II)
complexes are thus [Zn(1)(3)]2+ and [Zn(2)(3)]2+. In light of
the stoichiometric situation chosen, a combination of either
[Cu(1)(4)]+ and [Zn(2)(3)]2+ or [Cu(2)(4)]+ and [Zn(1)(3)]2+
is expected. According to the results in section 2.2, [Cu(1)(4)]+
and [Cu(2)(4)]+ have comparable stabilities. However,
[Zn(2)(3)]2+ is thermodynamically more stable than [Zn(1)(3)]2+,
biasing the global system toward [Cu(1)(4)]+ and [Zn(2)(3)]2+
as the favored combination.
2.4. Supramolecular Trapezoid. Design Criteria. To dem-
onstrate the value of our 2-fold completive self-sorting protocol
for nanostructure fabrication, we decided to exploit the selection
observed in section 2.3 for the fabrication of a five-component
supramolecular isosceles trapezoid. A look at the required
binding situation in an isosceles trapezoid (Figure 1) suggested
to control the required linkage of the unequal sides a and b to
c by utilizing the preferential formation of [Cu(2)(4)]+ and
[Zn(1)(3)]2+ (Scheme 4). As a result, the coordination properties
of 1 and 2 were implemented into the bisphenanthrolines 524
and 6 ()sides a and b), whose lengths (Chart 2) were varied in
a way to allow the unstrained construction of the trapezoid.
Likewise, the binding motifs of 3 and 4 were instigated into
the phenanthroline-terpyridine hybrid 7, which represents side
c in our approach.
Figure 5. Partial 1H NMR (400 MHz, 298 K) spectra of (a) [Cu(2)(4)](PF6),
(b) [Cu(2)(3)](PF6), and (c) self-sorting library with 2, 3, and 4 in the
presence of Cu+ in CD3CN.
coordinated by four donor atoms, even if the surrounding ligands
provide more than four of those.23 Accordingly, in [Cu(2)(3)]+
the terpyridine ligand 3 ought to behave as a bischelate 2,2′-
bipyridine, leaving one pyridine unit unused. As the 2,2′-
bipyridine is a weaker binding ligand than 4, [Cu(2)(4)]+ is
expected to be the most stable complex.
2.3. A1A2D1D2D3D4 Self-Sorting. Combining the insight from
both experiments above, we designed a 2-fold completive library
from 1-4 and Zn2+/Cu+, with the goal that self-sorting would
make full use of all constituents leading exclusively to
[Cu(1)(4)]+ and [Zn(2)(3)]2+. One has to note that self-sorting
in AD1D2D3 libraries had been effectively providing
[Zn(2)(3)]2+, but it was not successful with [Cu(1)(4)]+ as the
latter was formed in an even amount together with [Cu(2)(4)]+.
The expectation for A1A2D1D2D3D4 self-sorting was, however,
thatsgiven the right stoichiometrysthe full depletion of ligands
2 and 3 through formation of [Zn(2)(3)]2+ would drive the
formation of [Cu(1)(4)]+. These considerations suggested use
of an equimolar mixture of 1-4 (1:1:1:1) with the same amount
of Zn2+ and Cu+ (1:1). After refluxing this reaction mixture in
acetonitrile for 2 h, the resultant solution was characterized by
1H NMR, ESI-MS, and DPV. Theoretically, 20 different
complexes, all enumerated in Scheme 4, may form. Reward-
ingly, ESI-MS of the reaction mixture furnished evidence of a
highly successful self-sorting, furnishing basically only two
products. In line with such a statement, the ESI-MS showed
only two major species, a doubly charged one at m/z ) 374.7
corresponding to [Zn(2)(3)]2+ and a single one at m/z ) 659.2
corresponding to [Cu(1)(4)]+ (Figure 6).
Synthesis of the new ligand 6 was carried out along a general
procedure described by our group.24 After introducing the
aromatic groups at positions 2 and 9 of the phenanthroline 8,
two phenanthroline subunits 11 were connected by means of a
Sonogashira cross-coupling25 reaction as described in Scheme
5. For the synthesis of 7, 3-ethynylphenanthroline (12) was
reacted with the terpyridine unit 13 under solvent-free Sono-
gashira conditions.25
1
In the diagnostic region of the H NMR spectrum only one
set of signals was observed for each shielded phenanthroline 1
and 2, suggesting their involvement in only one type of complex.
A triplet at 6.87 ppm along with a doublet at 6.02 ppm
(characteristic of [Zn(2)(3)]2+) and a singlet at 5.95 ppm
(characteristic of [Cu(1)(4)]+) clearly demonstrated the dominant
presence of the heteroleptic species [Zn(2)(3)]2+ and [Cu(1)(4)]+
in solution (Supporting Information), in full agreement with the
ESI-MS data.
Despite the use of two selective heteroleptic coordination
motifs, implemented into 5-7, formation of trapezoid T is not
Scheme 4. A1A2D1D2D3D4 Type Self-Sorting Using Two Metal Ions and Four Ligands
9
16548 J. AM. CHEM. SOC. VOL. 131, NO. 45, 2009

From 2-Fold Completive to Intergrated Self-Sorting
A R T I C L E S
Figure 6. ESI-MS of an equimolar mixture of 1-4 in the presence of Zn2+ and Cu+ in MeCN.
components together and by reacting them for 2 h to achieve
self-correction. For the construction of the trapezoid T, though,
we decided to reduce mismatch and high kinetic barriers for
self-correction by applying a sequential addition protocol. Thus,
a mixture of 6 and 7 was first treated with Zn2+ (1:2:2 ratio) in
CD3CN. After sonication at 60 °C for 2 h, ligand 5 and Cu+
(1:2 ratio) were added, and the resulting mixture was further
sonicated at the same temperature for 8 h. The clear orange-
red solution was characterized by ESI-MS, 1H NMR, and DPV.
ESI-MS spectra showed dominantly peaks corresponding to
T ) [Cu2Zn2(5)(6)(7)2](OTf)4(PF6)2; as in the spectral region
of m/z ) 150-2000 a series of intense peaks (at m/z ) 614.4,
766.5, 995.6 and 1376.9) corresponding to the trapezoid was
detected (Figure 8). The most abundant peaks showed up at
m/z ) 614.4 and at m/z ) 766.5 that correspond to T after the
loss of six and five counteranions, respectively, i.e.,
[Cu2Zn2(5)(6)(7)2]6+ and [Cu2Zn2(5)(6)(7)2(OTf)]5+. Except the
signals characteristic for the trapezoid, some fragmentations
were also observed in negligible percentage. The five-charged
Figure 7. Differential pulse voltammograms of [Cu(1)(4)]+ (black),
[Cu(2)(4)]+ (red), a mixture of [Cu(2)(4)]+ and [Cu(1)(4)]+ (green), and
the self-sorting library (blue). All measurements were done in acetonitrile
with 0.1 M nBu4NPF6 as electrolyte against a Ag wire as a quasi-reference
electrode and 2,4,6-triphenylpyrylium tetrafluoroborate as internal standard
(scan rate of 20 mV s-1 and a pulse height of 2 mV).
peak at m/z
)
519.0 should be associated with
[CuZn2(6)(7)2(MeCN)]5+, the doubly charged one at m/z )
637.5 with [Cu2(5)(MeCN)2]2+, the triply charged peak at m/z
) 660.2 with [CuZn(6)(7)]3+, and the doubly charged one at
m/z ) 957.4 with [Zn(6)(7)]2+. Most importantly, we did not
observe any signals corresponding to the large rectangle RL )
necessarily warranted (Scheme 6). Theoretically, three ligands
in the presence of two different types of metal ions may lead to
several self-assembled structures. However, considering the
stability of the different complexes as a result of the chosen
heteroleptic coordination motifs (see sections 2.1-2.3), one may
expect the formation of three architectures, as depicted in Scheme
6: a trapezoid T, a small rectangle RS ()[Cu2Zn2(5)2(7)2]6+), and
a large rectangle RL () [Cu2Zn2(6)2(7)2]6+). If we predict the
relative stability of each architecture, then RS is expected to be
the least stable entity due to the fact that it contains the motif
[Zn(1)(3)](OTf)2 being less stable than [Zn(2)(3)](OTf)2.
[Cu2Zn2(6)2(7)2]6+ or the small rectangle RS
)
[Cu2Zn2(5)2(7)2]6+. The absence of characteristic signals for RS
(m/z ) 568.8, 712.5, 926.8, and 1285.2; Supporting Information)
allowed us to conclude that the formation of the trapezoid is
basically quantitative.
The exclusive formation of the trapezoid T was further
evidenced from the 1H NMR spectrum, in particular analyzing
the signals of bisphenanthroline 6 as it contains several protons
(b-H, b′-H, c-H, and methoxy protons) with peaks in a diagnostic
Preparation of the Trapezoid. The self-sorting of small
building units in sections 2.1 and 2.2 was effected by mixing
Figure 8. ESI-MS spectrum of a solution of trapezoid T in acetonitrile along with isotopic distributions (black, experimental; red, calculated) for
[Cu2Zn2(5)(6)(7)2](OTf)5+
.
9
J. AM. CHEM. SOC. VOL. 131, NO. 45, 2009 16549

A R T I C L E S
Mahata and Schmittel
Chart 2. Ligands Used for Synthesis of the Supramolecular Trapezoid T
Scheme 5. Synthesis of Ligands 6 and 7
region. Due to the stereogenic axis formed at the HETPHEN
1H NMR spectra of the trapezoid and the rack showed that
diagnostic signals (methoxy-, b-, and b′-H of ligand 6) appeared
in identical regions of both spectra: in the spectrum of rack
[Zn2(6)(3)2](OTf)4 (Figure 9a) these protons appeared at
2.76-2.81 ppm (methoxy protons) and at 6.83-6.89 ppm (b-,
b′-H), whereas in T they showed up at 2.71-2.85 ppm (methoxy
protons) and at 6.71-6.89 ppm (b-, b′-H). This agreement
confirms that in T the terpyridine moiety of 7 is exclusively
linked via a zinc metal center with 6. There is, however, a
difference between T and the rack. In T, due to the existence
of two diastereomers, the methoxy groups are diastereotopic
and hence eight singlets are expected. Due to overlap in the 1H
NMR, only five singlets were observed, though. Similarly, three
out of four triplets for b-protons and b′-protons were observed.
binding centers, two diastereomers are possible in T (similar
A
1
to RL and RLB), as indicated by the H NMR spectra (Figure
9c). Two singlets at 6.37 and at 6.40 ppm (c-H) supported the
presence of two diastereomeric trapezoids T in a ca. 1:1 ratio.
To further corroborate the constitutional connectivity of the
ligands at the metal centers of T, we synthesized the rack
assembly [Zn2(6)(3)2](OTf)4 (Scheme 7) as it mimics the
[Zn(terpy of 7)(6)]2+ centers of T closely. The rack was afforded
from a mixture of 6,3, and Zn2+ in CD3CN at 60 °C for 2 h
using the appropriate stoichiometry. A comparison between the
(23) Medlycott, E. A.; Hanan, G. S. Chem. Commun. (Cambridge) 2007,
4884.
(24) Schmittel, M.; Michel, C.; Wiegrefe, A.; Kalsani, V. Synthesis 2001,
1561.
For further corroboration of the structure of T, we prepared
(25) Liang, Y.; Xie, Y.-X.; Li, J.-H. J. Org. Chem. 2006, 71, 379.
the rectangle RL from a mixture of 6 and 7 with Cu+ and Zn2+
9
16550 J. AM. CHEM. SOC. VOL. 131, NO. 45, 2009

From 2-Fold Completive to Intergrated Self-Sorting
A R T I C L E S
Scheme 6. Synthesis of Self-Sorted Supramolecular Trapezoid T
Scheme 7. Synthesis of Rack Assembly [Zn2(6)(3)2](OTf)4
An evaluation of the methoxy peaks of 6 clearly establishes
the absence of any large rectangles RL in the self-sorting mixture
containing T (Figure 9b,c). Accordingly, in the 1H NMR of RL
the methoxy protons of 6 appear in two regions: for
[M(terpy)(6)]n+ centers they appear at 2.68-2.82 ppm, whereas
for [M(phen)(6)]n+ centers they show up at 3.19-3.39 ppm. In
the self-sorting mixture containing T no signals are present in
the region of 3.19-3.39 ppm. Unfortunately, the small rectangle
RS did not provide diagnostic 1H NMR signals for comparison.
However, since T is clearly not contaminated with RL, which
should be accompanied by RS in the same stoichiometric
amount, the presence of RS can be excluded as well.
We also studied the redox behavior of T. It is well-known in
the literature that despite a similar electronic environment the
redox potential may vary with geometry (e.g., angles between
ligands).18b As T exists in two diastereomeric forms with its
different geometric settings, two dissimilar potentials may be
expected. Experimentally, for T a broad peak was observed
(Figure 11), but deconvolution allowed the determination of two
individual redox potentials at 0.61 and 0.67 VSCE that may be
assigned to the two diastereomers. Both values are quite distinct
from E1/2 ) 0.44 VSCE found for [Cu(1)(4)](PF6). However, such
differences are not unusual and possibly due to the electronic
discrepancy of the ligands as well as the different geometric
scenarios. For a similar HETPHEN complex, a constraint
(1:1:1:1) in acetonitrile at 60 °C for 12 h. The ESI-MS of the
assembly was quite clean, exhibiting only peaks corresponding
1
to RL (Supporting Information). In contrast, the H NMR was
complicated (Figure 9b) due to the presence of several isomers
(two pairs of constitutionally different diastereomers; Figure 10).
In rectangle RL, two different types of metal ions can coordinate
A
B
to the same bisphenanthroline (as in the case of RL and RL
)
C
or to different bisphenanthrolines (as in the case of RL and
RLD). Such a complicated scenario is not possible in T as Cu+
will preferentially coordinate to 5 and Zn2+ to 6.
1
Figure 9. Partial H NMR (400 MHz, 298 K) spectra of (a) rack [Zn2(6)(3)2](OTf)4, (b) large rectangle RL, and (c) trapezoid T in CD3CN. Due to the
1
existence of four diastereomers of RL, the H NMR is complicated.
9
J. AM. CHEM. SOC. VOL. 131, NO. 45, 2009 16551

A R T I C L E S
Mahata and Schmittel
Figure 10. Four isomers of the large rectangle RL (two pairs of constitutionally different diastereomers).
Figure 11. Differential pulse voltammgram of trapezoid T in acetonitrile
with 0.1 M nBu4NPF6 as electrolyte against a Ag wire as a quasi-
Figure 12. Energy-minimized structure of the supramolecular trapezoid
T (hydrogen atoms are removed for clarity). Counter anions are not included.
reference electrode and dimethylferrocene as internal standard (scan rate
of 20 mV s-1 and a pulse height of 2 mV).
triangle, the oxidation potential was reported even higher (E1/2
) 0.83 VSCE).26 From the deconvolution we also determined
the individual population of each diastereomers ((51 ( 1):(49
( 1)). Both 1H NMR integration and deconvoluted DPV spectra
thus suggest formation of both diastereomers in equal amounts
(Supporting Information).
To get some insight into the geometry of the trapezoid, we
performed force field computations (MM+ as implemented into
Hyperchem 7.52) and molecular dynamics simulations. For the
computations, the long alkyl chains of ligand 6 were replaced
by methyl groups and no symmetry constraints were set. The
energy minimized structure of T is depicted in Figure 12. It
exhibits a Zn2+-Zn2+ distance of 1.68 nm and a Cu+-Cu+
distance of 1.25 nm. The short Zn2+-Cu+ distances amount to
1.60 and 1.62 nm, whereas the diagonal Zn2+-Cu+ distances
are 2.18 and 2.14 nm.
In summary, we have been able to prepare a trapezoid in
solution from the self-assembly of ligands 5-7 in the presence
of the metal ions, copper(I) and zinc(II). Considering the relative
stability of the architectures T, RS, and RL, RS is expected to
be the least stable entity due to the fact that it contains the motif
[Zn(1)(3)](OTf)2 being less stable than [Zn(2)(3)](OTf)2. On
the other hand, T and RL have the same situation about the
Zn2+ complex and a similar set up at the Cu+ complexation
site. Due to the difference in length of 5 and 6 in T the geometry
at the metal coordination centers is not perfect. Thus, along a
reasonable stability sequence the large rectangle RL is expected
to be the most stable entity: RL > T > RS. For the global reaction
outcome, however, not the individual energy of a species, but
the total energy of the ensemble is of importance. As we started
off with the same amount of 5 and 6, the formation of rectangle
RL will be paralleled by an equal amount of RS.
Figure 13. Relative energy of the rectangles RL and RS and of the trapezoid
T.
On the basis of the Gibbs free energy of formation of the
trapezoid ∆Gf (T) and the averaged free energy of formation
∆Gfav of the two rectangles, ∆Gfav ) [∆Gf (RS) + ∆Gf (RL)]/
2), three different scenarios, as depicted in a qualitative manner
in Figure 13, are possible. If ∆Gfav of the two rectangles is
comparable with the Gibbs free energy of formation ∆Gf (T)
of the trapezoid, i.e., 2∆∆Gf ∼ 0, then all three architectures
will form in parallel (Figure 13b).
In the case of ∆Gfav > ∆Gf(T) (Figure 13a), the system will
prefer to exist in trapezoid form. In such a case, one molecule
of RS and one molecule of RL will dynamically reassemble into
two molecules of T and the system will gain an overall energy
of 2∆∆Gf. On the contrary, there will be an equimolar mixture
of two rectangles without occurrence of T if ∆Gfav < ∆Gf(T)
(Figure 13c).
Level of Complexity. At this point it may be instructive to
evaluate the level of complexity reached with the preparation
of T, because, clearly, not all self-sorting systems are of the
same quality. As a quantifiable criterion for the level of
complexity, the degree of self-sorting (M) may be defined as
(26) Schmittel, M.; Mahata, K. Chem. Commun. (Cambridge) 2008, 2550.
9
16552 J. AM. CHEM. SOC. VOL. 131, NO. 45, 2009

From 2-Fold Completive to Intergrated Self-Sorting
A R T I C L E S
3
3
8.4 Hz, 2 H, a-H), 7.32 (t, J ) 8.4 Hz, 1 H, b-H), 7.57 (dd, J )
M ) P/P0 with P representing the number of possibilities and
P0 representing the number of produced complexes in the
mixture. For the clean generation of a mononuclear HETPHEN
or HETTAP complex, a situation typical for a AD1D2 type self-
sorting, P ) 3 and thus M becomes 3. Self-sorting described in
sections 2.1 and 2.2 is of the AD1D2D3 type. Here M ) 6, as
only one product is formed out of six possibilities. For the self-
sorting described in section 2.3 involving a A1A2D1D2D3D4 type
setting providing two products, M ) 10. Unfortunately, in the
case of the trapezoid the determination of the level of complexity
along the above formula becomes impracticable as there is an
infinite number of oligomeric aggregates possible.
3
3
8.0 Hz, J ) 4.4 Hz, 1 H, 8-H), 7.74 (d, J ) 8.8 Hz, 1 H, 6-H),
3
3
4
7.77 (d, J ) 8.8 Hz, 1 H, 5-H), 8.20 (dd, J ) 8.0 Hz, J ) 1.8
Hz, 1 H, 7-H), 8.35 (s, 1 H, 4-H), 9.19 (dd, 3J ) 4.4 Hz, 4J ) 1.8
Hz, 1 H, 9-H); 13C NMR (100 MHz, CDCl3) δ -0.3, 55.9, 99.6,
102.5, 104.0, 119.0, 120.9, 122.8, 126.1, 126.7, 126.8, 129.1, 129.6,
135.8, 138.4, 145.0, 146.3, 150.5, 158.3, 158.5; IR (KBr) ν 3420,
3000, 2956, 2899, 2835, 2360, 2149, 1616, 1599, 1588, 1549, 1489,
1474, 1453, 1433, 1411, 1401, 1306, 1290, 1250, 1223, 1175, 1111,
1057, 1035, 1027, 997, 906, 860, 842, 829, 822, 787, 763, 736,
691, 639; ESI-MS m/z (%) 413.2 (100) [M + H]+. Anal. Calcd for
C25H24N2O2Si: C, 72.78; H, 5.86; N, 6.79. Found: C, 72.49; H,
5.80; N, 6.74.
2,9-Bis(2,6-dimethoxyphenyl)-3-(trimethylsilanylethynyl)[1,10]-
1
phenanthroline (10). Yield 60%; mp > 260 °C; H NMR (400
3. Conclusions
MHz, CDCl3) δ 0.05 (s, 9 H, SiMe3), 3.69 (s, 6 H, OMe), 3.73 (s,
6 H, OMe), 6.61 (d, 3J ) 8.4 Hz, 2 H, [a/a′]-H), 6.65 (d, 3J ) 8.4
In conclusion, we have utilized steric effects, π-π interactions,
electronic effects, and metal-ion interactions to control integrative
self-sorting toward the formation of the trapezoid T. To the best
of our knowledge, T is the first supramolecular trapezoid, a
dynamic entity that contains three different bifunctional ligands
and two different metal ions. Such structural diversity in a small
aggregate may be applied to combine molecular subunits, even
leading to emergent molecular machines.27 Furthermore, such
diversity may be translated into 3D architectures, as 3D architec-
tures are often combinations of 2D architectures. For example, a
cube is a collection of several squares.
To develop a self-sorting strategy for T, the individual
coordination units of the three bifunctional ligands had to be
optimized. Along with a search to improve formation of
mononuclear heteroleptic complexes in mixtures of ligands and
metal ions, we have investigated increasingly more complex
libraries starting from AD1D2 type self-sorting (M ) 3), to
AD1D2D3 self-sorting (M ) 6) toward 2-fold completive self-
sorting described in a A1A2D1D2D3D4 library (M ) 10).
3
3
Hz, 2 H, [a/a′]-H), 7.29 (t, J ) 8.4 Hz, 1 H, b-H), 7.29 (t, J )
3
3
8.4 Hz, 1 H, b′-H), 7.62 (d, J ) 8.0 Hz, 1 H, 8-H), 7.74 (d, J )
3
3
8.8 Hz, 1 H, 6-H), 7.80 (d, J ) 8.8 Hz, 1 H, 5-H), 8.21 (d, J )
8.0 Hz, 1 H, 7 H), 8.35 (s, 1 H, 4-H); 13C NMR (100 MHz, CDCl3)
δ -0.3, 56.1, 56.4, 99.0, 102.7, 104.5, 105.1, 119.3, 120.6 (2C),
125.7, 126.2 126.9 (2C), 128.0, 129.7 (2C), 135.3, 138.6, 145.0,
146.0, 155.2, 157.4, 158.5 (2C); IR (KBr) ν 3417, 3001, 2957, 2939,
2898, 2836, 2152, 1643, 1616, 1599, 1590, 1538, 1505, 1473, 1458,
1432, 1412, 1397, 1359, 1305, 1286, 1250, 1214, 1185, 1173, 1112,
1068, 1036, 1023, 997, 912, 891, 859, 843, 781, 761, 732, 643;
ESI-MS m/z (%) 549.3 (100) [M + H]+. Anal. Calcd for
C33H32N2O4Si·H2O: C, 69.94; H, 6.05; N, 4.94. Found: C, 70.35;
H, 5.75; N, 4.93.
2,9-Bis(2,6-dimethoxyphenyl)-3-ethynyl[1,10]phenanthroline
(11). Yield 97%; mp > 260 °C; H NMR (400 MHz, CDCl3) δ
3.05 (s, 1 H, ethynyl), 3.71 (s, 6 H, OMe), 3.73 (s, 6 H, OMe),
1
3
3
6.63 (d, J ) 8.4 Hz, 2 H, [a/a′]-H), 6.66 (d, J ) 8.4 Hz, 2 H,
[a/a′]-H), 7.30 (t, 3J ) 8.4 Hz, 1 H, [b/b′]-H), 7.31 (t, 3J ) 8.4 Hz,
3
3
1 H, [b/b′]-H), 7.64 (d, J ) 8.0 Hz, 1 H, 8-H), 7.76 (d, J ) 8.8
3
3
Hz, 1 H, 6-H), 7.82 (d, J ) 8.8 Hz, 1 H, 5-H), 8.22 (d, J ) 8.0
Hz, 1 H, 7 H), 8.42 (s, 1 H, 4-H); 13C NMR (100 MHz, CDCl3) δ
56.2, 56.4, 80.9, 81.3, 104.6, 105.1, 118.9, 119.7, 120.4, 125.6,
126.4, 126.9, 127.0, 128.0, 129.8, 129.9, 135.4, 139.9, 145.3, 146.0,
155.3, 156.8, 158.5 (2C); IR (KBr) ν 3414, 3285, 3002, 2939, 2904,
2836, 1641, 1620, 1598, 1590, 1539, 1505, 1473, 1458, 1431, 1413,
1398, 1358, 1304, 1286, 1251, 1215, 1173, 1109, 1065, 1033, 1022,
988, 918, 892, 842, 781, 766, 731, 652; ESI-MS m/z (%) 477.2
(100) [M + H]+. Anal. Calcd for C30H24N2O4 ·H2O: C, 72.86; H,
5.30; N, 5.66. Found: C, 72.68; H, 5.16; N, 5.42.
4. Experimental Section
4.1. General Methods. All commercial reagents were used
without further purification. The solvents were dried with appropri-
ate desiccants and distilled prior to use. Silica gel 60 (70-230 mesh)
was used for column chromatography. 1H NMR and 13C NMR were
recorded on a Bruker Avance 400 MHz spectrometer using the
deuterated solvent as the lock and residual solvent as the internal
reference. NMR measurements were carried out at 298 K. The
following abbreviations were utilized to describe peak patterns: s
) singlet, d ) doublet, t ) triplet, and m ) multiplet. The
numbering of the carbon atoms of the molecular formulas shown
in the Experimental Section is only used for the assignments of
the NMR signal and is not in accordance with the IUPAC
nomenclature rules. Electrospray ionization mass spectra were
recorded on a Thermo-Quest LCQ Deca. Differential pulse volta-
mmetry was measured on a Parstat 2273 in dry acetonitrile. Melting
points were measured on a Bu¨chi SMP-20 and are uncorrected.
Infrared spectra were recorded using a Varian 1000 FT-IR instru-
ment. Elemental analysis measurements were done using an EA
3000 CHNS. Precursors for 624 and 728 were synthesized according
to known procedures. The energy minimized structure was com-
puted with the MM+ force field as implemented in Hyperchem 7.52.
4.2. Characterization and Preparation of Compounds. Syn-
thesis of the new ligand 6 was carried out along a general procedure
described by our group.24
Synthesis of 6. 2,9-Bis(2,6-dimethoxyphenyl)-3-ethynyl[1,10]-
phenanthroline (11; 200 mg, 420 µmol), 1,4-bis(decyloxy)-2,5-
diiodobenzene (135 mg, 210 µmol), TBAF ·3H2O (800 mg, 2.54
mmol), and trans-PdCl2(PPh3)2 (9.50 mg, 13.5 µmol) were com-
bined in a Schlenk flask under nitrogen atmosphere. The solid
mixture was stirred at 80 °C for 12 h. Then, it was cooled, dissolved
with dichloromethane, and washed successively with aqueous KOH
(100 mL) and water (5 × 200 mL). After drying over Na2SO4, the
organic solvent was removed under reduced pressure. The crude
product was purified using column chromatography (SiO2; 19:1
CH2Cl2/EtOAc) affording 6 as yellow solid. Yield 66%; mp 247
°C; H NMR (400 MHz, CD2Cl2) δ 0.88 (t, J ) 6.8 Hz, 6 H,
CH3), 1.30-1.51 (m, 24 H, CH2), 1.54-1.59 (m, 4 H, CH2),
1.79-1.86 (m, 4 H, CH2), 3.71 (s, 12 H, OCH3), 3.72 (s, 12 H,
OCH3), 3.87 (t, 3J ) 6.8 Hz, 4 H, OCH2), 6.30 (s, 2 H, c-H), 6.71
(d, J ) 8.4 Hz, 4 H, [a/a′]-H), 6.73 (d, J ) 8.4 Hz, 4 H, [a/a′]-
H), 7.38 (t, 3J ) 8.4 Hz, 2 H, [b/b′]-H), 7.41 (t, 3J ) 8.4 Hz, 2 H,
[b/b′]-H), 7.58 (d, 3J ) 8.2 Hz, 2 H, 8-7.83 (d, 3J ) 8.8 Hz, 2 H,
6-H), 7.89 (d, J ) 8.8 Hz, 2 H, 5-H), 8.30 (d, J ) 8.2 Hz, 2 H,
7-H), 8.42 (s, 2 H, 4-H); 13C NMR (100 MHz, CD2Cl2) δ 14.5,
1
3
3
3
3
3
2-(2,6-Dimethoxyphenyl)-3-(trimethylsilanylethynyl)[1,10]-
phenanthroline (9). Yield 74%; mp 211 °C; 1H NMR (400 MHz,
3
CDCl3) δ 0.04 (s, 9 H, SiMe3), 3.68 (s, 6 H, OMe), 6.63 (d, J )
(28) Dumur, F.; Mayer, C. R.; Dumas, E.; Marrot, J.; Se´cheresse, F.
(27) Gale, P. A. Philos. Trans. R. Soc. London, A 2000, 358, 431.
Tetrahedron Lett. 2007, 48, 4143.
9
J. AM. CHEM. SOC. VOL. 131, NO. 45, 2009 16553

A R T I C L E S
Mahata and Schmittel
23.3, 26.7, 29.8, 30.0, 30.1, 30.3, 30.4, 32.5, 56.5, 56.6, 70.2, 91.8,
93.0, 104.3, 104.4, 114.3, 117.9, 119.1, 120.0, 121.3, 126.3 (2C),
127.6, 127.7, 128.5, 130.3, 130.4, 136.3, 138.5, 145.5, 146.6, 153.5,
156.3, 157.5, 158.7, 159.0; IR (KBr) ν 3426, 3006, 2926, 2853,
2838, 1636, 1616, 1599, 1590, 1538, 1500, 1473, 1459, 1431, 1391,
1357, 1304, 1285, 1276, 1250, 1214, 1173, 1111, 1064, 1033, 1022,
996, 910, 891, 844, 781, 767, 732, 655; ESI-MS m/z (%) 1339.8
(100) [M + H]+. Anal. Calcd for C86H90N4O10 ·2H2O: C, 75.08;
H, 6.89; N, 4.07. Found: C, 75.11; H, 6.71; N, 4.04.
(s, 3 H, CH3), 1.50 (s, 3 H, CH3), 1.58 (s, 3 H, CH3), 1.62 (s, 3 H,
CH3), 1.67 (s, 3 H, CH3), 1.69 (s, 6 H, CH3), 1.71 (s, 3 H, CH3),
1.73 (s, 3 H, CH3), 1.76 (s, 3 H, CH3), 1.78 (s, 3 H, CH3), 1.82 (s,
3 H, CH3), 2.71 (s, 3 H, OCH3), 2.72 (s, 3 H, OCH3), 2.73 (s, 6 H,
OCH3), 2.84 (s, 9 H, OCH3), 2.85 (s, 3 H, OCH3), 3.59 (t, 3J ) 6.4
Hz, 2 H, OCH2), 3.61 (t, 3J ) 6.4 Hz, 2 H, OCH2), 5.87-6.08 (m,
12 H, mesityl-, a-, a′-H), 6.37 (s, 1 H, c-H) 6.40 (s, 1 H, c-H) 6.73
(t, 3J ) 8.0 Hz, 1 H, [b/b′]-H), 6.75 (t, 3J ) 8.0 Hz, 1 H, [b/b′]-H),
3
6.87 (t, J ) 8.4 Hz, 2 H, [b/b′]-H), 7.46-7.50 (m, 4 H, f′′-H),
Synthesis of 7. A procedure similar to that for the preparation
of 6 was followed by reacting the corresponding precursors 12 (100
mg, 489 µmol) and 13 (213 mg, 489 µmol) with trans-PdCl2(PPh3)2
(20.0 mg, 28.5 µmol) and TBAF ·3H2O (925 mg, 2.93 mmol). Yield
7.62-7.66 (m, 4 H, g′′-H), 7.76-7.88 (m, 4 H, 8′′-, a′′-H), 7.92
(d,3J ) 8.4 Hz, 2 H, a′′-H), 7.96-8.08 (m, 8 H, 5′′-, 6′′-, 8-, 8′-H),
8.17-8.25 (m, 9 H, b′′-, e′′-,6′-H), 8.29 (d,3J ) 9.2 Hz, 1 H, 6′-
H), 8.31 (d,3J ) 9.2 Hz, 1 H, 5′-H), 8.36 (d,3J ) 9.2 Hz, 1 H,
5′-H), 8.42 (d,3J ) 9.2 Hz, 2 H, 6-H), 8.49-8.51 (m, 10 H, 4′′-,
5-, [2′′/9′′]-, d′′-H), 8.61 (s, 2 H, c′′-H), 8.64 (d,3J ) 7.2 Hz, 2 H,
[2′′/9′′]-H), 8.68 (s, 2 H, c′′-H), 8.72-8.75 (m, 2 H, 7′′-H), 8.82
1
3
83%; mp 242 °C; H NMR (400 MHz, CDCl3) δ 7.38 (ddd, J )
3
4
3
7.6 Hz, J ) 4.8 Hz, J ) 1.2 Hz, 2 H, f′′-H), 7.66 (d, J ) 8.2
Hz,3J ) 4.4 Hz, 1 H, 8′′-H), 7.78 (d, 3J ) 8.4 Hz, 2 H, a′′-H), 7.81
(d, 3J ) 8.8 Hz, 1 H, 6′′-H), 7.85 (d, 3J ) 8.8 Hz, 1 H, 5′′-H), 7.90
(dt, 3J ) 7.6 Hz, 4J ) 2.0 Hz, 2 H, e′′-H), 7.97 (d, 3J ) 8.4 Hz, 2
H, b′′-H), 8.27 (dd, 3J ) 8.2 Hz, 4J ) 1.8 Hz, 1 H, 7′′-H), 8.44 (d,
3
3
(d, J ) 8.4 Hz, 1 H, [7/7′]-H), 8.84 (d, J ) 8.4 Hz, 1 H, [7/7′]-
3
H), 8.90 (s, 1 H, [4/4′]-H), 9.04 (d, J ) 8.4 Hz, 1 H, [7/7′]-H),),
9.04 (d, J ) 8.4 Hz, 1 H, [7/7′]-H), 9.06 (s, 1 H, [4/4′]-H), 9.07
3
4J ) 2.0 Hz, 1 H, 4′′-H), 8.69 (d, J ) 7.6 Hz, 2 H, d′′-H), 8.75
3
(s, 1 H, [4/4′]-H), 9.11 (s, 1 H, [4/4′]-H); IR (KBr) ν 3475, 3072,
2924, 2853, 2361, 2341, 2211, 1617, 1602, 1590, 1575, 1549, 1500,
1476, 1428, 1403, 1363, 1278, 1256, 1224, 1160, 1111, 1070, 1031,
1018, 914, 843, 791, 747, 736, 725, 694, 668, 659, 638; ESI-MS
m/z (%) 614.4 (100), [M - 2PF6, 4OTf]6+, 766.5 (90), [M - 2PF6,
3OTf]5+, 995.6 (18), [M - 2PF6, 2OTf]4+, 1376.9 (8) [M - 2PF6,
3
4
5
(ddd, J ) 4.8 Hz, J ) 2.0 Hz, J ) 1.0 Hz, 2 H, g′′-H), 8.78 (s,
3
4
2 H, c′′-H), 9.22 (dd, J ) 4.4 Hz, J ) 1.8 Hz, 1 H, 9′′-H), 9.33
(d, 4J ) 2.0 Hz, 1 H, 2′′-H); 13C NMR (100 MHz, CD2Cl2) δ 88.4,
93.6, 119.0, 119.8, 121.7, 123.8 (2C), 124.5, 126.6 127.9, 128.0,
128.4, 129.7, 132.9, 136.5, 137.4, 138.6, 139.5, 145.5, 146.6, 149.6,
149.7, 151.0, 152.4, 156.5, 156.7; IR (KBr) ν 3408, 2360, 2207,
1654, 1604, 1585, 1566, 1542, 1514, 1502, 1467, 1443, 1421, 1388,
1264, 1099, 1076, 1039, 991, 906, 831, 788, 729, 687, 661; ESI-
OTf]3+
. Anal. Calcd for C226H188Br2Cu2F24N18O22P2S4Zn2 ·
2H2O: C, 58.91; H, 4.20; N, 5.47; S, 2.78. Found: C, 58.45; H,
3.69; N, 5.28; S, 2.51.
MS m/z (%) 512.3 (100) [M
C35H21N5 ·CH2Cl2: C, 72.49; H, 3.89; N, 11.74. Found: C, 72.52;
H, 3.79; N, 11.46.
+
H]+. Anal. Calcd for
Acknowledgment. We are thankful to the Deutsche For-
schungsgemeinschaft and the University of Siegen for financial
support of this research. Support from the DAAD for K.M.
(Promotionsabschlussstipendium) is gratefully acknowledged.
Self-Assembly of Trapezoid T. At ambient atmosphere, a
mixture of 6 (5.20 mg, 3.88 µmol), 7 (3.97 mg, 7.76 µmol), and
Zn(OTf)2 (2.82 mg, 7.76 µmol) was refluxed in acetonitrile for 2 h
until a clear yellow solution was obtained. After addition of solid
[Cu(MeCN)4](PF6) (2.89 mg, 7.76 µmol) and 5 (4.13 mg, 3.88
µmol) the solution was further refluxed for 8 h. The resulting
orange-red solution was characterized without further purification.
Supporting Information Available: 1H NMR, 13C NMR, and
ESI spectra of all relevant compounds and details of the DPV
investigations and the deconvolution. This material is available
free of charge via the Internet at http://pubs.acs.org.
1
Yield quantitative; mp > 260 °C; H NMR (400 MHz, CD3CN) δ
3
0.78 (t, J ) 6.8 Hz, 6 H, CH3), 1.17-1.20 (m, 28 H, CH2), 1.37
(s, 3 H, CH3) 1.37-1.40 (m, 4 H, CH2), 1.45 (s, 3 H, CH3), 1.48
JA907185K
9
16554 J. AM. CHEM. SOC. VOL. 131, NO. 45, 2009