Rational construction of 2D and 3D borromean arrayed organic crystals by hydrogen-bond-directed self-assembly

Full text of DOI:10.1002/anie.200806198

Angewandte
Chemie
DOI: 10.1002/anie.200806198
Organic Crystals
Rational Construction of 2D and 3D Borromean Arrayed Organic
Crystals by Hydrogen-Bond-Directed Self-Assembly**
Yong-Biao Men, Junliang Sun, Zhi-Tang Huang, and Qi-Yu Zheng*
Crystal engineering, which relies on intermolecular interac-
tions to prepare the crystalline organic solids and coordina-
tion polymers with desired properties, has become an
important field of chemistry and material science.[1] The
construction of large networks with appropriate packing
geometries and topologies from well-designed building blocks
is one of the main goals in crystal engineering. This field has
thus become important, as the properties of materials depend
on the overall crystal structures.[2] To date, crystal structures
are not entirely predictable owing to the numbers of non-
covalent interactions with little energy difference.[3] There are
only a few cases for which the actual crystal structure was
deduced ahead of knowledge of the chemical composition.[4]
Especially for the entanglement of molecular crystals, which
are expected to be 2D or 3D networks, the rationalization and
prediction of the structures are more difficult. One way to
understand the topologies of crystals is to compare the known
phase diagram of very large systems, in which the vagaries of
individual energetic terms are often averaged away.[5] On the
other hand, the design and construction of molecular net-
works in the crystalline phase may be realized by using the
supramolecular synthon approach[6] or the reticular synthesis
method;[7] however, the desired final frameworks can not be
absolutely guaranteed.[8]
were prepared by the connection of inorganic metal ions,
metal clusters, or anions.[10,11] Predictable construction of
Borromean networks, however, still remains a challenge.
Herein we report the rational construction of Borromean
three-fold 2D!2D entangled layers and n-Borromean 2D!
3D entangled infinite layers with pure organic supramolecular
synthons based on hydrogen-bond-directed self assembly.
To build a Borromean structure, the 63 honeycomb net
(hcb net) is essential. Our synthetic strategy was to construct
the 2D large honeycomb-like networks based on the assembly
of Y-shaped building blocks through the supramolecular
synthons I, II, and III (Scheme 1a). Pure trimesic acid (TMA)
facilitates assembly into a prototypal hcb net through
The Borromean system, which contains nontrivial three-
ring links (no ring is interlocked to another without the help
of the third one), is a very intriguing pattern of crystal
entanglements, because of the topological complexity, struc-
tural integrity, and aesthetic beauty. Molecular Borromean
rings have been realized by Stoddart et al. with an all-in-one
assembly strategy.[9] As far as we know, only a handful of
structures have been reported that consist of three-fold
entangled 63 nets with Borromean links. [10,11] These structures
Scheme 1. a) Hydrogen bond synthons of carboxylic acid (I), acid–
alcohol/water (II), and acid–base (III). b) The structures of TCA and of
four linkers. c) The topology of a six-fold phenyl embrace in a TCA
dimer.
[*] Y.-B. Men, Prof. Z.-T. Huang, Prof. Q.-Y. Zheng
Beijing National Laboratory for Molecular Sciences, CAS Key
Laboratory of Molecular Recognition and Function, Institute of
Chemistry, Chinese Academy of Sciences
Beijing 100190 (China)
carboxylic acid dimer synthon I,[12] and is solvated with
water or methanol to form a pseudohexagonal net with mixed
synthon I and II.[13] Cocrystallization of TMA with 4,4’-
bipyridine (bipy) or 1,2-bis(4-pyridyl)ethane (bpea) results in
the formation of honeycomb grids with even larger channels
or cavities by the heterosynthon III.[14] However, it is
impossible to form a Borromean topology with fully planar
networks. Furthermore, the honeycomb cavities should be
large enough to ensure that three hexagonal rings can
penetrate each other.[15] Therefore, the rigid and trigonal
pyramidal molecule 1,3,5-tris(4-carboxyphenyl)adamantane
(TCA)[16] was taken into the main skeleton for the construc-
tion of Borromean linked organic crystals, which plays an
Fax: (+86)10-6255-4449
E-mail: zhengqy@iccas.ac.cn
Dr. J. Sun
Structural Chemistry and Berzelii Centre EXSELENT on Porous
Materials, Stockholm University
10691 Stockholm (Sweden)
[**] This work was supported by the National Natural Science
Foundation of China (No. 20672118), the Major State Basic
Research Development Program of China (No. 2007CB8008005,
2008CB617501), and the Chinese Academy of Sciences.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200806198.
Angew. Chem. Int. Ed. 2009, 48, 2873 –2876
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2873

Communications
important role in directing the interweaving and the packing
respectively. In all these five kinds of rings shown in Figure 1,
the six apexes are non-coplanar owing to the trigonal
pyramidal structure, which form the vertexes of two triangles.
The triangle formed by apexes at positions 1, 3, and 5 is
parallel to the other triangle formed by apexes at positions 2,
4, and 6, with relative distance of 5.7, 6.0, 7.7, 7.2, and 8.4 ꢁ,
respectively.
Self-inclusion or entanglement is facilitated by such a
trigonal pyramidal structure of TCA and the chair conforma-
tion of these hexagonal nets, resulting in closed-packed
structures.[18] In fact, in the crystal structures of TCA,
[TCA·methanol], [2TCA·3phenazine], and [2TCA·bipy], all
adjacent TCA pairs form a well-known supramolecular motif
called a six-fold phenyl embrace (Scheme 1c),[19] which is
of the crystals. Furthermore, linear or approximately linear
linkers, such as methanol, phenazine, bipy, and 4,4’-azopyr-
idine (azpy), were chosen as the guiding skeletons for larger
and more complicate entangled frameworks (Scheme 1b).
Five structures, namely TCA, [TCA·methanol], [2TCA·3-
phenazine], [2TCA·bipy], and [2TCA·azpy] were determined
by single crystal X-ray diffraction.[17] They all possess 2D
hexagonal or pseudohexagonal networks (Figure 1; Support-
ꢀ
stabilized by multiple edge-to-face C H···p interactions
between phenyl rings from both TCA molecules. Similar to
argentophilic, aurophilic, hydrogen-bonding, halogen-bond-
ing, or p–p interactions found in reported Borromean
links,[10,11] this motif could be considered as a key supra-
molecular synthon, which further connects the 63 net into
more complicated layers or even 3D polycatenated structures
(see below).
As expected, for TCA alone, three hcb nets of TCA are
interpenetrated to form a Borromean weave through multiple
ꢀ
C H···p interactions. Three hcb nets lie on top of another, as
shown in Figure 2a in green!red!blue!green, which
repeats in a cyclic fashion. Thus, three nets in the Borromean
system are linked in such a way that none of them is bonded to
another, but they are not separable. These layered Borro-
mean weaves are further stacked with a small shift along the
layer. As shown in Figure 2c, every third Borromean weave
goes back to the original position, which is similar to the
ABCABC stacking in cubic close packing. In [TCA·metha-
nol], the only difference is that every third Borromean weave
is replaced by a new Borromean weave (Figure 2b), which
consists of three [TCA·methanol] hcb nets in Figure 1b (see
Figure 1. Ball- and stick models showing five kinds of hexagonal or
pseudohexagonal nets found in a) TCA, b) [TCA·methanol] (which also
contains TCA hexagonal nets), c) [2TCA·3phenazine], d) [2TCA·bipy],
and e) [2TCA·azpy]. C gray, N blue, O red, H white. Semitransparent
green triangles indicate that alternative apexes of TCAs are situated at
different levels with respect to the mean plane of the hexagon.
ing Information, Figure S5–S9). In TCA, the molecule
predictably self-assembles into an infinite 2D hcb net with
30 ꢀ 30 ꢁ cavities through synthon I. In [TCA·methanol],
there are two kinds of hcb nets, and in each unit cell, six nets
consist of TCA molecules, whereas the other three (30 ꢀ 30 ꢁ
cavities) consist of both TCA and methanol molecules
assembled through synthon II. Unlike the 2:3 cocrystal of
trimesic acid with bipy reported by Zaworotko and Shar-
ma,[14a] only the hcb nets with 2:1 stoichiometry were
observed in cocrystals of TCA with phenazine, bipy, and
azpy, even upon dissolving a 2:3 ratio of TCA and the linkers
in the solvent. Owing to two additional phenazine molecules
in the residual volume, the stoichiometry of the entire
structure is 2:3 in [2TCA·3phenazine]. In these three
structures, corresponding linkers were only inserted at one
third of hcb edges by the supramolecular synthon III as shown
in Figure 1c–e, and the final pseudohexagonal nets have
larger internal dimensions of 30 ꢀ 33, 30 ꢀ 40, and 30 ꢀ 37 ꢁ,
Figure 2. a) Space-filling model of one Borromean weave consisting of
three TCA hcb nets in TCA. b) Space-filling model of one Borromean
weave consisting of three TCA·methanol hcb nets in [TCA·methanol].
c) Topological view of four separable TCA Borromean weaves in TCA.
Three hcb nets in the same Borromean weave are shown as red, blue
and green.
2874 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 2873 –2876

Angewandte
Chemie
also Supporting Information, Figure S10). To the best of our
knowledge, [TCA·methanol] is the only reported structure
containing two different Borromean weaves.
(Figure 4a). In contrast to the 2D Borromean and n-
Borromean links above, [2TCA·azpy] has an inclined poly-
catenation: twelve identically orientated hydrogen-bonded
Similar Borromean weaves were formed in [2TCA·3phe-
nazine] and [2TCA·bipy] (Figure 3a,b). However, owing to
the non-equivalent edge lengths of hcb nets (Figure 1c,d),
Figure 4. a) One unit cell of [2TCA·azpy]; azpy molecules are shown
using a space-filling model and TCA is shown with lines. b) The
topology of inclined polycatenated networks in [2TCA·azpy]; the main
hexagonal ring is shown red, and the twelve hexagonal rings inter-
locked with it are shown green and blue.
hcb nets thread through the same void of the thirteenth hcb
net resulting in a 3D inclined polycatenated network (Fig-
ure 4b). The formation of these two inclined hcb nets, might
be due to the small parallel shift of 1.1 ꢁ between two
pyridine moieties of azpy, which results in two orientated azpy
linkers, as well as two inclined hcb nets in [2TCA·azpy].
In conclusion, we have successfully constructed five
hydrogen-bonded 63 honeycomb nets. Although the topolo-
gies of the 2D networks are same, they are different in the way
they entangle. 2D!2D Borromean entangled layers, mixed
2D!2D Borromean entangled layers, and 2D!3D n-Borro-
mean entangled infinite layers have been obtained by the
crystallization of TCA and cocrystallization of TCA with
methanol, phenazine, and 4,4’-bipyridine. However, the
cocrystalliation of TCA and 4,4’-azopyridine formed a 3D
inclined polycatenation network owing to two orientated azpy
linkers. It is the first time that Borromean linked organic
crystals have been constructed systematically based on
hydrogen-bonding directed self-assembly.
Figure 3. a) Space-filling model of one Borromean weave consisting of
three TCA·phenazine hcb nets in [2TCA·3phenazine]. b) Space-filling
view of one Borromean weave consisting of three TCA·bipy hcb nets in
[2TCA·bipy]. c) Topological view of six adjacent hcb nets in [2TCA·phe-
nazine], which is similar to [2TCA·bipy]. Three thick hcb nets and three
thin hcb nets form two Borromean weaves, which further interlock
each other by the TCA dimers between thick green and thin blue nets
around pink circles. Three hcb nets in the same Borromean weave are
shown as red, blue, and green.
half of the TCA molecules in the blue and green hcb nets do
not form TCA dimers (six-fold phenyl embraces) inside this
Borromean weave. Instead, these TCAs form dimers with
TCAs from either up or down Borromean weaves (Fig-
ure 3c). Thus, all Borromean weaves are locked together
through this kind of TCA dimer, resulting in a 3D poly-
catenated net. A particularly interesting aspect in this entire
3D network is the entanglement of three adjacent hcb
networks as a Borromean weave, as shown using the thick
red, thick green, and thin blue net in Figure 3c. This new type
of network is a so-called n-Borromean link, and only two
examples have been reported to date.[10,11a,c] Figure 3c shows a
simple illustration of n-Borromean links, where the absolute
heights are in order of thin green!thin red!thin blue!
thick green!thick red!thick blue, but the adjacent lower
network is always above the higher one (see also the
Supporting Information, Figure S11).
Received: December 19, 2008
Published online: March 12, 2009
Keywords: Borromean rings · crystal engineering ·
.
hydrogen bonds · self-assembly · topology
[1] a) G. R. Desiraju, Angew. Chem. 1995, 107, 2541 – 2558; Angew.
Chem. Int. Ed. Engl. 1995, 34, 2311 – 2327; b) Making Crystals by
Design: Methods, Techniques and Applications (Eds.: D. Braga,
F. Greponi), Wiley-VCH, Weinheim, 2007; c) Frontiers in Crystal
Engineering (Eds.: E. R. T. Tiekink, J. J. Vittal), Wiley, Chi-
chester, 2006; d) Crystal Design: Structure and Function Per-
spectives in Supramolecular Chemistry, Vol. 7 (Ed.: G. R. Desir-
aju), Wiley, Chichester, 2003.
In the above four structures, TCA molecules always form
a six-fold phenyl embrace with each other, whereas such a
dimer was not found in [2TCA·azpy], although the hcb nets of
[2TCA·azpy] are similar to those of the above four structures
[2] a) M. W. Hosseini, Acc. Chem. Res. 2005, 38, 313 – 323; b) J. D.
Wuest, Chem. Commun. 2005, 5830 – 5837.
[3] J. D. Dunitz, Chem. Commun. 2003, 545 – 548.
[4] T. Beyer, T. Lewis, S. L. Price, CrystEngComm 2001, 3, 178 – 213.
Angew. Chem. Int. Ed. 2009, 48, 2873 –2876
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2875

Communications
¯
[5] S. Lee, A. B. Mallik, Z. Xu, E. B. Lobkovsky, L. Tran, Acc.
[17] Crystal data for TCA: triclinic, P1, a = 17.1909(3), b =
17.5114(3), c = 17.5337(3) ꢁ, a = 60.1413(4), b = 77.3032(6),
g = 84.4332(3)8, V= 4465.21(13) ꢁ3, Z = 6, T= 173 K, F(000) =
Chem. Res. 2005, 38, 251 – 261.
[6] a) G. R. Desiraju, Crystal Engineering. The Design of Organic
Solids, Elsevier, Amsterdam, 1989; b) G. R. Desiraju, Angew.
Chem. 2007, 119, 8492 – 8508; Angew. Chem. Int. Ed. 2007, 46,
8342 – 8356.
[7] O. M. Yaghi, M. OꢂKeefe, N. W. Ockwig, H. K. Chae, M.
Eddaoudi, J. Kim, Nature 2003, 423, 705 – 714.
1572,
1calcd = 1.108 gcmꢀ3
.
m(MoKa) = 0.076 mmꢀ1
(l =
0.71069 ꢁ), 30154 measured reflections, 16882 independent
reflections (Rint = 0.0341), 1000 refined parameters, R1 = 0.0691,
wR2 = 0.1868 for 8941 reflections with I > 2s(I), GOF = 1.030.
The diffuse electron density due to the disordered and uniden-
tified moieties was treated with SQUEEZE routine within the
PLATON software package (A. L. Spek, Acta Crystallogr. Sect.
A 1990, 46, C34). The void volume is listed at the end of each of
[8] R. Robson, Dalton Trans. 2008, 5113 – 5131.
[9] K. S. Chichak, S. J. Cantrill, A. R. Pease, S. H. Chiu, G. W. V.
Cave, J. L. Atwood, J. F. Stoddart, Science 2004, 304, 1308 – 1312.
[10] a) L. Carlucci, G. Ciani, D. M. Proserpio, CrystEngComm 2003,
5, 269 – 279; b) L. Carlucci, G. Ciani, D. M. Proserpio, Coord.
Chem. Rev. 2003, 246, 247 – 289; c) I. A. Baburin, V. A. Blatov, L.
Carlucci, G. Ciani, D. M. Proserpio, CrystEngComm 2008, 10,
1822 – 1838.
[11] a) M.-L. Tong, X.-M. Chen, B.-H. Ye, L.-N. Ji, Angew. Chem.
1999, 111, 2376 – 2379; Angew. Chem. Int. Ed. 1999, 38, 2237 –
2240; b) D. B. Leznoff, B.-Y. Xue, R. J. Batchelor, F. W. B.
Einstein, B. O. Patrick, Inorg. Chem. 2001, 40, 6026 – 6034;
c) S. Muthu, J. H. K. Yip, J. J. Vitta, J. Chem. Soc. Dalton Trans.
2002, 4561 – 4568; d) M. P. Suh, H. J. Choi, S. M. So, B. M. Kim,
Inorg. Chem. 2003, 42, 676 – 678; e) R. Liantonio, P. Metrangolo,
T. Pilati, G. Resnati, Cryst. Growth Des. 2003, 3, 355 – 361; f) L.
¯
the crystal information files. [TCA·methanol]: trigonal, P3, a =
17.598(3), c = 15.334(3) ꢁ, V= 4112.8(12) ꢁ3, Z = 6, T= 173 K,
F(000) = 1680, 1calcd = 1.280 gcmꢀ3. m(MoKa) = 0.090 mmꢀ1 (l =
0.71073 ꢁ), 9617 measured reflections, 4838 independent reflec-
tions (Rint = 0.0284), 353 refined parameters, R1 = 0.0856, wR2 =
0.2058 for 3253 reflections with I > 2s(I), GOF = 1.001. The
diffuse electron density due to the disordered and unidentified
moieties was treated with SQUEEZE routine within the
¯
PLATON software package. [2TCA·3phenazine]: triclinic, P1,
a = 9.805(2), b = 13.971(3), c = 14.357(3) ꢁ, a = 75.84(3), b =
84.43(3), g = 84.14(3)8, V= 1891.8(7) ꢁ3, Z = 2, T= 173 K, F-
(000) = 806, 1calcd = 1.346 gcmꢀ3
.
m(MoKa) = 0.089 mmꢀ1 (l =
0.71073 ꢁ), 12451 measured reflections, 6660 independent
reflections (Rint = 0.0262), 524 refined parameters, R1 = 0.0778,
wR2 = 0.1861 for 5337 reflections with I > 2s(I), GOF = 1.011.
´
Dobrzanska, H. G. Raubenheimer, L. J. Barbour, Chem.
Commun. 2005, 5050 – 5052; g) R. Liantonio, P. Metrangolo, F.
Meyer, T. Pilati, W. Navarrini, G. Resnati, Chem. Commun.
2006, 1819 – 1821; h) X.-Q. Lu, M. Pan, J.-R. He, Y.-P. Cai, B.-S.
Kang, C.-Y. Su, CrystEngComm 2006, 8, 827 – 829; i) X.-L.
Zhang, C.-P. Guo, Q.-Y. Yang, W. Wang, W.-S. Liu, B.-S. Kang,
C.-Y. Su, Chem. Commun. 2007, 4242 – 4244; j) X.-L. Zhang, C.-
P. Guo, Q.-Y. Yang, T.-B. Lu, Y.-X. Tong, C.-Y. Su, Chem. Mater.
2007, 19, 4630 – 4632; k) J. Li, L. Song, S. Du, Inorg. Chem.
Commun. 2007, 10, 358 – 361; l) P. Byrne, G. O. Lloyd, N. Clarke,
J. W. Steed, Angew. Chem. 2008, 120, 5845 – 5848; Angew. Chem.
Int. Ed. 2008, 47, 5761 – 5764.
¯
[2TCA·bipy]: triclinic, P1, a = 10.4648(6), b = 11.5920(8), c =
13.3415(11) ꢁ, a = 88.298(6), b = 85.013(5), g = 78.412(4)8, V=
1579.32(19) ꢁ3, Z = 2, T= 113 K, F(000) = 606, 1calcd
=
1.208 gcmꢀ3
.
m(MoKa) = 0.082 mmꢀ1 (l = 0.71070 ꢁ), 19846
measured reflections, 7477 independent reflections (Rint
=
0.0270), 392 refined parameters, R1 = 0.0481, wR2 = 0.149 for
5461 reflections with I > 2s(I), GOF = 1.098. The diffuse elec-
tron density due to the disordered and unidentified moieties was
treated with SQUEEZE routine within the PLATON software
package. [2TCA·azpy]: monoclinic, P2/c, a = 22.228(4), b =
11.809(2), c = 12.144(2) ꢁ, b = 105.17(3)8, V= 3076.8(11) ꢁ3,
Z = 4, T= 173 K, F(000) = 1240, 1calcd = 1.271 gcmꢀ3. m(MoKa) =
0.087 mmꢀ1 (l = 0.71069 ꢁ), 9918 measured reflections, 5401
independent reflections (Rint = 0.0378), 397 refined parameters,
R1 = 0.0836, wR2 = 0.1909 for 4391 reflections with I > 2s(I),
GOF = 1.011. CCDC 713783, 713784, 713785, 713786 and 713787
contain the supplementary crystallographic data for this paper.
These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
[12] D. J. Duchamp, R. E. Marsh, Acta Crystallogr. Sect. B 1969, 25,
5 – 19.
[13] a) S. H. Dale, M. R. J. Elsegood, S. J. Richards, Chem. Commun.
2004, 1278 – 1279; b) B. R. Bhogala, P. Vishweshwar, A. Nangia,
Cryst. Growth Des. 2002, 2, 325 – 328.
[14] a) C. V. K. Sharma, M. J. Zaworotko, Chem. Commun. 1996,
2655 – 2656; b) T. R. Shattock, P. Vishweshwar, Z. Wang, M. J.
Zaworotko, Cryst. Growth Des. 2005, 5, 2046 – 2049.
[15] H. L. Frisch, E. Wassermann, J. Am. Chem. Soc. 1961, 83, 3789 –
3795.
[18] a) B. Moulton, M. J. Zaworotko, Chem. Rev. 2001, 101, 1629 –
1658; b) M. J. Zaworotko, Chem. Commun. 2001, 1 – 9.
[19] I. Dance, M. Scudder, Chem. Eur. J. 1996, 2, 481 – 486.
[16] D. N. Chin, D. M. Gordon, G. M. Whitesides, J. Am. Chem. Soc.
1994, 116, 12033 – 12044.
2876 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 2873 –2876