Shape-persistant macrocycles with terpyridine units: Synthesis, characterization, and structure in the crystal

Article abstract of DOI:10.1021/ja034029p

The synthesis of a variety of shape-persistent macrocycles with either one (1a-d, 2) or two (opposing) terpyridine units (3, 4, 5a-c) and inner diameters of up to 2 nm is described. The sequences are mainly based on transition metal cross-coupling reactio

Full text of DOI:10.1021/ja034029p

Published on Web 05/17/2003
Shape-Persistant Macrocycles with Terpyridine Units:
Synthesis, Characterization, and Structure in the Crystal
Christian Grave,† Dieter Lentz,*,† Andreas Scha¨fer,† Paolo Samor`ı,‡,§
Ju¨rgen P. Rabe,‡ Peter Franke,† and A. Dieter Schlu¨ter*,†
Institut fu¨r Chemie, Freie UniVersita¨t Berlin, Takustrasse 3, D-14195 Berlin, Germany, and
Institut fu¨r Physik, Humboldt UniVersita¨t zu Berlin, Newtonstrasse 15
D-12489 Berlin, Germany
Received January 3, 2003; E-mail: adschlue@chemie.fu-berlin.de
Abstract: The synthesis of a variety of shape-persistent macrocycles with either one (1a-d, 2) or two
(opposing) terpyridine units (3, 4, 5a-c) and inner diameters of up to 2 nm is described. The sequences
are mainly based on transition metal cross-coupling reactions and, whenever appropriate, compared with
one another regarding their respective efficiency. Typical overall yields and amounts prepared range from
8% (4) to 27% (3) and 25 mg (1a) to 290 mg (1b), respectively. For solubility and processing of the targeted
cycles, all precursors have already been decorated with flexible side chains (hexyloxy or hexyloxymethyl).
The cycles’ characterization is based on MALDI-TOF mass spectrometry, 2D NMR spectroscopy, and/or
low-temperature single-crystal X-ray diffraction. Their packing in the crystal is discussed in terms of both
number and length of side chains. Cycle 1d was physisorbed into an ordered structure at the solution-
HOPG interface and investigated by scanning tunneling microscopy (STM).
flexible macrocycles such as cycloalkanes3 their cyclic back-
bones’ conformational rigidity creates a defined interior that is
Introduction
In the last decade there was a great deal of research on shape-
persistent macrocycles on the nanometer scale.1,2 In contrast to
well separated from an exterior. This rigidity together with a
noncollapsable and, thus, usable interior are the main factors
that render these macrocycles interesting candidates for various
supramolecular aspects/applications including lyotropic and
thermotropic behavior,4 organization into and transport through
porous molecular crystals,1a-c,5 and pattern formation at
interfaces.1d,6 Besides representatives such as cyclodextrins,1e,7
cyclopeptides,1f,8 metalacycles,1g-i,9 etc.,2a,b macrocycles, espe-
cially with backbones consisting of only sp- or sp2-hybridized
atoms, have attracted considerable interest.1a,c,d,j-w,2c,10 Moore’s
trailblazing phenylene acetylene macrocycles,11 as well as
† Freie Universita¨t Berlin.
‡ Humboldt Universita¨t zu Berlin.
§ New address: Istituto per la Sintesi Organica e la Fotoreattivita`, CNR
Bologna, Via Gobetti 101, 40129 Bologna, Italy.
(1) For a selection of the most recent examples, see: (a) Henze, O.; Lentz,
D.; Scha¨fer, A.; Franke, P.; Schlu¨ter, A. D. Chem. Eur. J. 2002, 357-365.
(b) Werz, T. B.; Staeb, T. H.; Benisch, C.; Rausch, B. J.; Rominger, F.;
Gleiter, R. Org. Lett. 2002, 4, 339-342. (c) Campbell, K.; Kuehl, C. J.;
Ferguson, M. J.; Stang, P. J.; Tykwinski, R. R. J. Am. Chem. Soc. 2002,
124, 7266-7267. (d) Ba¨uerle, P.; Mena-Osteritz, E.; Fuhrmann, G.; Kaiser,
A.; Ammann, M. Polym. Mater. Sci. Eng. 2002, 86, 34. (e) Wakao, M.;
Fukase, K.; Kusumoto, S. J. Org. Chem. 2002, 67, 8182-8190. (f)
Sanderson, J. M.; Yazdani, S. Chem. Commun. 2002, 1154-1155. (g)
Newkome, G. R.; Cho, T. J.; Moorefield, C. N.; Cush, R.; Russo, P. S.;
Godinez, L. A.; Saunders: M. J.; Mohapatra, P. Chem. Eur. J. 2002, 8,
2946-2954. (h) Schweiger, M.; Seidel, S. R.; Arif, A. M.; Stang, P. J.
Inorg. Chem. 2002, 41, 2556-2559. (i) Takahashi, R.; Kobuke, Y. J. Am.
Chem. Soc. 2003, 125, 2372-2373. (j) Iyoda, M.; Nakao, K.; Kondo, T.;
Kuwatani, Y.; Yoshida, M.; Matsuyama, H.; Fukami, K.; Nagase, S.
Tetrahedron Lett. 2001, 42, 6869-6872. (k) Campbell, K.; McDonald, R.;
Ferguson, M. J.; Tykwinski, R. R. Organometallics 2003, 22, 1353-1355.
(l) Schmittel, M.; Ganz, A.; Fenske, D. Org. Lett. 2002, 4, 2289-2292.
(m) Baxter, P. N. W., Chem. Eur. J. 2002, 8, 5250-5264. (n) Yu, L.;
Lindsey, J. S. J. Org. Chem. 2001, 66, 7402-7419. (o) Rucareanu, S.;
Mongin, O.; Schuwey, A.; Hoyler, N.; Gossauer, A.; Amrein, W.; Hediger,
H.-U. J. Org. Chem. 2001, 66, 4973-4988. (p) Schafer, L. L.; Nitschke,
J. R.; Mao, S. S. H.; Liu, F.-Q.; Harder, G.; Haufe, M.; Tilley, T. D. Chem.
Eur. J. 2002, 8, 74-83. (q) Tobe, Y.; Utsumi, N.; Kawabata, K.; Nagano,
A.; Adachi, K.; Araki, S.; Sonoda, M.; Hirose, K.; Naemura, K. J. Am.
Chem. Soc. 2002, 124, 5350-5364. (r) Nielsen, M. B.; Diederich, F. Chem.
Rec. 2002, 2, 189-198. (s) Fischer, M.; Ho¨ger, S. Eur. J. Org. Chem.
2003, 441-446. (t) Liu, C.-H.; Tour, J. J. Org. Chem. 2002, 67, 7761-
7768. (u) An, D. L.; Nakano, T.; Orita, A.; Otera, J. Angew. Chem., Int.
Ed. 2002, 41, 171-173. (v) Srinivasan, M.; Sankararaman, S.; Hopf, H.;
Varghese, B. Eur. J. Org. Chem. 2003, 660-665. (w) Shen, X.; Ho, D.
M.; Pascal, R. A., Jr. Org. Lett. 2003, 5, 369-371. (x) Chiu, S.-H.; Pease,
A. R.; Stoddart, J. F.; White, A. J. P.; Williams, D. J. Angew. Chem., Int.
Ed. 2002, 41, 270-274. (y) de Meijere, A.; Kozhushkov, S. I. Chem. Eur.
J. 2002, 8, 3195-3202. (z) Higuchi, M.; Kanazawa, H.; Yamamoto, K.
Org. Lett. 2003, 5, 345-347.
(2) For recent reviews, see: (a) de Meijere, A.; Kozhushkov, S. I. Top. Curr.
Chem. 1999, 201, 1-42. (b) Ho¨ger, S. J. Polym. Sci., Part A: Polym. Chem.
1999, 37, 2685-2698. (c) Grave, C.; Schlu¨ter, A. D. Eur. J. Org. Chem.
2002, 3075-3098.
(3) See, for example: Keul, H.; Ho¨cker, H. Cycloalkanes and Related
Oligomers and Polymers. In Large Ring Molecules; Semlyen, J. A., Ed.; J.
Wiley & Sons: Chichester, 1996; Chapter 10, pp 375-406.
(4) See, for example: (a) Kimura, M.; Wada, K.; Ohta, K.; Hanabusa, K.;
Shirai, H.; Kobayashi, N. J. Am. Chem. Soc. 2001, 123, 2438-2439. (b)
Ho¨ger, S.; Enkelmann, V.; Bonrad, K.; Tschierske, C. Angew. Chem., Int.
Ed. 2000, 39, 2268-2270. (c) Mindyuk, O. Y.; Stetzer, M. R.; Heiney, P.
A.; Nelson, J. C.; Moore, J. S. AdV. Mater. 1998, 10, 1363-1366.
(5) See, for example: (a) Mu¨ller, P.; Uso´n, I.; Hensel, V.; Schlu¨ter, A. D.;
Sheldrick, G. M. HelV. Chim. Acta 2001, 84, 778-785. (b) Venkataraman,
D.; Lee, S.; Zhang, J.; Moore, J. S. Nature 1994, 371, 591-593.
(6) See, for example: (a) Ho¨ger, S.; Bonrad, K.; Mourran, A.; Beginn, U.;
Mo¨ller, M. J. Am. Chem. Soc. 2001, 123, 5651-5659. (b) Kro¨mer, J.; Rios-
Carreras, I.; Fuhrmann, G.; Musch, C.; Wunderlin, M.; Debaerdemaeker,
T.; Mena-Osteritz, E.; Ba¨uerle, P. Angew. Chem., Int. Ed. 2000, 112, 3623-
3628. (c) Mindyuk, O. Y.; Stetzer, M. R.; Gidalevitz, D.; Heiney, P. A.;
Nelson, J. C.; Moore, J. S. Langmuir 1999, 15, 6897-6900.
(7) Harada, A. Cyclodextrins. In Large Ring Molecules; Semlyen, J. A., Ed.;
J. Wiley & Sons: Chichester, 1996; Chapter 11, pp 407-432.
(8) Scheraga, H. A. Cyclic Peptides and Loops in Proteins. In Large Ring
Molecules; Semlyen, J. A., Ed.; J. Wiley & Sons: Chichester, 1996; Chapter
3, pp 99-112.
(9) Leininger, S.; Olenyuk, B.; Stang, P. J. Chem. ReV. 2000, 100, 853-908.
9
10.1021/ja034029p CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 6907-6918
6907

A R T I C L E S
Grave et al.
phenylene12 and phenylene butadiynylene ones,13 are the basic
structural types here. They have been described in sizes of up
to approximately 3 nm (diameter) and with different side-chain
patterns.2
of their supramolecular and materials chemistry related proper-
ties, the optimum synthesis strategy and the availability are
discussed in some detail. Solubility is an essential feature for
characterization and processing of conformationally rigid com-
pounds. For some of these cycles, the solubilities in chloroform
were quantified, therefore, and qualitatively correlated with
structural factors such as side-chain pattern and number of
terpyridine units per cycle. Furthermore, first insights into these
cycles’ packing behavior both in the 3D of a single crystal and
in the 2D of self-assembled monolayers adsorbed on highly
oriented pyrrolytic graphite (HOPG) are provided by X-ray
diffraction and scanning tunneling microscopy investigations,
respectively.
There has recently been a tendency to extend the field of
shape-persistent and nanosized macrocycles to those with
exo- or endocyclically oriented heteroatoms as anchor
groups.1a,c,d,j-p,2c The incorporation of anchor groups such as
sulfur in thiophene1d,j,6b or the nitrogens in pyridine,14
bipyridines,1a,15 and terpyridines16 widens the cycles’ applicabil-
ity considerably. Macrocycles with exocyclically oriented anchor
sites may be used, for example, as rigid scaffolds to hold metals
at a defined distance1c,17 or for metal-mediated linear (1D), areal
(2D), or even spatial (3D) constructs in which the cycles serve
as Tinkertoy building blocks and the metals as connectors.1c,l
Endocyclic anchor groups, in turn, should allow one to either
specifically bind guests inside a cycle’s cavity1m-o,s,18 or induce
incorporation of, for example, metal clusters. In combination
with the cycles’ presumed inherent ability to form superstruc-
tures with channels by aggregation, these thoughts could lead
to a new approach to atomically thin metal wires.19
Results and Discussion
Synthesis. The target structures reported here are hexagonal
and carry either 46 (1a-d, 3), 50 (4), or 58 (2, 5a-c) atoms in
the core, thus being the largest shape-persistent pyridine-
containing macrocycles described so far (Figure 1).14,16 To allow
a stepwise approach to metal complexes, cycles with either one
or two terpyridine units were prepared. In the former case the
terphenyl unit serves as a geometrically similar placeholder for
terpyridyl. The different side-chain patterns ought to mediate
solubility and influence aggregation behavior. The tetrahydro-
pyranyl (THP) groups in 1c and 5b serve as protecting groups
for OH to allow easy modifications on an already existing
macrocycle without having to carry the modification through
parts of the sequence. Except 4, which was cyclized via
oxidative alkynyl-alkynyl homocoupling,23 all other structures
were cyclized via Sonogashira (alkynyl-aryl) cross-coupling.24
An aspect to be considered in macrocyclic chemistry is the
number of precursor compounds that are used in the final
cyclization step. The more of them employed, the lower the
cyclization yield will normally be, but their syntheses’ effort
will also decrease. In contrast, the fewer precursors used, the
higher their syntheses’ effort will be and cyclization efficiency
will normally increase. For each individual cycle the efficiency
maximum of these two opposing effects has to be individually
assessed. From a comprehensive evaluation of the literature,
by and large, one can conclude that regarding cycles containing
heteroaromatic units (like terpyridenes) a cyclization strategy
involving two cyclization precursors seems to be the optimum.2c
Equipped with AA and BB functionalities, respectively, they
lead to cycles in yields of 20-25% irrespective of whether the
two “halves” have similar sizes or not. Considerations to
generate heteroaromatic cycles from just one precursor com-
pound (which obviously must be of the AB-type) are syntheti-
cally unattractive, because, for the cases considered here, this
would require appropriately unsymmetrically functionalized
terpyridines. They, however, do not exist yet.25,26 The target
2,2′-Bipyridine and 2,2′:6′,2′′-terpyridine are widely used as
ligands for metal complexation;20 these complexes show
interesting photophysical and electrochemical properties.21 Their
incorporation into macrocyclic backbones may suit especially
well the purposes described above. We have synthesized,
therefore, a number of 5,5′-substituted bi- and 5,5′′-substituted
terpyridine building blocks22 for the construction of shape-
persistent macrocycles, a few of which have already been
obtained.1a,15,16 The present contribution describes the synthesis
of a whole set of phenylacetylene macrocycles with either one
or two terpyridine units. They differ in size and side-chain
pattern. To provide a solid basis for a systematic exploration
(10) For a highlight on “polyunsaturated cyclophanes”, see: Bodwell, G. J.;
Satou, T. Angew. Chem., Int. Ed. 2002, 41, 4003-4007.
(11) Zhang, J.; Pesak, D. J.; Ludwick, J. L.; Moore, J. S. J. Am. Chem. Soc.
1994, 116, 4227-4239.
(12) (a) Hensel, V.; Schlu¨ter, A. D. Chem. Eur. J. 1999, 5, 421-429. (b)
Schlu¨ter, A. D.; Hensel, V.; Liess, P.; Lu¨tzow, K. NATO ASI Series, Series
C: Mathematical and Physical Sciences (1999), 499 (Modular Chemistry),
241-250. (c) Hensel, V.; Lu¨tzow, K.; Jacob, J.; Gessler, K.; Saenger, W.;
Schlu¨ter, A. D. Angew. Chem., Int. Ed. Engl. 1997, 36, 2654-2656.
(13) (a) Tobe, Y.; Utsumi, N.; Nagano, A.; Naemura, K. Angew. Chem., Int.
Ed. 1998, 37, 1285-1287. (b) Ho¨ger, S.; Meckenstock, A.-D.; Pellen, H.
J. Org. Chem. 1997, 62, 4556-4557. (c) de Meijere, A.; Kozhushkov, S.;
Haumann, T., Boese, R.; Puls, C.; Cooney, M. J.; Scott, L. T. Chem. Eur.
J. 1995, 1, 124. (d) Boldi, A. M.; Diederich, F. Angew. Chem., Int. Ed.
Engl. 1994, 33, 486. (e) Anderson, H. L.; Sanders, J. K. M. Angew. Chem.,
Int. Ed. Engl. 1990, 29, 1400-1403.
(14) Tobe, Y.; Nagano, A.; Kawabata, K.; Sonoda, M.; Naemura, K. Org. Lett.
2000, 2, 3265-3268.
(15) Henze, O.; Lentz, D.; Schlu¨ter, A. D. Chem. Eur. J. 2000, 2362-2367.
(16) Lehmann, U.; Schlu¨ter, A. D. Eur. J. Org. Chem. 2000, 3483-3487.
(17) Campbell, K.; McDonald, R.; Tykwinski, R. R. J. Org. Chem. 2002, 67,
1133-1140.
(18) Nakash, M.; Sanders, J. K. M. J. Chem. Soc., Perkin Trans. 2 2001, 2189-
2194.
(19) See, for example: (a) Hong, B. H.; Bae, S. C.; Lee, C.-W.; Jeong, S.;
Kim, K. S. Science 2001, 294, 348-351. (b) Gudiksen, M. S.; Lauhorn,
L. J.; Wang, J.; Smith, D. C.; Lieber, C. M. Nature 2002, 415, 617-620.
(20) For related reviews see: (a) Schubert, U. S.; Eschbaumer, C. Angew. Chem.,
Int. Ed. 2002, 41, 2892-2926. (b) Kaes, C.; Katz, A.; Hosseini, M. W.
Chem. ReV. 2000, 100, 3553-3590. (c) Cargill Thompson, A. M. W. Coord.
Chem. ReV. 1997, 160, 1-52. (d) Constable, E. C. AdV. Inorg. Chem.
Radiochem. 1986, 30, 69-121.
(23) For oxidative alkynyl-alkynyl homocoupling reactions catalyzed by Pd2+
,
see, for example: (a) Liu, Q.; Burton, D. J. Tetrahedron Lett. 1997, 38,
4371-4374. (b) Takano, S.; Sugihara, T.; Ogasawara, K. Synlett 1990,
453-454. (c) Rossi, R.; Carpita, A.; Bigelli, C. Tetrahedron Lett. 1985,
26, 523-526.
(24) (a) Sonogashira, K. Coupling Reactions Between sp2 and sp Carbon Centers
in ComprehensiVe Organic Synthesis; Trost, B. M., Ed.; Pergamon: New
York, 1991; Vol 3, Chapter 2.4. (b) Sonogashira, K.; Thoda, Y.; Hagihara,
N. Tetrahedron Lett. 1975, 16, 4467-4470.
(21) Sauvage, J.-P.; Collin, J.-P.; Chambron, J.-C.; Guillerez, S.; Coudret, C.;
Balzani, V.; Barigelletti, F.; DeCola, L.; Flamigni, L. Chem. ReV. 1994,
94, 993-1019.
(22) (a) Manickam, G.; Schlu¨ter, A. D. Synthesis 2000, 3, 442-446. (b)
Manickam, G.; Schlu¨ter, A. D. Eur. J. Org. Chem. 2000, 3475-3481. (c)
Henze, O.; Lehmann, U.; Schlu¨ter, A. D. Synthesis 1999, 4, 683-687. (d)
Lehmann, U.; Henze, O.; Schlu¨ter, A. D. Chem. Eur. J. 1999, 5, 854-
859.
(25) For a recent review about unsymmetrical bi- and terpyridines, see: Chelucci,
G.; Thummel, R. P. Chem. ReV. 2002, 102, 3129-3170.
(26) For a procedure to generate an unsymmetrically functionalized phenan-
throline building block for a similar purpose, see: Liu, S.-X.; Michel, C.;
Schmittel, M. Org. Lett. 2000, 2, 3959-3962.
9
6908 J. AM. CHEM. SOC. VOL. 125, NO. 23, 2003

Shape-Persistant Macrocycles with Terpyridine
A R T I C L E S
Scheme 1. Stepwise Synthesis of Terpyridine Unit 14aa
a Reagents and conditions: (a) 1. TMS-acetylene, Pd(PPh3)4/CuI, TEA/
toluene; 2. MeOH, (b) Pd(PPh3)4/CuI, TEA/toluene, and (c) Pd(PPh3)4,
toluene.
Due to the lower symmetry of macrocycles with heteroaro-
matic units, repetitive strategies toward the precursor molecules
that otherwise proved to be very useful in macrocycle synthe-
sis11,12 are not feasible here. The cycles’ smallest repeat unit
(if there are any) has already the size of a cyclization precursor.
The alternative building block strategy for a modular approach
to two half-ring precursors has been successfully employed.1a,15,16
The synthetic sequences to the cycles 1-5 are divided into
Schemes 1-7. Scheme 1 describes a stepwise construction of
the symmetrical cycle precursor 14a with an early generation
of the terpyridine unit (in 11), Scheme 2 a more favorable
synthesis of 14 and some derivatives with a late generation of
the terpyridine unit (in 14), Scheme 3 the assembly of cycles
1a-d from the differently sized “half-rings” 20a-c as AA-
type and 21a,b as BB-type components, respectively, Scheme
4 the assembly of the large precursor 26b, again with a late
generation of the terpyridine unit (in 26a), and its cyclization
with the BB-type component 21a to the cycle 2, Scheme 5 the
assembly of cycle 3 from the two similarly sized half-rings 30
as AA-type and 33b as BB-type component with an early
generation of the terpyridine, Scheme 6 the analogous assembly
of the larger cycles 5a-c from precursors 33b,c and 35a,b, and,
finally, Scheme 7 the oxidative cyclization of the half-ring
precursor 33b to cycle 4. In the following the different synthesis
parts will be briefly commented on.
Figure 1. Synthesized macrocycles with one (1, 2) or two terpyridine units
(3-5).
structures presented in Figure 1 were, therefore, accessed by
cyclization of AA- and BB-type “half-ring” precursors with
either a terphenyl or terpyridine unit.
9
J. AM. CHEM. SOC. VOL. 125, NO. 23, 2003 6909

A R T I C L E S
Grave et al.
Scheme 2. Synthesis of Half-Rings 20a-c with the Terpyridine
Scheme 3. Ring-Closure of 46-Membered Cycles 1a-d with One
Assembly in the Last Stepa
Terpyridine Unita
a Reagents and conditions: Pd(PPh3)4/CuI, TEA/toluene, high dilution.
Scheme 1. The iodobromopyridine 6 was coupled with
TMS-acetylene under Sonogashira conditions. One-pot depro-
tection and recrystallization yielded the bromo,ethynyl-func-
tionalized pyridine 7. Its reaction with the bromoiodobenzene
8 under Sonogashira conditions gave product 9 by selective
coupling at the iodo site of 8, thus leaving the two bromo
functions of both coupling partners unaffected. In analogy to a
known terpyridine synthesis22d the pyridyl bromo function of 9
was then reacted under Stille cross-coupling conditions with
bis-stannylated pyridine 10 to selectively give terpyridine 11.
Some bipyridine homocoupling product was removed during
isolation. The bis-stannylated pyridine 10 is a key compound
in the present work.27 All terpyridines reported here were
synthesized by coupling it with an ethynyl-substituted bromo-
functionalized pyridine (9, 19a-c, 25, 27, 31a-c). The yields
of the respective products (11, 14a-c, 26a, 28a, 32a-c) vary
between 31 and 70%, with a tendency to decreasing values for
increasing substituent’s size of the bromo pyridine. The reported
yields of terpyridines prepared by the same strategy depend on
the substitution of the pyridines used and at 14-55% are
somewhat lower.16,22d,28 Sonogashira cross-coupling of 11 with
ethynyl compound 12a gave a mixture of the desired product
14a, starting material 11, monocoupled 13, and the homocoupled
15, with the separation of the former three being extremely
tedious and associated with an enormous loss of material,
although after several columns pure 14a could finally be
obtained. Different conditions (DIPEA/toluene or propylamine)
did not improve the results. Thus, the relatively early generation
of the terpyridine unit in the sequence was considered disad-
vantageous as long as the separation problem could not be
solved. Possible ways to solve the problem would be either the
application of a larger excess of 12a or the transformation of
bromo into the much more reactive iodo functionality. The
former was unattractive because 12’s synthesis involves six steps
and its recovery was expected to be too complicated due to
homocoupling side reactions. The latter option was also not
seriously considered because of the known incompatibility of
the terpyridine unit with many iodination reagents.29 We,
therefore, decided to move the generation of the terpyridine unit
(27) Yamamoto, Y.; Yanagi, A. Chem. Pharm. Bull. 1982, 30, 1731-1737.
(28) For a recent, improved synthetic method via a central 2,6-diiodopyridine
building block, see: Colasson, B. X.; Dietrich-Buchecker, C.; Sauvage,
J.-P. Synlett 2002, 271-272.
(29) See, for example: (a) Henze, O. Ph.D. Thesis, FU Berlin, Germany, 2000,
http://www.diss.fu-berlin.de/2000/47/. (b) Lehmann, U. Ph.D. Thesis, FU
Berlin, Germany, 1999.
a Reagents and conditions: (a) Pd(PPh3)4/CuI, TEA; (b) NaOH, methanol/
CH2Cl2; (c) Pd(PPh3)4, toluene; and (d) Bu4NF, THF.
9
6910 J. AM. CHEM. SOC. VOL. 125, NO. 23, 2003

Shape-Persistant Macrocycles with Terpyridine
A R T I C L E S
Scheme 4. Synthesis of 58-Membered Cycle 2 with One
Scheme 5. Synthesis of 46-Membered Cycle 3 with Two
Terpyridine Unita
Terpyridine Unitsa
a Reagents and conditions: (a) Pd(PPh3)4/CuI, TMS-acetylene, TEA/
toluene; (b) Pd(PPh3)4, toluene; (c) NaOH, methanol/CH2Cl2; (d) Pd(PPh3)4/
CuI, TEA, or TEA/toluene; (e) Pd(PPh3)4/CuI, TEA/toluene, high dilution;
and (f) Bu4NF, THF.
Scheme 2. The alternative sequence has the same number of
reaction steps. Here first the pyridines 19a-c, which already
contain one complete side and two corners, were generated via
a series of coupling/deprotection steps. Then the terpyridine unit
was assembled by reacting these pyridines with the bisstannyl
10 to give 14a-c, respectively, in yields from 31 to 55%. Their
purification from monocoupled side product and homocoupled
bipyridine was easily possible due to the large enough difference
in polarity between these compounds with different numbers
of pyridyl substituents.
a Reagents and conditions: (a) Pd(PPh3)4/CuI, TEA; (b) NaOH, methanol/
CH2Cl2; (c) Pd(PPh3)4, toluene; (d) Bu4NF, THF; and (e) Pd(PPh3)4/CuI,
TEA, high dilution.
to the latest possible step and, in the case of incomplete
terpyridine formation, exploit the polarity differences of main
and side products for an easy separation. Another strategy that
avoids different side products with terpyridine units is available30
but was not applied in the context of the present work.
(30) Grave, C. Ph.D. Thesis, FU Berlin, Germany, 2002, http://www.diss.fu-
berlin.de/2002/156/.
9
J. AM. CHEM. SOC. VOL. 125, NO. 23, 2003 6911

A R T I C L E S
Grave et al.
Scheme 6. Synthesis of 58-Membered Cycles 5a-c with Two
Scheme 4. The 58-membered macrocycle 2 with one terpyr-
idine unit was synthesized analogously, i.e., by late assembly
of the terpyridine unit. Synthesis and purification of the large
terpyridyl percursor 26a did not cause any problems. The
cyclization yield was also in a normal range (18%) despite the
considerable increase in cycle size. It is therefore believed that
this strategy could be equally well applied to even larger cycles.
Terpyridine Unitsa
Scheme 5 describes a somewhat different strategy for the
shape-persistent cycle 3 with two terpyridine units. For one of
its two cyclization precursors (33b) the terpyridine unit is formed
at the latest possible step, similarly to above. The other precursor
(30) was obtained by coupling 28b with an excess of the diiodide
29. The synthesis of the iodo-functionalized half-ring 30 starts
from 28a, which is easily generated by Stille coupling of 10
and 27. The yield of 70% for the ethynyl-substituted terpyridine
28a is remarkable. Deprotected terpyridine 28b is reacted with
a large excess of 29 to yield 30. This suppressed the undesired
and difficult to separate off side products. In all cases, no
monocoupled side products were detected. Probably, terpyridine
compounds that carry free ethynyl units are totally consumed
in side reactions (i.e., homocoupling to linear compounds with
six pyridine units) during the reaction or workup. Different from
14a in Scheme 2, the purification of halo-functionalized 30 via
column chromatography is not problematic, and most of 29
could be recovered during workup.31 The same holds true for
35a,b (Scheme 6). Ring closure was carried out under the above
conditions and afforded the 46-membered cycle 3 in a yield of
27%.
a Reagents and conditions: (a) Pd(PPh3)4/CuI, TEA/toluene and (b)
Pd(PPh3)4/CuI, TEA/toluene, high dilution.
Scheme 7. Synthesis of 40-Membered Cycle 4 by Oxidative
Alkynyl-Alkynyl Couplinga
a Conditions: Pd(PPh3)2Cl2/CuI, O2, THF/piperidine, pseudo-high dilu-
tion.
Scheme 6. From a Sonogashira coupling of terpyridine 33b
with 1,4-diiodobenzene or 34, the larger iodinated ring precursor
35a or 35b, respectively, was synthesized. Ring closure of 33
and 35 under the above conditions led to 58-membered
macrocycles 5a-c in yields between 11 and 19%. These
somewhat lower yields are probably due to the cycle’s lower
solubilities leading to a precipitation of product during prepara-
tive GPC.
Scheme 7. Finally, the 50-membered macrocycle 4 with two
terpyridine units was obtained by Pd2+-catalyzed oxidative
ethynyl-ethynyl coupling of 33b.23 Also here, the low solubility
of the macrocycle in practically all common organic solvents
is probably responsible for the low yield. The reaction conditions
were therefore not optimized.
Figure 2. Typical GPC chromatogram of the raw product of a ring closure
reaction (of cycle 1b). The dominant peaks can be assigned to the macrocyle
1b and cyclic side product [1b]2.
Scheme 3. The ring-closure reactions of compounds 20a-c
and 21a,b were done under high dilution conditions (ca. 0.0015
M) in sealed vessels as previously described.15 The product
mixture contained the 46-membered macrocycles 1a-d, linear
and/or cyclic oligomers (e.g., [1a]2-[1c]2, the cyclic oligomers
with the double molar mass of the corresponding macrocycles
1a-c, see below), and some insoluble material. The raw mix-
tures were analyzed by GPC, and both the cycle fraction and
the next higher molar mass fraction were isolated by preparative
GPC (Figure 2). The typical yields for this second fractions
were lower than those for the targeted cycles by approximately
a factor of 2. The yields of 17-21% of the macrocycles are in
the range of what has been described in the literature for similar
cycles.1a,15,16,31 The oligomers may be interesting candidates for
studying their potential to form helical secondary structures, a
question that is presently of some interest.32
Comments Regarding Cycles’ Characterization. The char-
acterization of the macrocycles is not complete in all cases
because of occasional solubility problems (13C NMR spectra
could not be recorded for 3, 4, and 5a,b) and sometimes extreme
line broadening of the NMR signals. Additionally the chemical
1
shifts in both H and 13C NMR spectra sometimes showed a
strong concentration and temperature dependence.1q,6a,13a,14,33
The MALDI-TOF spectra for 1b, 1c, 2, 3, 4, and 5a-c
showed the expected signals at m/z ) [M + H] or [M + Na/
K]. For 1c, 3, and 5c low-intensity peaks, which appeared at
higher values, could not be assigned. They do not correlate to
open-chain dimers expected from incomplete reaction and may
represent either adducts or impurities (not observed by GPC
with UV detection). For cycles 1a and 1d FAB(+) measure-
ments were done; the structure of these compounds was further
proven by X-ray diffraction (see below).
(31) Schmittel, M.; Ammon, H. Synlett 1999, 6, 750-752.
(32) See, for example: Gong, B. Chem. Eur. J. 2001, 7, 4336-4342.
9
6912 J. AM. CHEM. SOC. VOL. 125, NO. 23, 2003

Shape-Persistant Macrocycles with Terpyridine
A R T I C L E S
Figure 3. Comparison of the 1H NMR spectra (CDCl3, 500 MHz) of cycle 3 (top) and cyclic side product [3]2 (bottom). The numbered peaks were assigned
by HETCOR and COSY measurements; + ) solvent.
Unfortunately, correct data from elemental analysis could be
obtained for neither of the cycles. This was attributed to
complexed solvents and/or salts. To produce solvent-free NMR
spectra, for example, the samples were treated with deuterated
solvents until all protonated solvent was exchanged.
By and large, the characterization of these compounds
contained in the second GPC fraction (see Figure 2) was more
complicated. They showed a high solubility but unfortunately
could not be crystallized into single crystals. From the NMR
spectra it was concluded that they do not have end groups and
therefore should be cyclic. This was also supported by MALDI-
TOF mass spectrometric analyses which showed the expected
m/z values [M + H] and [M + Na/K] for a compound with a
The 58-membered macrocycles with two terpyridine units and
only four side chains, 5a and 5b, showed extremely broad
1
signals in their H NMR spectra. For 5a when measured in
1
double molar mass. Figure 3 shows the H NMR spectra of
concentrated CDCl3 solutions, signals of the terpyridine unit
(8.0-9.0 ppm) broadened so much that they virtually disap-
peared in the baseline. Upon increasing dilution (not quantified),
they could be made visible but were still very broad. The
spectrum’s resolution could be further enhanced when deuterated
trifluoracetic acid was added. The solubility of 5a improved
with decreasing pH. The concentration and pH dependence of
the signal broadening are probably due to aggregation; the latter
was observed before for similar cycles.31
macrocycle 3 and the compound of double molar mass contained
in fraction 2. As can be seen, the spectra are very similar in
their simplicity and do not show end groups. It is assumed that
fraction 2 is the cycle built from four precursors. It cannot,
however, be excluded that it has a catenane structure with two
intertwined cycles 3.
Solubility. All ring precursors are well soluble in common
solvents such as CHCl3, THF, or toluene. When these com-
pounds are transformed into the more rigid cycles, solubility
can be expected to decrease.35 Quantitative numbers for the
solubility of the macrocycles were determined in chloroform,
with an additional measurement for 1a in toluene (Table 1).
For each cycle two measurements were made whose results
deviate because of practical problems during measurement. For
the poorly soluble cycles, very small amounts had to be weighed.
For the well-soluble cycles, the solutions became viscous and
extremely difficult to handle, as they tended to dry and
precipitate. The average values, therefore, are rough estimates.
For the remaining 58-membered cycle, 5c, characterization
was much easier, presumably due to its larger number of
solubilizing side chains, which not only increased the cycle’s
solubility but also decreased its tendency to aggregate. If the
spectra were recorded directly after sample preparation, well-
resolved signals were obtained, and characterization of 5c by
1H, 13C, and HETCOR NMR spectrometry was allowed.
For all cyclization reactions, defined higher oligomeric
fractions could be observed in analytic GPC. For all cycles
except 1d, a second fraction with twice the mass of the dimeric
macrocycle was isolated. MALDI or NMR measurements, as
far as they could be obtained, prove the cyclic nature of this
higher fraction. A definitive decision, whether these cyclic
oligomers would be monocycles (from four precursors) or
catenanes of two intertwined macrocycles, however, cannot be
drawn from these results.34
(33) See, for example: Lahiri, S.; Thompson, J. L.; Moore, J. S. J. Am. Chem.
Soc. 2000, 122, 11315-11319.
(34) For a work dealing with a similar problem, see: (a) Samor´ı, P.; Ja¨ckel, F.;
U¨ nsal, O¨ .; Godt, A.; Rabe, J. P.ChemPhysChem 2001, 2, 461-464. (b)
U¨ nsal, O¨ .; Godt, A. Chem. Eur. J. 1999, 5, 1728-1733.
(35) Moore, J. S.; Zhang, J.; Wu, Z.; Venkataraman, D.; Lee, S. Macromol.
Symp. 1994, 77 (International Symposium on New Macromolecular
Architectures and Supramolecular Polymers 1993), 295-231.
9
J. AM. CHEM. SOC. VOL. 125, NO. 23, 2003 6913

A R T I C L E S
Grave et al.
a
Table 1. Solubilities of Macrocycles in CHCl3
entry
measured solubility [mg/mL]
average [mg/mL]
1a
1aa
1b
1d
2
4.4, 4.4
7.3, 3.4
388, 247
371, 378
>325b, >225b
2.5, 2.6
∼4
∼5
∼300
∼350
>300b
∼3
3
4
7.0, 4.9
∼6
b
a Solubility in toluene. Values were measured from an unsaturated
solution.
Because of the partially large differences between them, some
conclusions can nevertheless be drawn. The most stunning
observation is that the cycles are either very poorly (1a, 3, 4)
or extremely well soluble (1b, 1d, 2). A change in solvent from
CHCl3 to toluene has, at least for 1a, practically no effect.
Three factors were considered to possibly have an impact on
the cycles’ solubility: number and nature of side chains, number
of terpyridine units, and ring size. A cycle carrying a terphenyl
unit with one exocyclic hexyloxy side chain (1a) has a much
lower solubility than the same cycle with a terphenyl unit
carrying four hexyl side chains (1d). Obviously, the larger the
number of flexible chains per cycle, the larger is its solubility.
Size, however, does not seem to play an important role. If one
compares the 46-membered 1b and its 58-membered analogue
2, both of which have the same number of side chains but differ
by two phenylacetylene units (ratio ring members/side chains
9.2 for 1b, 11.6 for 2), there is practically no difference in
solubility.
Terpyridine units generally decrease solubility; cycles with
two such units are generally poorly soluble (cf. 1b with 3). This
is not, as one may think, an incorrect observation due to a
smaller number of side chains of 3. For example, for poorly
soluble 3 (with two terpyridine units), the ratio ring members/
side chains is 11.5, and the ratio aromatics/side chains is 2.5,
while the well-soluble 2 (with one terpyridine unit) has nearly
the same values of 11.6 and 2.4, respectively. An influence of
the higher symmetry point group of the cycles with two
terpyridines (D2h) versus those with one such unit (C2V) cannot
be excluded, however.
Perhaps the most interesting observation here results from
the comparison of cycles 1a and 1b, which differ by a factor of
80 to the advantage of 1b. The only difference between these
cycles is that 1a carries four hexyloxy side chains in its corners,
while 1b has four hexyloxymethyl side chains. Even if one
expects a better solubility in the case of the benzyl ether, where
the oxygen is separated from the aromatic and free to rotate,
while for the phenol ether, the aromatic-oxygen bond has
partially double bond character due to resonance effects, this
large effect is astonishing. Other than 3 and 4, which are nearly
insoluble in common solvents even under heating, 1a dissolves
easily in warm THF or benzene but precipitates, however, under
cooling.
Figure 4. ORTEP36 diagram of 1a, 50% probability ellipsoids: (a) view
perpendicular to the macrocycle; (b) view parallel to the macrocycle; solvent
molecules and hydrogen atoms omitted for clarity.
structure determination of the macrocycles containing two
terpyridine units failed, all cycles containing one terpyridine
unit (1a-d, 2) form nice crystals, and the structure of three of
these compounds (1a, 1b, and 1d) could be determined by X-ray
crystallography. This was done at low temperature to prevent
disintegration of the crystals by loss of solvate molecules. All
three compounds crystallize in the triclinic space group P1h with
one molecule in the asymmetric unit.
Although the three systems structurally studied possess the
same cyclic backbone, the number and lengths of their side
chains and the solvent used for the crystallization have strong
effects not only on the crystal packing but also on the distortion
of the cycle from planarity. 1a carries five hexyloxy side chains,
and 1b one hexyloxy and four hexyloxymethyl side chains. This
minor differencestogether with the fact that 1a and 1b were
crystallized from different solvents, benzene and toluene,
respectivelysleads to a significantly different packing, while
the backbones of the individual cycles show a similar conforma-
tion and form almost planar sheets, as can be seen in a side
view of the molecules 1a and 1b (Figures 4b and 5b). Cycle
1b, however, possesses great positional freedom perpendicular
to the plane sheet, as expressed by the large thermal ellipsoids.
One of the aromatic rings of the terphenyl unit is tilted out of
plane (for 1a, C4g-C5g-C1h-C6h 147.2°, Figure 4a; for 1b,
C6e-C1e-C6d-C5d -149.6°, Figure 5a). As observed in most
sterically unrestricted terpyridines, the NCCN torsion angles are
close to 180° (for 1a, N1c-C1c-C5b-N1b 173.7°, N1b-
C1b-C3a-N1a -175.6°; for 1b, N1-C5a-C6a-N2 166.8°,
N2-C10a-C11a-N3 178.5°).20c,37 The data do not allow an
unambiguous assignment of the nitrogen atoms, which may even
be disordered.
The 1H NMR spectra of the well-soluble cycles show strong
shift dependencies on concentration and temperature. This
indicates aggregation, which will be the subject of future
investigation.
X-ray Crystallography. Interestingly, there is a striking
difference in the tendency of the macrocycles to crystallize.
Whereas all attempts to get suitable crystals for an X-ray
The largest distance across the hole inside the macrocycle is
found between the acetylenic carbon atoms (for 1a, C4l-C10l
1.992 nm; for 1b, C1c-C7g 1.919 nm). The shortest inner
(36) ORTEP3 for Windows. L. J. Ferrugia, J. Appl. Crystallogr. 1997, 30, 565.
(37) See, for example: (a) Bessel, C. A.; See, R. F.; Jameson, D. L.; Churchill,
M. R.; Takeuchi, K. J. J. Chem. Soc., Dalton Trans. 1992, 3223-3228.
(b) Constable, E. C.; Khan, F. K.; Marquez, V. E.; Raithby, P. R. Acta
Crystallogr. 1992, C48, 932-934.
9
6914 J. AM. CHEM. SOC. VOL. 125, NO. 23, 2003

Shape-Persistant Macrocycles with Terpyridine
A R T I C L E S
Figure 5. ORTEP36 diagram of 1b, 50% probability ellipsoids: (a) view
perpendicular to the macrocycle; (b) view parallel to the macrocycle;
hydrogen atoms and solvent molecules omitted for clarity.
diameter for 1a is C4c-C5h 1.248 nm and for 1b C3a-C4d
1.200 nm. Similar to Moore’s all-hydrocarbon phenylacetyl-
enemacrocycles5b and our recently published bipyridine-contain-
ing macrocycles,1a,15 both compounds form layered struc-
tures.38
Figure 6. Space-filling model of the packing (SCHAKAL9939) of 1a, view
along 0-1-1. Solvent molecules are colored yellow. (a) One layer. Different
colors were used to distinguish molecules generated by the following
symmetry operations: x, y, z red; -2+x, 1+y, z green; -x, 2-y, -z violet;
2-x, 1-y, -z gray; 2+x, -1+y, z blue; 2-x, -1-y, 1-z orange; -x, -y,
1-z turquoise. (b) Shingle-roof type cut through consecutive layers.
Compound 1a‚3C6H6 crystallizes with three molecules of
benzene per formula unit. Figure 6a presents a view along
0-1-1 on seven adjacent molecules of 1a within one layer.
Solvate molecules are colored in yellow for clarity. Within the
layers the molecules show close contacts at the terpyridine units
and four of the five O-hexyl side chains. The almost linear side
chain D sticks between the almost linear chain I and chain K
of a neighboring macrocycle. Chain K does not show the
favorable all-trans conformation of extended alkanes, but is
restricted by a neighbored benzene molecule. The chain G leans
on one aromatic ring of a terphenyl unit of a neighboring
macrocycle. This results in the above-mentioned torsion of the
aromatic ring H while all other aromatics are in a plane with
the cyclic backbone. The only side chain with no close contact
to neighboring molecules within the same layer is chain E, which
has interestingly very large thermal parameters and is near the
solvate molecules with great positional freedom, as can be seen
in Figure 6b. All in all, the packing within one layer reminds
one of interacting gear wheels. The remaining free volume is
filled with side chains and solvate molecules interlocking
adjacent layers. The sides of the cycles overlay in all directions,
whereby subsequent layers are shifted against each other (Figure
6b). Removal of the solvent molecules would create channels
almost along 0-1-1. However, these are not perpendicular to
the plane of the molecules.
Compound 1b‚2C7H8 crystallizes with two molecules of
toluene in the asymmetric unit. Figure 7a shows one layer of
molecules in a view along the crystallographic c axis. Again,
the molecules within the layer are in close contact at the
terpyridine units, forming parallel “ribbons” of rings at a closest
possible distance. These ribbons are held at a distance by
interdigitation of the antiparallely oriented two in-plane side
chains of each ring (assigned as C and E for one ring). The
methyl group of the hexyl chain E turns the benzene ring D
out of the macrocycle’s plane. Further close contacts are visible
at the starting point of chain B. The remaining side chains point
above and below the backbone of the macrocycle, linking
together adjacent layers, as can be seen in Figure 7b. The side
chains H and G are oriented almost perpendicular to the
macrocycle’s backbone. Chain H, which shows the largest
disorder, penetrates the interior of a macrocyclus of the next
layer and reaches into the space between the sides of two
adjacent macrocycles in the next but one layer; chain G shows
opposite behavior.
Compound 1d‚3C6H6 contains three molecules of benzene
per formula unit. The cycle carries eight side chains which do
not exclusively point exocyclically. This results in a totally
different crystal structure, with the macrocycles, instead of
(38) For another recent example, see: Ho¨ger, S.; Morrison, D. L.; Enkelmann,
V. J. Am. Chem. Soc. 2002, 124, 6734-6736.
9
J. AM. CHEM. SOC. VOL. 125, NO. 23, 2003 6915

A R T I C L E S
Grave et al.
Figure 8. ORTEP36 diagram of 1d, 50% probability ellipsoids: (a) view
perpendicular to the macrocycle, all but one solvent molecule omitted for
clarity; (b) view parallel to the macrocycle; solvent molecules and hydrogen
atoms omitted for clarity.
Figure 7. Space-filling model of the packing (SCHAKAL9939) of 1b: (a)
view along 0-0-1, solvent molecules colored in yellow for clarity. Different
colors were used to distinguish molecules generated by the following
symmetry operations: x, y, z red; -1+x, 1+y, z green; 2-x, 2-y, 1-z
violet; 3-x, 1-y, 1-z gray; 1+x, -1+y, z blue; -x, -2-y, -z orange;
-1-x, -1-y, -z turquoise; (b) view roughly along 0-1-0, solvent
molecules omitted for clarity. Different colors were used to distinguish
molecules generated by the following symmetry operations: x, y, z red;
1-x, 1-y, 1-z yellow; -1+x, y, z green; and 2-x, 1-y, 1-z blue.
forming planar sheets, being heavily distorted from planarity
(C1l-C2l-C5l-C6l 25°, C6h-C1h-C5g-C6g 70°, Figure 8).
The inner diameters, however, are quite well in accordance with
those of 1a,b (C4l-C10l 1.840 nm, C4c-C5h 1.343 nm), and
this holds true also for the terpyridine conformation (N1a-C3a-
C1b-N1b 176.2°, N1b-C5b-C1c-N1c -171.1°).
Figure 9. Packing diagram (ORTEP36) of 1d, view along 1-0-0, 20%
probability ellipsoids. Side chains and hydrogen atoms were omitted for
clarity.
The packing follows a more simple AB layer pattern along
the a axis and differs from the other cycles because there is
overlap between the cycles only in the direction of the b axis
(Figure 9). Thus macrocyclic backbone and side-chain areas
alternate along the c axis. Within the side-chain area remain
two neighboring channels which are shared by the most highly
disordered components of the structure. They are filled with
the two benzene solvate molecules per formular unit, which lie
roughly perpendicular to each other and are separated by the
outer side chain F. One solvate molecule has extremely large
thermal parameters and thus great positional freedom. The inner
F and H side chains stretch through the interior space of the
cycle with one benzene molecule sandwiched between them.
One K chain from a cycle of the next layer also reaches into
the cycle. The remaining side chains pack into the space between
the cycles.
Scanning Tunneling Microscopy. While much endeavor has
been addressed recently to understanding and controlling the
physisorption of phenylene-based rodlike and disklike molecules
into ordered nanostructures at the solution-HOPG (highly
oriented pyrolitic graphite) interface,40 only few studies have
been reported on the formation of ordered architectures of
(39) SCHAKAL: E. Keller, J. Appl. Crystallogr. 1989, 22, 12-22.
(40) (a) Samor´ı, P.; Fechtenko¨tter, A.; Bo¨hme, T.; Ja¨ckel, F.; Mu¨llen, K.; Rabe,
J. P. J. Am. Chem. Soc. 2001, 123, 11462-11467. (b) Samor´ı, P.; Severin,
N.; Mu¨llen, K.; Rabe, J. P. AdV. Mater. 2000, 12, 579-582. (c) Stabel, A.;
Herwig, P.; Mu¨llen, K.; Rabe, J. P. Angew. Chem., Int. Ed. Engl. 1995,
34, 303-307. (d) Askadskaya, L.; Boeffel, C.; Rabe, J. P. Ber. Bunsen-
Ges. Phys. Chem. 1993, 97, 517-521. (e) Rabe, J. P.; Buchholz, S. Phys.
ReV. Lett. 1991, 66, 2096-2099.
9
6916 J. AM. CHEM. SOC. VOL. 125, NO. 23, 2003

Shape-Persistant Macrocycles with Terpyridine
A R T I C L E S
is not elongated but rather appears as a spherical spot suggests
that the two external pyridine rings are nonplanar with respect
to both the central one and to the basal plane of the HOPG
surface. The alkoxy and alkyl side chains are likely to be
arranged according to a 2-fold symmetry, as observed for other
“hairy disks”.40c,d In the present case, the high conformational
mobility at the surface, which occurs on a time scale faster than
the scanning frequency, and/or their electronic states, which are
far away from the Fermi level of the HOPG, did not allow us
to resolve their structures. The overall molecular arrangement
on the surface of graphite is loosely packed. This means that
the gain in enthalpy upon adsorption on HOPG is not very
high.40a This is most likely the reason for the low stability in
time of the visualized 2D crystals which indeed revealed a high
dynamics in the seconds time scale. Moreover, the nonplanar
conformation, which has been observed by X-ray diffraction in
the crystal of 1d, makes its adsorption at surfaces expensive
from an entropy viewpoint, since a planarization needs to occur.
Figure 10. STM image of compound 1d recorded in current mode at the
solution-HOPG interface. Tunneling parameters Ut ) 640 mV, It ) 36
pA.
Ordered nanostructures of 5a and 5c could not be observed
on HOPG. The reason for this might be found in the thermo-
dynamics of the physisorption at the solid-liquid interface.
These two macrocycles are bigger than 1d and consequently
have a larger cavity in the middle. This larger hole represents
a lack in enthalpy gain, which is one basis for the physisorption
at surfaces. As already noticed for 1d the low stability could
be ascribed to a low gain in enthalpy. Upon increasing the size
of the macrocycles, this enthalpic gain becomes too small to
overcome the loss in entropy associated with transition from
the 3D of a solution to the 2D of the surface and to the
translational entropy needed to create a tightly packed supramo-
lecular architecture at surfaces.
conjugated macrocycles that were based on phenylene-
ethynylene6a or thiophene units.1d,6b,41 Also taking into account
that little is known regarding the self-assembly at surfaces of
pyridine-based conjugated systems,42 we have considered it most
important to extend our exploration on the aforementioned
pyridine-based macrocycles to their behavior in ultrathin films.
The STM investigations have been executed at the solution-
HOPG interface on macrocycles 1d, 5a, and 5c. Among these
three different macrocycles investigated, only compound 1d
could physisorb into highly ordered nanostructures at the
solution-HOPG interface. The STM current image is shown
in Figure 10. In the same figure the unit cell of the molecular
arrangement and the model of the packing are superimposed.
The adsorbate unit cell possesses an oblique geometry: it
amounts to a ) 3.7 ( 0.2 nm, b ) 3.5 ( 0.2 nm, R ) 54 ( 3°.
The contrast in the STM current images is mainly ruled by
the difference between the energies of the molecular states of
the adsorbate and the Fermi level of the substrate. Thus, due to
a larger energy difference, aliphatic chains on HOPG generally
appear darker than moieties with π-electrons, provided that the
moieties are lying equally flat on the substrate.34a In our example
of a three-dimensional (non perfectly flat) object, the contrast
depends additionally on the degree of spatial overlap of the
electronic states of the adsorbate with those of the substrate.40a,43
In view of these facts, one can try to interpret the observed
contrast in the current STM images displayed in Figure 10. An
angular bright segment is visible and can be ascribed to the
terphenylene unit. With a slightly lighter contrast, in the
positions that are placed at the edges of the drawn unit cells,
the unsubstituted terpyridine segments are located. Their lower
contrast, if compared to the terphenylene units, can be explained
with both the different electronic structure of the moiety and a
different coupling with the substrate’s electronic states. This
latter effect would be due to a diverse location along the z axis
of the conjugated orbitals with respect to the HOPG states. In
particular, the fact that the contrast of the terpyridine segments
Conclusions
A set of large terpyridine-containing macrocycles with 46
(1a-d, 3), 50 (4), and 58 atoms (2, 5a-c) in the cyclic core
was synthesized using a flexible building block approach. This
approach allowed one to systematically vary the cycles’ sizes
and side-chain patterns and has the potential to be applied to
even larger cycles. In most cases it proved best to generate the
terpyridine units at a late stage of the synthetic sequence. The
macrocycles’ solubility, a significant factor for both their
handling and characterization, is strongly influenced by even
minor variations in the cycles’ side-chain pattern. The three
cycles 1a, 1b, and 1d could be grown into single crystals and
their structures solved by low-temperature X-ray diffraction.
They form layered structures, and the cycles of consecutive
layers form columnar stacks that are filled with solvent
molecules and flexible side chains. The respective detailed
packing strongly depends on the side-chain pattern. Apart from
cycles 1a and 1b, whose cores are planar, cycle 1d in the crystal
significantly deviates from the planar geometry. Monolayers of
this cycle on HOPG were investigated by STM. Though the
resolution of the images obtained is not very high, a packing
model for this “2D crystal” could nevertheless be derived. Both
the image’s nonoptimal resolution and the monolayer’s low
stability in time were attributed to 1d’s presumed relatively low
adsorptive attraction on HOPG, which may be caused by
preferred nonplanar conformations as they were observed in the
single crystal.
(41) Mena-Osteritz, E.; Ba¨uerle, P. AdV. Mater. 2001, 13, 243-246.
(42) (a) Ziener, U.; Lehn, J.-M.; Mourran, A.; Mo¨ller, M. Chem. Eur. J. 2002,
8, 951-957. (b) Semenov, A.; Spatz, J. P.; Mo¨ller, M.; Lehn, J.-M.; Sell,
B.; Schubert, D.; Weidl, C. H.; Schubert, U. S. Angew. Chem., Int. Ed.
1999, 28, 2547-2550.
(43) Lazzaroni, R.; Calderone, A.; Bre´das, J. L.; Rabe, J. P. J. Chem. Phys.
1997, 107, 99.
9
J. AM. CHEM. SOC. VOL. 125, NO. 23, 2003 6917

A R T I C L E S
Grave et al.
replacement parameters. The aromatic rings of 1b and the solvent
molecules of all compounds were refined as rigid bodies. The hydrogen
atoms were included at geometrically calculated positions and refined
using a “riding model” with isotropic thermal parameters (U(H) )
1.5Ueq(C) for methyl groups and U(H) ) 1.2Ueq(C) for all other H
atoms). Crystal data, experimental conditions, and refinement values
are listed in the Supporting Information. ORTEP36 for Windows and
SCHAKAL9939 were used for preparing the graphical representations.
The molecular structures of 1a, 1b, and 1d are depicted in Figures 4,
5, and 8. Packing diagrams are presented in Figures 6, 7, and 9.
Experimental Section
Synthesis. Compounds 6,29a 10,27 12b,15 16b,15 22,29a 29,29b and 3444
were prepared according to literature procedures. For the synthetic
procedures and analytical data for all other compounds, please see the
Supporting Information.
Solubility Measurements. Concentrated solutions of the compounds
were gained by shaking an excess of material in CHCl3 or toluene
overnight. After filtration, defined volumes (ca. 0.5-1 mL for the less
soluble, ca. 0.01 mL for the better soluble substances) were transferred
with an Eppendorf pipet to a probe glass, the solvent was evaporated,
and the residue was dried in high vacuum and weighed.
X-ray Analysis. Crystals of 1a suitable for X-ray crystallography
were obtained by crystallization from hot benzene. Attempts to get
suitable crystals from THF or toluene failed. Compounds 1b and 1d
are considerably more soluble than 1a in common solvents and
crystallize less readily. Crystals were obtained by slow diffusion of
methanol into a solution in toluene (for 1b) or benzene (for 1d) at
ambient temperature.
STM Investigation. Almost saturated solutions of the macrocycles
1d, 5a, and 5c in 1,2,4-trichlorobenzene were applied to the basal plane
of a freshly cleaved HOPG surface. By changing the tunneling
parameters during the STM imaging, namely, the voltage applied to
the tip (Ut) and the average tunneling current (It), it was possible to
switch between the visualization of the adsorbate layer to that of the
underlying HOPG substrate with a lattice resolution. This enabled us
to calibrate the images of the molecular adsorbate with respect to the
underlying HOPG.
As all three compounds crystallize with solvent molecules and
therefore tend to disintegrate with loss of the included solvate, the
crystals were mounted out of saturated solutions at low temperature
on the top of a glass fiber and the data collection was performed at
low temperature using a Bruker-AXS SMART CDD diffractometer.
A total of 600 frames (∆ω ) 0.3°) for each run were collected for
three φ positions (0°, 90°, and 240°), resulting in 1800 frames for each
data set. The data were reduced to Fo2 and corrected for the absorption
effects using SAINT45 and SADABS,46 respectively. The structures were
solved using direct methods (SIR9247 1a and 1b, and SHELXS48 1d)
and refined using full matrix least squares (SHELXL 9748).
All non-H atoms were refined anisotropically. The hexyl side chains
were refined with the help of similar distance restraints for the 1,2 and
1,3 bond distances as well as rigid bond restraints for the anisotropic
Scanning tunneling microscopy investigations were carried out using
a picoAmp-Nanoscope IIIa (Digital Instruments), which allows the
detection of a tunneling current of 1 pA. This latter characteristic is an
essential prerequisite for studying in a noninvasive way organic layers
that are known to be poorly conductive. Images were recorded with a
scan rate of ca. 13-18 line/s with a resolution of 512 × 512 pixels.
The Pt/Ir (85%, 15%) STM tips were prepared by electrochemical
etching.
Acknowledgment. This work was supported by the Deutsche
Forschungsgemeinschaft (Sfb 448, TPs A1 and B5) and the
Fonds der Chemischen Industrie. We thank Ms. E. Franzus and
Dr. G. Holzmann for their competent help with EI and FAB
mass spectrometric analyses and Ms. C. Zimmermann for her
help with GPC analyses.
(44) Hensel, V. Ph.D. Thesis, FU Berlin, Germany, 1998.
(45) SAINTPLUS Software Reference Manual, Version 5.054; Bruker-AXS:
Madison, WI, 1997-1998.
(46) Blessing, R. H. Acta Crystallogr., Sect. A 1995, 51, 33-38. SADABS; Bruker
AXS, 1998.
(47) SIR92, A program for crystal structure solution: Altomare, A.; Cascarano,
G.; Giacovazzo, C.; Guagliardi, A. J. Appl. Crystallogr. 1993, 26, 343-
350.
(48) Sheldrick, G. M. SHELX97 [includes SHELXS97, SHELXL97, CIFTAB],
Programs for Crystal Structure Analysis (Release 97-2); Institut fu¨r
Anorganische Chemie der Universita¨t: Tammanstr. 4, D-3400 Go¨ttingen,
Germany, 1998.
Supporting Information Available: Synthetic procedures and
analytical data for the compounds prepared; X-ray crystal-
lographic files in CIF format for compounds 1a, 1b, and 1d.
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA034029P
9
6918 J. AM. CHEM. SOC. VOL. 125, NO. 23, 2003