A triangle-square equilibrium of metallosupramolecular assemblies based on Pd(II) and Pt(II) corners and diazadibenzoperylene bridging ligands

Article abstract of DOI:10.1021/ja004360y

Tetraaryloxy-substituted diazadibenzoperylene bridging ligands 1a,b were employed in transition metal-directed self-assembly with Pd(II) and Pt(II) phosphane triflates 2a,b which resulted in complex dynamic equilibria between molecular triangles 3a-d and molecular squares 4a-d in solution. Characterization of the equilibria and assignment of the metallacycles was accomplished by ¹H and ³¹P{¹H} NMR spectroscopy in combination with electrospray ionization Fourier transform ion cyclotron mass spectrometry (ESI-FTICR-MS). It was found that the equilibria depend on several factors, such as the metal ion (Pd²⁺ or Pt²⁺), the solvent, and the steric demand of the phenoxy substituents of the diazadibenzoperylene ligands 1a,b. Introduction of bulky tert-butyl groups in 1b shifts the equilibrium significantly in the direction of the molecular squares. Molecular dynamics simulations of the triangle and square structures revealed critical steric effects and restricted conformational flexibilities of the phosphane and diazadibenzoperylene ligands that help explain the distinct dynamic behavior observed in variable-temperature NMR studies. Concentration-dependent UV/vis and fluorescence spectroscopy revealed the limited stability of the assemblies and confirmed the reversible nature of the dynamic equilibria.

Full text of DOI:10.1021/ja004360y

5424
J. Am. Chem. Soc. 2001, 123, 5424-5430
A Triangle-Square Equilibrium of Metallosupramolecular Assemblies
Based on Pd(II) and Pt(II) Corners and Diazadibenzoperylene
Bridging Ligands
†
‡
‡
,†
Armin Sautter, Dietmar G. Schmid, G u1 nther Jung, and Frank W u1 rthner*
Contribution from the Department of Organic Chemistry II, UniVersity of Ulm, Albert-Einstein-Allee 11,
D-89081 Ulm, and the Department of Organic Chemistry, UniVersity of T u¨ bingen,
Auf der Morgenstelle 18, D-72076 T u¨ bingen, Germany
ReceiVed December 27, 2000. ReVised Manuscript ReceiVed March 15, 2001
Abstract: Tetraaryloxy-substituted diazadibenzoperylene bridging ligands 1a,b were employed in transition
metal-directed self-assembly with Pd(II) and Pt(II) phosphane triflates 2a,b which resulted in complex dynamic
equilibria between molecular triangles 3a-d and molecular squares 4a-d in solution. Characterization of the
1
31
1
equilibria and assignment of the metallacycles was accomplished by H and P{ H} NMR spectroscopy in
combination with electrospray ionization Fourier transform ion cyclotron mass spectrometry (ESI-FTICR-
2
+
2+
MS). It was found that the equilibria depend on several factors, such as the metal ion (Pd or Pt ), the
solvent, and the steric demand of the phenoxy substituents of the diazadibenzoperylene ligands 1a,b. Introduction
of bulky tert-butyl groups in 1b shifts the equilibrium significantly in the direction of the molecular squares.
Molecular dynamics simulations of the triangle and square structures revealed critical steric effects and restricted
conformational flexibilities of the phosphane and diazadibenzoperylene ligands that help explain the distinct
dynamic behavior observed in variable-temperature NMR studies. Concentration-dependent UV/vis and
fluorescence spectroscopy revealed the limited stability of the assemblies and confirmed the reversible nature
of the dynamic equilibria.
4
d,7b
Introduction
and applications such as anion recognition
and electrochemi-
4
k
cal sensing. Recently, the incorporation of functional dye
ligands, based on perylene bisimides and porphyrins, into
metallacycles opened up a promising avenue toward artificial
Transition metal-directed self-assembly of macrocycles and
cage compounds has received considerable attention lately,
leading to more and more complex architectures, for example,
1
4
f-g,10
1b,2,3
3,4,5
light-harvesting systems.
molecular triangles,
squares and rectangles,
pentagons
6
7
and hexagons, catenanes, nanotubes and polytubes, bowls and
8
9
(4) (a) Fujita, M.; Yazaki, J.; Ogura, K. J. Am. Chem. Soc. 1990, 112,
645-5647. (b) Stang, P. J.; Cao, D. H. J. Am. Chem. Soc. 1994, 116,
981-4982. (c) Stang, P. J.; Olenyuk, B. Acc. Chem. Res. 1997, 30, 502-
cages, cuboctahedra and dodecahedra among many others. The
possibility to use a wide variety of transition metal ions
diversifies this field considerably. The focus of current research
is on the introduction of functionality into these systems such
5
4
518. (d) Slone, R. V.; Yoon, D. I.; Calhoun, R. M.; Hupp, J. T. J. Am.
Chem. Soc. 1995, 117, 11813-11814. (e) Slone, R. V.; Hupp, J. T.; Stern,
C.; Albrecht-Schmitt; T. E. Inorg. Chem. 1996, 35, 4096-4097. (f)
W u¨ rthner, F.; Sautter, A. Chem. Commun. 2000, 445-446. (g) W u¨ rthner,
F.; Sautter, A.; Schmid, D. G.; Weber, P. J. A. Chem. Eur. J. 2001, 7,
894-902. (h) Woessner, S. M.; Helms, J. B.; Houlis, J. F.; Sullivan, B. P.
Inorg. Chem. 1999, 38, 4380-4381. (i) Sun, S.-S.; Silva, A. S.; Brinn, I.
M.; Lees, A. J. Inorg. Chem. 2000, 39, 1344-1345. (j) Campos-Fernandez,
C. S.; Clerac, R.; Dunbar, K. R. Angew. Chem., Int. Ed. 1999, 38, 3477-
3479. (k) Lahav, M.; Gabai, R.; Shipway, A. N.; Willner, I. Chem. Commun.
1999, 1937-1938.
as redoxactivity,4 magnetic or luminescence properties,
,5
4j
4,5,10
*
To whom correspondence should be addressed.
Department of Organic Chemistry II, University of Ulm.
Department of Organic Chemistry, University of T u¨ bingen.
†
‡
(
1) (a) Lehn, J.-M. Supramolecular Chemistry, VCH: Weinheim, 1995.
b) Leininger, S.; Olenyuk, B.; Stang, P. J. Chem. ReV. 2000, 100, 853-
08 and references therein.
2) (a) Hall, J.; Loeb, S. J.; Shimizu, G. K. H.; Yap, G. P. A. Angew.
(
9
(
Chem., Int. Ed. 1998, 37, 121-123. (b) Thompson, A.; Rettig, S. J.; Dolphin,
D. Chem. Commun. 1999, 631-632. (c) Habereder, T.; Warchhold, M.;
N o¨ th, H.; Severin, K. Angew. Chem., Int. Ed. 1999, 38, 3225-3228. (d)
Lai, S.-W.; Chan, M. C.-W.; Peng, S.-M.; Che, C.-M. Angew. Chem., Int.
Ed. 1999, 38, 669-671. (e) Schnebeck, R.-D.; Freisinger, E.; Glahe, F.;
Lippert, B. J. Am. Chem. Soc. 2000, 122, 1381-1390. (f) Schnebeck, R.-
D.; Randaccio, L.; Zangrando, E.; Lippert, B. Angew. Chem., Int. Ed. 1998,
(5) (a) Benkstein, K. D.; Hupp, J. T.; Stern, C. L. J. Am. Chem. Soc.
1998, 120, 12982-12983. (b) Benkstein, K. D.; Hupp, J. T.; Stern, C. L.
Angew. Chem., Int. Ed. 2000, 39, 2891-2893. (c) Benkstein, K. D.; Hupp,
J. T.; Stern, C. L. Inorg. Chem. 1998, 37, 5404-5405. (d) Woessner, S.
M.; Helms, J. B.; Shen, Y.; Sullivan, B. P. Inorg. Chem. 1998, 37, 5406-
5407. (e) Hartmann, H.; Berger, S.; Winter, R.; Fiedler, J.; Kaim, W. Inorg.
Chem. 2000, 39, 4977-4980.
(6) (a) Stang, P. J.; Persky, N. E.; Manna, J. J. Am. Chem. Soc. 1997,
119, 4777-4778. (b) Matsumoto, N.; Motoda, Y.; Matsuo, T.; Nakashima,
T.; Re, N.; Dahan, F.; Tuchagues, J.-P. Inorg. Chem. 1999, 38, 1165-
1173. (c) Mamula, O.; von Zelewsky, A.; Bernardinelli, G. Angew. Chem.,
Int. Ed. 1998, 37, 289-293. (d) Hasenknopf, B.; Lehn, J.-M.; Baum, G.;
Kneisel, B. O.; Fenske, D. Angew. Chem., Int. Ed. Engl. 1996, 35, 1838-
1840.
3
7, 119-121. (g) Schnebeck, R.-D.; Freisinger, E.; Lippert, B. Chem.
Commun. 1999, 675-676. (h) Barbera, J.; Elduque, A.; Gimenez, R.; Oro,
L. A.; Serrano, J. L. Angew. Chem., Int. Ed. Engl. 1996, 35, 2832-2835.
(3) (a) Fujita, M.; Ogura, K. Bull. Chem. Soc. Jpn. 1996, 69, 1471-
1
482. (b) Fujita, M.; Ogura, K. Coord. Chem. ReV. 1996, 148, 249-264.
(c) Lee, S. B.; Hwang, S.; Chung, D. S.; Yun, H.; Hong, J.-I. Tetrahedron
Lett. 1998, 39, 873-876. (d) Sun, S.-S.; Lees, A. J. Inorg. Chem. 1999,
3
8, 4181-4182. (e) Sun, S.-S.; Lees, A. J. J. Am. Chem. Soc. 2000, 122,
(7) (a) Fujita, M. Acc. Chem. Res. 1999, 32, 53-61 and references
therein. (b) Aoyagi, M.; Biradha, K.; Fujita, M. J. Am. Chem. Soc. 1999,
121, 7457-7458. (c) Biradha, K.; Aoyagi, M.; Fujita, M. J. Am. Chem.
Soc. 2000, 122, 2397-2398.
8
956-8967. (f) Schnebeck, R.-D.; Freisinger, E.; Lippert, B. Eur. J. Inorg.
Chem. 2000, 1193-1200. (g) McQuillan, F. S.; Berridge, T. E.; Chen, H.;
Hamor, T. A.; Jones, C. J. Inorg. Chem. 1998, 37, 4959-4970.
1
0.1021/ja004360y CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/18/2001

J. Am. Chem. Soc., Vol. 123, No. 23, 2001 5425
There have been many concepts and ideas to predict which
species will self-assemble in solution,1,11 and these consider-
ations are mainly based on the information stored in the building
blocks, that is, the number of receptor groups, their positioning
in space, the electronic configuration of the metal ions, and the
specificity of molecular recognition. It appears from this work
that rational design of metallosupramolecular structures is a
reliable method. Thus, cis-protected square planar Pd(II) corners
will assemble to molecular squares in the presence of ditopic
linear bridging ligands, whereas angular units of 60° and 120°
tions and structural characteristics of both triangles and squares
and helped to understand the dynamics observed on the NMR
time scale. Furthermore, the optical absorption and fluorescence
properties as well as thermodynamic stabilities of the metal
complexes were investigated.
Results and Discussion
Preparation of Metallacycles and NMR Studies. Our initial
goal was the construction of functional molecular squares and
cages via cis-protected Pd(II) and Pt(II) phosphane triflates 2a,b
and the new highly fluorescent diazadibenzoperylene ligands
in combination with linear linkers suffice to form triangles and
hexagons, respectively.2
,3,6
However, there is an increasing
1
4
number of examples that deviate from the expected structures.
Often two or more species equilibrate when there is no
thermodynamically clearly favored species or when further
effects come into play such as steric strain and electrostatic
repulsion. Equilibration is facilitated by the reversible nature
of the bond-forming-bond-breaking process in the self-as-
sembly of metal-ligand systems that is strongly desired and
regarded as a self-repairing option to obtain perfect structures.
1a,b. Stoichiometric amounts of 1a or 1b and 2a or 2b were
reacted in dichloromethane at room temperature for 24 h to yield
orange complexes in 55-78% isolated yields (Scheme 1).15 In
all cases the bright greenish fluorescence of 1a,b is quenched
immediately upon mixing with 2a,b, and the solutions turn
orange. The isolated crystalline complexes were characterized
by elemental analyses (C, H, N, S) that proved the 1:1
stoichiometry for all combinations of diazadibenzoperylene
bridging ligands 1a,b and metal-phosphane corners 2a,b.16 On
the other hand NMR spectra of these complexes exhibited broad
The focus of our work is on large-sized molecular triangles
and molecular squares, where numerous examples for square
structures but considerably less for triangular structures are given
1
31
1
signals in the H and two rather broad singlets in the P{ H}
NMR spectra which did not indicate the formation of the
2
in the literature. Besides “pure” squares and triangles, also a
3
,4,13
expected molecular squares (Figure 1b).
few equilibria between molecular squares and triangles have
3
The room-temperature NMR spectra implied dynamic pro-
cesses on the NMR time scale. Therefore, variable-temperature
NMR spectra of the 1a-2a complex were recorded in deuterated
nitromethane allowing for a wide temperature window (Figure
been reported. One of the most useful corner units is given
with cis-dpppPd(II) and -Pt(II) which were introduced by Stang
et al.4b To date, these corners are reported to form exclusively
molecular squares with linear diazaligands.1
2,13
However, during
1
{
). Over the whole temperature range (253-353 K) the 31P-
14
coordination studies of diazadibenzoperylene ligands 1a,b with
these corner units we observed for the first time triangle-square
equilibria also for this system. The triangle-square equilibria
were characterized by a combination of variable-temperature
multinuclear NMR spectroscopy and ESI-FTICR (electrospray
ionization Fourier transform ion cyclotron) mass spectrometry.
Molecular modeling gave insight into the mutual steric interac-
1
H} spectra display two singlets between 10 and 11 ppm that
sharpen notably at higher temperatures. The signals are shifted
upfield by about 8 ppm with respect to the precursor complex
4
f,13
2
a, typical for aza-ligand coordination.
No uncoordinated
metal-phosphane corners could be detected. The presence of
3
1
two P resonances indicates two nonequivalent metal-phos-
1
phane corner units that are coordinated by 1a. The H NMR
spectra show broad and undefined signals at room temperature
(Figure 1b), but the signals sharpen at lower temperatures and
many defined resonances appear (Figure 1c). However, at higher
temperatures, the spectra become simpler, and at 353 K only
two sets of signals remain, two for the diazadibenzoperylene
ligands 1a and two for the dppp moiety (Figure 1a). Some
signals of the various phenyl groups overlap in the aromatic
region, but the R protons next to the coordinating nitrogens at
around 9 ppm give well-separated characteristic resonances. In
accordance there are two sets of signals present for the propyl
bridge of the dppp ligand that overlap themselves but that are
well-separated from all other signals. Notably, the integral ratios
of corresponding protons of the two sets of signals compare
(
8) (a) Yu, S.-Y.; Kusukawa, T.; Biradha, K.; Fujita, M. J. Am. Chem.
Soc. 2000, 122, 2665-2666. (b) Fujita, M.; Fujita, N.; Ogura, K.;
Yamaguchi, K. Nature 1999, 400, 52-55. (c) Fujita, M.; Yu, S.-Y.;
Kusukawa, T.; Funaki, H.; Ogura, K.; Yamaguchi, K. Angew. Chem., Int.
Ed. 1998, 37, 2082-2085.
(9) (a) Takeda, N.; Umemoto, K.; Yamaguchi, K.; Fujita, M. Nature
1
999, 398, 794-796. (b) Olenyuk, B.; Levin, M. D.; Whiteford, J. A.;
Shield, J. E.; Stang, P. J. J. Am. Chem. Soc. 1999, 121, 10434-10435. (c)
Olenyuk, B.; Whiteford, J. A.; Fechtenk o¨ tter, A.; Stang, P. J. Nature 1999,
3
98, 796-799.
10) (a) Funatsu, K.; Imamura, T.; Ichimura, A.; Sasaki, Y. Inorg. Chem.
998, 37, 1798-1804. (b) Drain, C. M.; Lehn, J.-M. J. Chem. Soc.. Chem.
(
1
Commun. 1994, 2313-2315. (c) Drain, C. M.; Nifiatis, F.; Vasenko, A.;
Batteas, J. D. Angew. Chem., Int. Ed. 1998, 37, 2344-2347. (d) Slone, R.
V.; Hupp, J. T. Inorg. Chem. 1997, 36, 5422-5423. (e) Slone, R. V.;
Belanger, S.; Hupp, J. T.; Guzei, I. A.; Rheingold, A. L. Coord. Chem.
ReV. 1998, 171, 221-243. (f) Slone, R. V.; Yoon, D. I.; Calhoun, R. M.;
Hupp, J. T. J. Am. Chem. Soc. 1995, 117, 11813-11814. (g) Iengo, E.;
Milani, B.; Zangrando, E.; Geremia, S.; Alessio, E. Angew. Chem., Int. Ed.
31
1
well to those in the P{ H} NMR spectrum.
The integral ratio of the 31P{ H} NMR signals was found to
be solvent-dependent, changing from 4:6 in deuterated aceto-
nitrile over 5:5 in acetone to about 6:4 in nitromethane. This
strongly supports the presence of a thermodynamic equilibrium
1
2
000, 39, 1096-1099.
(
11) (a) Stang, P. J. Chem. Eur. J. 1998, 4, 19-27. (b) Caulder, D. L.;
Raymond, K. N. Acc. Chem. Res. 1999, 32, 975-982. (c) Navarro, J. A.
R.; Lippert, B. Coord. Chem. ReV. 1999, 185-186, 653-667.
(
12) Hong et al. (ref 3c) reported a triangle-square equilibrium with
enPd(NO3)2 and trans-1,2-bis(4-pyridyl)ethylene. For reasons of charac-
(15) NMR experiments indicate that the self-assembly process is
1
31
1
terization by mass spectrometry they prepared the analogous [Pd(dppp)]-
quantitative because exactly the same H and P{ H} NMR spectra as for
the isolated complexes were observed after mixing stoichiometric amounts
of 1 and 2 in CDCl3 in the NMR tube. However, owing to the very good
solubility of the complexes in dichloromethane and the formation of
extremely fine precipitates upon addition of diethyl ether, the isolated yields
are only 55-78%.
[
(OTf)2] complex.
13) Stang, P. J.; Cao, D. H.; Saito, S.; Arif, A. M. J. Am. Chem. Soc.
995, 117, 6273-6283. Here, an example is reported with the diaza-ligand
(
1
pyridazine which did not yield a molecular square but a complex with two
pyrazines coordinated to one metal ion. It was reasoned that steric interaction
between the phosphane ligands of adjacent metals prevented square
formation.
(16) All attempts to grow single crystals suitable for X-ray analysis failed.
Only very small crystals of poor quality could be obtained which were
unstable and immediately cracked upon manipulation or removal from the
mother liquor.
(
14) W u¨ rthner, F.; Sautter, A.; Thalacker, C. Angew. Chem., Int. Ed.
2
000, 39, 1243-1245.

5426 J. Am. Chem. Soc., Vol. 123, No. 23, 2001
Sautter et al.
Scheme 1. Self-Assembly by Metal-Ligand Coordination of Phenoxy-Substituted Diazadibenzoperylenes 1a,b and
Metal-Phosphane Triflates 2a,b and Triangle-Square Equilibria 3a-d/4a-d
between two different species. Interestingly, dissociation occurs
at elevated temperatures (T > 320 K) in DMSO as no more
sharp signals could be detected but only a broad resonance
around δ ) 20 ppm.
are formed. However, there are differences in the integral ratios
of the two sets of signals for the Pd(II) and the Pt(II) complexes.
The ratio in CD NO is ca. 2:3 for 1a-2a (Pd(II)) and ca. 1:1
for 1a-2b (Pt(II)). This implies that not one species with two
nonequivalent metal-phosphane corner and diazadibenzop-
erylene ligand units is present, but that two different species
must be in equilibrium. This point of view is further supported
by complexes that contain the sterically more demanding tert-
butyl-substituted ligand 1b which exhibit inverted integral ratios
3
2
The NMR spectra of all other complexes between diazali-
gands 1a,b and metal corners 2a,b exhibit the same features as
those of complex 1a-2a, that is, temperature dependence and
1
31
two sets of signals at high temperatures both in the H and P-
{
1
H} NMR spectra, indicating that the same kind of species
(ca. 7:3 (1b-2a) and 8:2 (1b-2b)) with respect to the
complexes containing ligand 1a. Obviously, here the other
species is preferred, pointing to a pronounced steric influence
of the tert-butyl groups on the equilibrium. Figure 2 shows parts
1
31
1
of the H and P{ H} NMR spectra of 1b-2b which exhibit
simpler signal patterns (because of the 4-substituted phenoxy
groups) and additionally two characteristic singlets for the tert-
1
17
butyl groups in the H NMR.
Mass Spectrometry. ESI-FTICR-MS allowed the unambigu-
ous assignment of the species present in the equilibria. This
Figure 2. 1H and 31P{ H} NMR spectra of triangle-square equilibrium
1
Figure 1. Variable-temperature H (left; aromatic part) and 31P{ H}
right) NMR spectra of triangle-square equilibrium 3a/4a recorded in
CD NO ; (a) 353 K; (b) 303 K; (c) 253 K.
1
1
3d/4d recorded in CD NO at 353 K; (a) aromatic protons; (b) tert-
3
2
(
butyl protons; (c) phosphorus nuclei.
3
2

J. Am. Chem. Soc., Vol. 123, No. 23, 2001 5427
Figure 4. Superposition of signals arising from the triply charged
triangle [3b-3OTf]3 and a simply charged fragment (one perylene
ligand and one platinum phosphane corner). Both species exhibit the
same m/z ratio, but are clearly resolved and assigned by the ESI-FTICR-
MS method.
+
Figure 3. (a) ESI-FTICR-MS (acetone solution) of triangle-square
equilibrium 3b/4b; the insets show the resolved signals of (b) the triply
3
+
charged square [4b-3OTf] and (c) the doubly charged triangle [3b-
2
+
2
OTf] .
technique was used because of both its capability of achieving
extremely high mass resolution and the possibility of measuring
with a very high mass accuracy.18 With this method signals
arising from the isotopic distribution of multiply charged large
molecules generated by ESI can be resolved. From the isotopic
pattern the charge state can be derived, thus allowing the
calculation of the molecular weight. However, only the more
stable platinum complexes 1a-2b and 1b-2b could be
analyzed. The overview spectrum of system 1a-2b (acetone
solution) displays three main signals centered at m/z 2254.416
with a charge state of z ) 2, that is, separation of peaks by 0.5
mass units, m/z 1987.392 (z ) 3) and m/z 1453.308 (z ) 3)
(
Figure 3). Two signals are magnified to show the isotopic
distribution (Figure 3b and c) with a resolution of up to 78 000
fwhh, full width at half peak height). As the ionization of the
Figure 5. ESI-FTICR-MS (acetone solution) of triangle-square
equilibrium 3d/4d; the inset shows the calculated (top) and the
(
complexes occurs simply by the loss of triflate anions, the
corresponding deconvolution of the three signals confirms that
only two oligomers are present in the equilibrium of a molecular
mass of 4807 and 6409. These values correspond exactly to a
experimental (bottom) isotopic distribution of the square species [4d-
+
3
OTf]3
.
doubly charged species of two corners and two diaza-ligands
“half-square”), (iii) a triply charged triangular species, (iv) a
(1a-2b)3 trimer (3b) and a (1a-2b)4 tetramer (4b). Evidence
(
for the cyclic nature of the species has been provided by the
NMR studies allowing us to assign the two species to be the
molecular triangle 3b and the molecular square 4b.
4
2
-fold charged square species or (v) higher n-fold charged (1a-
b)n species. All of these species render the same m/z ratio but
can be distinguished by their charge state and isotopic distribu-
tions. Therefore, only the high resolution of the ESI-FTICR
method allows the unambiguous determination of the charge
states for highly charged species and therefore identification
and assignment of the actual species. Figure 4 shows the
overlaying of a singly charged species (spacing of the mass
peaks by one mass unit) and a triply charged species (spacing
of 1/3) corresponding to (i) and (iii), and the signal pattern
matches well with the overlaying of the corresponding calculated
spectra.
In contrast to the signals for the trimeric and tetrameric
species at m/z 2254 and m/z 1987, the signal of the triply charged
triangle [3b-3OTf]3+ at m/z 1453 is superimposed by a singly
charged species with the same m/z ratio, that is, one perylene
ligand and one platinum phosphane corner (Figure 4). This
signal (m/z 1453) serves as a good example to address the
general difficulty of unambiguous characterization of large
supramolecular structures of high symmetry by mass spectrom-
etry. It could in principle originate from several species different
in their charge state z, (i) a singly charged fragment consisting
of a platinum corner coordinated by one diaza-ligand, (ii) a
The results for complex 1b-2b are in accordance to those
for 1a-2b, showing signals in the mass spectrum that originate
from the triangle 3d and the square 4d. However, under the
applied electrospray conditions (saturated acetone solution at
room temperature) the two dominating intense signals in the
(17) Assignment of the signal sets in the NMR spectra at elevated
temperatures is tentatively on the basis of the observed inverse integral
ratio of the two sets of signals in the tert-butylphenoxy systems 3c-d/
4
c-d compared to the phenoxy systems 3a-b/4a-b. The sterically
demanding tert-butyl groups will favor the molecular square in the
equilibrium, and therefore the dominating set of signals in 3c-d/4c-d is
assigned to the molecular squares. In the equilibria 3a-b/4a-b the
dominating set of signals is assigned to the molecular triangles, as is outlined
in Figure 1 and Figure 2. This assignment is also in accordance with ESI-
FTICR-MS signal intensities.
3+
spectrum arise exclusively from square 4d, the [4d-3OTf]
4
+
species at m/z 2287.038 and the [4d-4OTf] species at m/z
1677.795, whereas the signal intensities of the triangular species
2
+
3+
[3d-2OTf] (m/z 2591.271), [3d-3OTf] (m/z 1677.555), and
[
4+
3d-4OTf] (m/z 1220.912) are only about 5-10% of those
(
18) (a) Amster, J. J. Mass Spectrom. 1996, 31, 1325-1337. (b) Marshall,
A. G.; Hendrickson, C. L.; Jackson, G. S. Mass Spectrom. ReV. 1998, 17,
-35.
for the square signals (Figure 5). The inset in Figure 5 compares
the measured isotopic distribution (bottom) of the 3-fold charged
1

5428 J. Am. Chem. Soc., Vol. 123, No. 23, 2001
Sautter et al.
the outer side of the triangle, thus making the latter energetically
more favorable despite the less ideal geometry at the metal
2
1
centers (Figure 6d). Another feature that becomes important
when we consider 3c/4c and 3d/4d is the smaller inner volume
of the triangle which should lead to a sterically very critical
encounter of the three tert-butylphenoxy groups at the top and
bottom inner sides. Therefore, the reverse preference for the
triangle/square species for the two phenoxy substituents ob-
served by NMR and ESI-FTICR-MS is well-explained by these
molecular modeling studies.
To explain the dynamic behavior observed in NMR spec-
troscopy we account the following features, which we also
believe to be of importance in governing the equilibria. (i) The
diazadibenzoperylene backbone is twisted as revealed by an
X-ray crystal structure of a closely related ligand with a twisting
Figure 6. Side and top views of energy minimized structures of square
a (a,c) and triangle 3a (b,d) obtained by molecular dynamics
simulations and MM2 force field geometry optimizations. The steric
crowding of the dppp-phenyl rings and the phenoxy groups of the
diazadibenzoperylene ligands is most pronounced for square 4a as
visualized in (c).
4
14
angle of about 25° in the solid state. In solution fast flipping
averages this twisting, and the high flexibility of the phenoxy
substituents afford the good solubility properties of these
molecules. (ii) The somewhat bulky phosphane (dppp) ligands
wrap around the coordinating pyridine units of the diazadiben-
zoperylenes with their phenyl rings such as a tweezer that results
in favorable π-π contacts. This interaction has been observed
square (resolution up to 65 000 fwhh) with the calculated one
(
top). A precise match between these two patterns could be
1
3
observed for all signals.
in X-ray crystal structures, and it has been reported that in
solution this interaction leads to broad NMR resonances of both
pyridine and dppp protons that show a distinct temperature
The signal centered at m/z 1677 is again a superposition of
trimeric and tetrameric as well as a singly charged species all
of the same m/z ratio as found for the system 3b/4b. An
additional measurement at ultrahigh resolution revealed the
presence of these three species at the given m/z ratio. It is of
course not possible to quantify the equilibria by ESI-MS, but it
is noteworthy that the ratios are in fairly good agreement with
the situation in solution assigned by NMR spectroscopy.
Further evidence for the presence of dynamic triangle-square
equilibria was obtained from an ESI-FTICR-MS exchange
experiment. The platinum complexes 3b/4b and 3d/4d were
mixed in acetone in a 1:1 stoichiometry. In the resulting ESI-
FTICR mass spectra mixed species of triangles and squares
containing both diaza ligands 1a and 1b could be unambiguously
identified by their masses and charge states.19
4
f,10g,13
dependence.
(iii) There is an inner and outer region of
1
the macrocycles and one should expect splitting of the H NMR
resonances of the phenoxy and perylene core protons at low
temperatures when the structure is rigid and rotation is slow on
the NMR time scale caused by steric hindrance and other
intermolecular interactions. High temperatures lead to a high
conformational flexibility (flipping of the perylene backbone,
fast dppp dynamics and rotations of the perylene ligands) with
fast processes on the NMR time scale pretending a high
symmetry of the two species, resulting in simple sets of NMR
resonances. At room-temperature coalescence is reached, and
at lower temperatures there is a loss of conformational flexibility,
and rotations and other dynamic processes become slow on the
NMR time scale and therefore lead to abundant sharp reso-
nances, owing to the reduction of symmetry.
Molecular Modeling. To gain insight into the dynamic
behavior observed in the NMR spectra of the triangular and
square species, and to find arguments why the (expected)
molecular squares equilibrate with triangles, we applied mo-
lecular dynamics simulations that afforded plausible energy
Optical Properties. The metallacycles 3a-d/4a-d exhibit
strong absorption with maxima at 512 ( 2 nm and 480 ( 2 nm
(CH Cl ) which originate from the diazadibenzoperylene chro-
2
2
2
0
minimized structures of both triangles and squares. The
obtained energetically minimized structures are depicted in
Figure 6.
mophore. From Figure 7a we see a bathochromic shift of 15 (
1 nm for the S -S transition of this chromophore in the UV/
0
1
vis absorption spectra upon metal coordination. In addition, there
are strong changes in the intensities of the spectral fine structure
of the phenoxy-substituted systems 3a/4a and 3b/4b compared
to the diazadibenzoperylene ligand 1a. For the tert-butylphen-
oxy-substituted systems 3c/4c and 3d/4d the intensity changes
in the spectral fine structure are less pronounced (Figure 7a).22
It is obvious that there is considerable steric strain in both
cyclic assemblies rendering these metallacycles thermodynami-
4
cally less stable than other squares reported by us and others
(for stability studies, see below). The diazadibenzoperylene
ligands have a highly twisted and nonflexible aromatic back-
bone14 which is further rigidified when incorporated into the
macrocycles. Herein π-π contacts are given between two
phenyl groups of the dppp ligand and the π-surface of the
diazadibenzoperylene ligands which typically stabilize tetrameric
assemblies.4 However, as shown in the side views of both
assemblies, it is exactly this structural feature which interferes
considerably with the spatial demand of the phenoxy substituents
of our diazadibenzoperylene ligands, especially within the square
scaffold (Figure 6a,c). Significantly more space is available at
(21) Entropy should also favor the formation of triangles as they assemble
from fewer components with respect to squares.
(22) It is not yet clear to us what exactly causes these spectral changes.
A possible reason for the changes in the spectral fine structures might be
rigidification of the diazadibenzoperylene chromophores in the assemblies
as suggested by molecular modeling and the observed NMR dynamics.
Because the chromophores are more tightly assembled in the triangles than
in the squares, stronger changes in the intensity of vibronic transitions are
reasonable for the phenoxy-substituted systems 3a/4a and 3b/4b because
of a higher triangle-to-square ratio compared to the tert-butylphenoxy-
substituted systems 3c/4c and 3d/4d. Of course, other effects may also be
involved, such as excitonic coupling between the chromophores. For the
squares, excitonic coupling is expected to be negligible because of the
orthogonal arrangement of adjacent chromophores and the large distances
of ca. 1.5 nm between opposite chromophores. In the triangles the angle
between adjacent chromophores is significantly smaller than 90°, and the
,13
(
19) Selected experimental m/z values of mixed intact triangle and square
3
+
2+
species: [(1a1b22b3)-3OTf] m/z 1602.813; [(1a21b2b3)-2OTf] m/z
3
+
2
366.561; [(1a31b2b4)-3OTf] m/z 2062.155.
20) CAChe 3.2, Oxford Molecular 1999. MD-MM2 simulation at 300
K using the CAChe 3.2 program.
(

J. Am. Chem. Soc., Vol. 123, No. 23, 2001 5429
Figure 8. Dilution study monitored by the concentration dependence
of the UV/vis absorption coefficient at 478 nm (3a/4a (9s) and 3b/
4b (0- - -)) and normalized integrated emission corrected for the optical
density at the excitation wavelength (3a/4a (bs) and 3b/4b (O- - -)).
performed with the platinum complex 3b/4b (CH2Cl2), and
similar curves were obtained (Figure 8). However, the concen-
tration range, where complex dissociation occurs, is at lower
-
6
concentrations with a point of inflection at 1 × 10 M,
providing evidence for the higher stability of platinum coordina-
tion compared to that of palladium coordination.
Conclusions
Transition metal-directed self-assembly of substituted diaza-
dibenzoperylene bridging ligands 1a,b with Pd(II)- and Pt(II)-
phosphane corners 2a,b resulted in unexpected equilibria
between molecular triangles 3a-d and molecular squares 4a-
d. Owing to complex dynamics on the NMR time scale,
identification and characterization of the species could only be
Figure 7. (a) UV/vis absorption spectra of 1a (s), 1b (- - -), 3b/4b
-‚-‚-‚), and 3d/4d (‚‚‚). (b) Fluorescence spectra of 1a (s), 1b (- - -),
Cl (ꢀ
values were calculated with respect to the concentrations of the
chromophores 1a,b).
(
3
b/4b (-‚-‚-‚) and 3d/4d (‚‚‚). All spectra were recorded in CH
2
2
1
accomplished by a combination of variable-temperature H and
In contrast to the previously reported perylene bisimide-based
molecular squares which exhibited fluorescence quantum yields
31
1
P{ H} NMR spectroscopy and ESI-FTICR mass spectrometry.
4
f-g
By employing different metal ions, Pd(II) and Pt(II), and two
diazadibenzoperylene ligands that differ only in their phenoxy
substituents, energetic as well as steric effects that govern the
triangle-square equilibria could be investigated. Molecular
modeling helped to draw a reasonable picture of the steric
situation of the triangular and square structures. The absorption
properties of the diazadibenzoperylene chromophores are af-
fected considerably by metal coordination, and fluorescence is
strongly quenched. The higher stability of the platinum com-
plexes with respect to the palladium complexes could be proven
by dilution studies of the complexes which revealed dissociation
processes in the micro- to nanomolar range.
around 90%,
here the intense diazadibenzoperylene fluo-
rescence (ΦF ) 0.75) is strongly quenched upon metal
coordination. Weak emission of the complexes occurs with
maxima at 540 ( 1 nm and 576 ( 2 nm and a quantum yield
5
6
which is by a factor of 10 to 10 lower than for 1a,b (Figure
b).
To obtain further information on the stability of the metal-
7
lacycles we followed the process of coordination by UV/vis
and fluorescence dilution studies (Figure 8). Upon diluting a
-4
10
M concentrated solution of 3a/4a (CH2Cl2), the maximum
absorption band and the high-energy shoulder are gradually
shifted to shorter wavelengths, eventually resulting in the
absorption spectrum of the free ligand 1a. At the same time
the intensity of the high-energy shoulder at λ ) 478 nm
decreases. Figure 8 shows a logarithmic plot of the extinction
coefficient at this indicative wavelength versus concentration,
resulting in a sigmoid curve that exhibits a narrow concentration
Experimental Section
Materials and Methods. Solvents and reagents were purchased from
Merck unless otherwise stated and purified and dried according to
standard procedures.23 [Pd(dppp)][(OTf)
]‚2H
O (dppp ) 1,3-bis-
2
2
-
5
-6
(diphenylphosphino)propane; OTf ) trifluoromethanesulfonate) and [Pt-
range between 10 and 10 M, where complex formation-
dissociation processes take place. The same dilution study
monitored by fluorescence spectroscopy reveals a similar
signature if the normalized integrated emission (that is corrected
for the optical density at the excitation wavelength) is plotted
versus concentration. For both graphs the points of inflection
13
(
dppp)][(OTf)
2
2
]‚2H O
were synthesized according to the literature.
UV/vis spectra were recorded on a Perkin-Elmer Lambda 40P
spectrometer. All ꢀ values are given with respect to the concentration
of the diazadibenzoperylene chromophores. Corrected fluorescence
1
spectra were recorded on a Perkin-Elmer LS 50B fluorometer. H NMR
spectra were recorded on a Bruker AMX 500 spectrometer at 500 MHz,
and chemical shifts are reported relative to the signal of residual
protonated nitromethane δ 4.33 ppm. The 31P{ H} NMR spectra were
recorded at 202 MHz, and chemical shifts are reported relative to
-6
are located at the same concentrations of 3 × 10 M, indicating
that the same species in the equilibria are detected by both
spectroscopic techniques. Finally, the same experiment was
1
chromophore to chromophore center distances are considerably smaller;
hence, spectral changes due to excitonic coupling might be of importance.
(23) Perrin, D. D.; Armarego, W. L. F. Purification of Laboratory
Chemicals, 2nd ed;, Pergamon Press: Oxford, 1980.

5430 J. Am. Chem. Soc., Vol. 123, No. 23, 2001
Sautter et al.
external 85% aqueous H PO δ 0.0 ppm. ESI-FTICR-MS measurements
3
4
Fluorescence (CH
2
Cl
2
, 293 K) 539, 576 nm. Anal. Calcd for
were carried out with a passively shielded 4.7 T APEXII-ESI/MALDI-
FT-ICR-mass spectrometer from Bruker Daltonik (Bremen, Germany).
Saturated solutions of the complexes in acetone were prepared. For
mass calculation, data acquisition, processing, apodization, and decon-
volution the mass spectrometry software XMASS version 5.0.7 (Bruker
Daltonik) was used. The number of data points was 512 K for
acquisition and 1 M for processing within a mass range of 600-4000
Da. To increase the signal-to-noise-ratio 100 scans were accumulated.
For the external four-point-calibration an ES Tuning Mix from Hewlett-
Packard (Waldbronn, Germany) was used.
(C77H N O F P
54 2 10 6 2
PtS
2
2 n
‚2H O) (n‚1638.5) C, 56.45; H, 3.57; N, 1.71;
S, 3.91. Found C, 56.47; H, 3.53; N, 1.65; S, 3.84.
Metallacycles 3c/4c. 5,6,12,13-Tetra(4-tert-butylphenoxy)-2,9-di-
azadibenzo-[cd,lm]perylene 1b (24.4 mg, 0.027 mmol) and [Pd(dppp)]-
[
(OTf)
2
]‚2H
2
O 2a (22.6 mg, 0.027 mmol) were reacted and worked up
as described for 3a, 4a to yield 33 mg (70%) of an orange solid. Mp
>
H
(
H
1
310 °C. H NMR (500 MHz, CD
3
3
NO
2
, 353 K) 3c: δ 9.15 (s, 12H,
R
), 7.75 (m, 24H, Hdppp), 7.51 (d, J ) 8.4 Hz, 24H, HPhO), (7.34 (s
3
br), 36H, Hdppp)), 6.92 (d, J ) 8.4 Hz, 24H, HPhO), 3.27 (s(br), 12H,
dppp), 2.39 (s(br), 6H, Hdppp), 1.49 (s, 108H, HtBu); 4c: δ 9.43 (s, 16H,
14
The synthesis of 1a has been recently described in detail. The same
synthetic protocol was used for the synthesis of the tert-butyl-substituted
analogue 1b (refer to Supporting Information).
3
H
R
), 7.63 (m, 32H, Hdppp), 7.38 (d, J ) 8.4 Hz, 32H, HPhO), (7.13 (s,
3
4
H
H, 16H, Hperylene)), 7.04 (m, 48H, Hdppp), 6.86 (d, J ) 8.4 Hz, 32H,
PhO), 3.22 (s(br), 16H, Hdppp), 2.39 (s(br), 8H, Hdppp), 1.37 (s, 36H,
Metallacycles 3a/4a. 5,6,12,13-Tetraphenoxy-2,9-diazadibenzo-
31
1
H
tBu). P{ H} NMR (202 MHz, CD
3
NO
Cl
dm cm )] 513 (67 300), 481 (43 800), 336 (23 400), 294 (70 200),
86 (61 500). Fluorescence (CH Cl , 293 K) 541, 575 nm. Anal. Calcd
for (C93 PdS ‚H O) (n‚1756.2) C, 63.61; H, 5.05; N, 1.60;
2
, 353 K, 85% H
3 4
PO ) 3c: δ
[
2
cd,lm]perylene 1a (50.0 mg, 0.072 mmol) and [Pd(dppp)][(OTf)
O 2a (61.3 mg, 0.072 mmol) were stirred at room temperature under
argon in CH Cl (10 mL) for 24 h. After filtration the solution was
2
]‚
-1
11.19 (s); 4c: δ 10.16 (s). UV-vis (CH
2
2
, 293 K) λmax (ꢀ) [nm (mol
H
2
3
-1
2
2
2
2
2
concentrated to ca. 1 mL, and an orange solid was precipitated upon
addition of diethyl ether. The product was separated by centrifugation,
washed twice with diethyl ether, and dried in vacuo to yield 86 mg
86 2 10 6
H N O F P
2
2
2
n
S, 3.65. Found C, 63.71; H, 5.11; N, 1.56; S, 3.37.
Metallacycles 3d/4d. 5,6,12,13-Tetra(4-tert-butylphenoxy)-2,9-di-
azadibenzo-[cd,lm]perylene 1b (36.9 mg, 0.04 mmol) and [Pt(dppp)]-
1
(
3
78%). Mp 304-306 °C dec. H NMR (500 MHz, CD
3
NO
2
, 353 K)
3
a: δ 9.11 (s, 12H, H
R
), 7.73 (m, 24H, Hdppp), 7.41 (t, 24H, J ) 7.5
Hz, HPhO), 7.37 (s, 12H), 7.22 (m, 12H, HPhO), 7.16 (m, 24H, Hdppp),
.09 (m, 12H, Hdppp), 6.92 (d, 24H, J ) 7.5 Hz, HPhO), 3.26 (s(br),
2H, Hdppp), 2.43 (s(br), 6H, Hdppp); 4a: δ 9.38 (s, 16H, H ), 7.62 (m,
2H, Hdppp), 7.38 (s, 16H), 7.34 (t, 32H, HPhO), 7.29 (m, 16H, Hdppp),
.27 (m, 16H, HPhO), 7.06 (m, 32H, Hdppp), 6.94 (d, 32H, HPhO), 3.21
s(br), 16H, Hdppp), 2.43 (s(br), 8H, Hdppp). P{ H} NMR (202 MHz,
CD NO , 353 K, 85% H PO ) 3a: δ 11.31 (s); 4a: δ 10.42 (s). UV-
vis (CH Cl , 293 K) λmax (ꢀ) [nm (mol dm cm )] 510 (68 100),
78 (49 100), 333 (30 900), 293 (83 200); Fluorescence (CH Cl , 293
K) 541, 575 nm. Anal. Calcd for (C77 PdS ‚H O) (n‚
531.8) C, 60.38; H, 3.68; N, 1.83; S 4.19. Found: C, 60.29; H, 3.89;
N, 1.68; S, 4.00.
Metallacycles 3b/4b. 5,6,12,13-Tetraphenoxy-2,9-diazadibenzo-
cd,lm]perylene 1a (50.0 mg, 0.072 mmol) and [Pt(dppp)][(OTf) ]‚2H
b (67.6 mg, 0.072 mmol) were reacted and worked up as described
for 3a, 4a to yield 75 mg (64%) of an orange solid. Mp >310 °C. H
NMR (500 MHz, CD NO , 353 K) 3b: δ 9.09 (s, 12H, H ), 7.73 (m,
4H, Hdppp), 7.43 (t, 24H, HPhO), 7.40 (s, 12H), 7.32-7.05 (m, obscured
and overlap with signals of 4b, 12H + 24H + 12H, HPhO + Hdppp
[
(OTf)
2
2
]‚2H O 2b (37.7 mg, 0.04 mmol) were reacted and worked up
as described for 3a, 4a to yield 41 mg (55%) of an orange solid. Mp
>
H
3
3
4
3
7
1
3
7
1
3 2
310 °C. H NMR (500 MHz, CD NO , 353 K) 3d: δ 9.11 (s, 12H,
R
3
R
), 7.76 (m, 24H, Hdppp), 7.52 (d, J ) 8.4 Hz, 24H, HPhO), (7.19 (m,
3
6H, Hdppp)), (7.16 (s, 12H, Hperylene)), 6.92 (d, J ) 8.4 Hz, 24H, HPhO),
.33 (s(br), 12H, Hdppp), 2.39 (s(br), 6H, Hdppp), 1.50 (s, 108H, HtBu);
3
1
1
(
3
d: δ 9.40 (s, 16H, H
R
), 7.64 (m, 32H, Hdppp), 7.39 (d, J ) 8.4 Hz,
3
2
3
4
3
2H, HPhO), 7.36 (s, 4H, 16H, Hperylene), 7.06 (m, 48H, Hdppp), 6.87 (d,
-
1
3
-1
2
2
3
J ) 8.4 Hz, 32H, HPhO), 3.28 (s(br), 16H, Hdppp), 2.39 (s(br), 8H, Hdppp),
4
2
2
31
1
1
H
λ
3
.38 (s, 36H, HtBu). P{ H} NMR (202 MHz, CD
PO ) 3d: δ -11.31 (s); 4d: δ -12.07 (s). UV-vis (CH
max (ꢀ) [nm (mol dm cm )] 513 (68 500), 481 (45 700), 334 (29
00), 294 (78 400). Fluorescence (CH Cl , K) 540, 578 nm. Anal. Calcd
PtS ‚2H O) (n‚1862.9) C, 59.96; H, 4.87; N, 1.50;
3
NO
2
, 353 K, 85%
H
54
N
2
O
10
F P
6 2
2
2
n
3
4
2
Cl , 293 K)
2
1
-1
3
-1
2
2
for (C93
86 2 10 6
H N O F P
2
2
2
n
[
2
2
2
O
S, 3.44. Found C, 60.25; H, 4.88; N, 1.50; S, 3.50.
1
Acknowledgment. F. W. is indebted to the Fonds der
Chemischen Industrie and BMBF for a Liebig Grant and to DFG
for a Habilitation Grant. We gratefully acknowledge the Ulmer
Universit a¨ tsgesellschaft for a startup grant for this project, BASF
AG and Degussa-H u¨ ls AG for the donation of chemicals.
3
2
R
2
+
H
H
H
dppp), 6.93 (d, 24H, HPhO), 3.32 (s(br), 12H, Hdppp), 2.40 (s(br), 6H,
dppp); 4b: δ 9.35 (s, 16H, H ), 7.63 (m, 32H, Hdppp), 7.37 (t, 32H,
PhO), 7.35 (s, 16H), 7.32-7.05 (m, obscured and overlap with signals
R
of 3b, 16H + 32H + 16H, HPhO + Hdppp + Hdppp), 6.95 (d, 32H, HPhO),
Supporting Information Available: Preparation and physi-
cal properties of 1b (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
3
1
1
3
.27 (s(br), 16H, Hdppp), 2.40 (s(br), 8H, Hdppp). P{ H} NMR (202
MHz, CD NO , 353 K, 85% H PO ) 3b: δ -11.14 (s); 4b: δ -11.65
s). UV-vis (CH Cl , 293 K) λmax (ꢀ) [nm (mol dm cm )] 510 (69
00), 478 (49 300), 335 (25 200), 324 (24 900), 294 (79 900).
3
2
3
4
-
1
3
-1
(
9
2
2
JA004360Y