Molecular tectonics. Selective exchange of cations in porous anionic hydrogen-bonded networks built from derivatives of tetraphenylborate

Article abstract of DOI:10.1021/ja042233m

Tetraphenylmethanes with multiple hydrogen-bonding sites are known to associate to form robust porous supramolecular networks. Analogous anionic networks can be built from the corresponding tetraphenylborates. Crystallization of the tetraphenylphosphonium salt of tetraphenylborate 2 produces an anionic network in which 74% of the volume is available for including cations and neutral guests. Other salts of anion 2 with diverse cations crystallize consistently to form the same network, whereas a neutral analogue of anion 2, tetraphenylmethane 1, produces an uncharged network that is far less open. Cations can be exchanged in single crystals of salts of tetraphenylborate 2 with retention of crystallinity and with selectivities similar to those observed in typical zeolites. Together, these observations provide new strategies for making ordered molecular materials by design, and they reveal that constructing such materials from charged subunits offers special advantages.

Full text of DOI:10.1021/ja042233m

Published on Web 04/02/2005
Molecular Tectonics. Selective Exchange of Cations in Porous
Anionic Hydrogen-Bonded Networks Built from Derivatives of
Tetraphenylborate
Nadia Malek, Thierry Maris, Michel Simard, and James D. Wuest*
Contribution from the De´partement de Chimie, UniVersite´ de Montre´al,
Montre´al, Que´bec H3C 3J7 Canada
Received December 24, 2004; E-mail: james.d.wuest@umontreal.ca
Abstract: Tetraphenylmethanes with multiple hydrogen-bonding sites are known to associate to form robust
porous supramolecular networks. Analogous anionic networks can be built from the corresponding
tetraphenylborates. Crystallization of the tetraphenylphosphonium salt of tetraphenylborate 2 produces an
anionic network in which 74% of the volume is available for including cations and neutral guests. Other
salts of anion 2 with diverse cations crystallize consistently to form the same network, whereas a neutral
analogue of anion 2, tetraphenylmethane 1, produces an uncharged network that is far less open. Cations
can be exchanged in single crystals of salts of tetraphenylborate 2 with retention of crystallinity and with
selectivities similar to those observed in typical zeolites. Together, these observations provide new strategies
for making ordered molecular materials by design, and they reveal that constructing such materials from
charged subunits offers special advantages.
Introduction
When tectons can form multiple interactions that are strong
and directional, such as hydrogen bonds, crystallization normally
Learning how to predict the structures and properties of
molecular crystals is a major challenge in modern science.1 This
challenge is creating a surge of excitement in the field of crystal
engineering,2 an area of rich opportunity where much new
knowledge awaits discovery and exploitation. Reaching the goal
of making crystalline molecular materials by design will require
sustained effort and creative use of tools drawn from many
separate disciplines, including crystallography, statistics, theory,
and synthesis.
generates open networks.6-15 Volume unfilled by the networks
themselves is then occupied by included guests or by indepen-
dent interpenetrating networks.16,17 Systematic studies of deriva-
tives of tetraphenylmethane have provided particularly strong
support for these generalizations.9-11 For example, crystalliza-
tion of tetraphenylmethane 1 from HCOOH/dioxane is directed
In this effort, systematic comparisons of how related mol-
ecules crystallize will continue to be a particularly valuable
source of deeper understanding. From such comparisons has
emerged the notion that certain functional groups in molecules
can play a dominant role in crystallization, by taking part in
attractive intermolecular interactions according to reliable
patterns. When multiple sticky groups are incorporated in
properly designed molecules, their interactions can be considered
to determine the resulting structures. Molecules of this type have
been called tectons, from the Greek word for builder,3 and the
term molecular tectonics has been used to refer to the art and
science of constructing ordered molecular materials from
tectonic subunits.4,5
by hydrogen bonding of its diaminotriazine groups according
to established motifs,6-9 thereby producing a robust three-
(6) Laliberte´, D.; Maris, T.; Wuest, J. D. J. Org. Chem. 2004, 69, 1776.
(7) Laliberte´, D.; Maris, T.; Sirois, A.; Wuest, J. D. Org. Lett. 2003, 5, 4787.
(8) Sauriat-Dorizon, H.; Maris, T.; Wuest, J. D.; Enright, G. D. J. Org. Chem.
2003, 68, 240.
(9) Brunet, P.; Simard, M.; Wuest, J. D. J. Am. Chem. Soc. 1997, 119, 2737.
(10) Fournier, J.-H.; Maris, T.; Wuest, J. D. J. Org. Chem. 2004, 69, 1762.
(11) Laliberte´, D.; Maris, T.; Wuest, J. D. Can. J. Chem. 2004, 82, 386. Saied,
O.; Maris, T.; Wuest, J. D. J. Am. Chem. Soc. 2003, 125, 14956. Fournier,
J.-H.; Maris, T.; Wuest, J. D.; Guo, W.; Galoppini, E. J. Am. Chem. Soc.
2003, 125, 1002. Brunet, P.; Demers, E.; Maris, T.; Enright, G. D.; Wuest,
J. D. Angew. Chem., Int. Ed. 2003, 42, 5303. Le Fur, E.; Demers, E.; Maris,
T.; Wuest, J. D. Chem. Commun. 2003, 2966. Fournier, J.-H.; Maris, T.;
Simard, M.; Wuest, J. D. Cryst. Growth Des. 2003, 3, 535.
(1) Dunitz, J. D. Chem. Commun. 2003, 545. Desiraju, G. R. Nature Materials
2002, 1, 77. Gavezzotti, A. Acc. Chem. Res. 1994, 27, 309. Maddox, J.
Nature 1988, 335, 201.
(2) Braga, D. Chem. Commun. 2003, 2751. Biradha, K. CrystEngComm 2003,
5, 374. Hollingsworth, M. D. Science 2002, 295, 2410. Crystal Engineering:
From Molecules and Crystals to Materials; Braga, D., Grepioni, F., Orpen,
A. G., Eds.; Kluwer: Dordrecht, Netherlands, 1999. Desiraju, G. R. Crystal
Engineering: The Design of Organic Solids; Elsevier: Amsterdam, 1989.
(3) Simard, M.; Su, D.; Wuest, J. D. J. Am. Chem. Soc. 1991, 113, 4696.
(4) Mann, S. Nature 1993, 365, 499.
(12) For reviews, see: Steiner, T. Angew. Chem., Int. Ed. 2002, 41, 48. Prins,
L. J.; Reinhoudt, D. N.; Timmerman, P. Angew. Chem., Int. Ed. 2001, 40,
2382.
(5) Wang, X.; Simard, M.; Wuest, J. D. J. Am. Chem. Soc. 1994, 116, 12119.
9
5910
J. AM. CHEM. SOC. 2005, 127, 5910-5916
10.1021/ja042233m CCC: $30.25 © 2005 American Chemical Society

Molecular Tectonics
A R T I C L E S
dimensional network in which each tecton forms a total of 16
hydrogen bonds with eight neighbors. About 45% of the volume
is available for including guests, which lie in parallel channels
and can be exchanged in single crystals or even partly removed
without loss of crystallinity.
In many ways, such porous molecular networks are strikingly
similar to zeolites, and they may find analogous applications
in separation, catalysis, and other areas of technology. However,
zeolitic networks are anionic, whereas porous molecular net-
works that have been studied so far have rarely carried a net
charge.15,18 We have now synthesized tetraphenylborate 2, which
is a charged analogue of tetraphenylmethane 1, and we have
(13) For other recent studies of crystal engineering using molecules designed
to associate by hydrogen bonding, see: Uemura, K.; Kitagawa, S.; Fukui,
K.; Saito, K. J. Am. Chem. Soc. 2004, 126, 3817. Sada, K.; Inoue, K.;
Tanaka, T.; Tanaka, A.; Epergyes, A.; Nagahama, S.; Matsumoto, A.;
Miyata, M. J. Am. Chem. Soc. 2004, 126, 1764. Maspoch, D.; Domingo,
N.; Ruiz-Molina, D.; Wurst, K.; Tejada, J.; Rovira, C.; Veciana, J. J. Am.
Chem. Soc. 2004, 126, 730. Baudron, S. A.; Avarvari, N.; Canadell, E.;
Auban-Senzier, P.; Batail, P. Chem. Eur. J. 2004, 10, 4498. Angeloni, A.;
Crawford, P. C.; Orpen, A. G.; Podesta, T. J.; Shore, B. J. Chem. Eur. J.
2004, 10, 3783. Murata, T.; Morita, Y.; Fukui, K.; Sato, K.; Shiomi, D.;
Takui, T.; Maesato, M.; Yamochi, H.; Saito, G.; Nakasuji, K. Angew. Chem.,
Int. Ed. 2004, 43, 6343. Soldatov, D. V.; Moudrakovski, I. L.; Ripmeester,
J. A. Angew. Chem., Int. Ed. 2004, 43, 6308. Du, Y.; Creighton, C. J.;
Tounge, B. A.; Reitz, A. B. Org. Lett. 2004, 6, 309. Bhogala, B. R.;
Vangala, V. R.; Smith, P. S.; Howard, J. A. K.; Desiraju, G. R. Cryst.
Growth Des. 2004, 4, 647. Balasubramaniam, V.; Wong, L. G. J.; Vittal,
J. J.; Valiyaveettil, S. Cryst. Growth Des. 2004, 4, 553. Dale, S. H.;
Elsegood, M. R. J.; Coombs, A. E. L. CrystEngComm 2004, 6, 328. Vinodu,
M.; Goldberg, I. CrystEngComm 2004, 6, 215. Aakero¨y, C. B.; Desper, J.;
Helfrich, B. A. CrystEngComm 2004, 6, 19. Alshahateet, S. F.; Nakano,
K.; Bishop, R.; Craig, D. C.; Harris, K. D. M.; Scudder, M. L. CrystEng-
Comm 2004, 6, 5. Brown, R. J.; Camarasa, G.; Griffiths, J.-P.; Day, P.;
Wallis, J. D. Tetrahedron Lett. 2004, 45, 5103. Pedireddi, V. R.;
SeethaLekshmi, N. Tetrahedron Lett. 2004, 45, 1903. Martin, S. M.;
Yonezawa, J.; Horner, M. J.; Macosko, C. W.; Ward, M. D. Chem. Mater.
2004, 16, 3045. Kumar, D. K.; Jose, D. A.; Dastidar, P.; Das, A. Chem.
Mater. 2004, 16, 2332. Barton, O. G.; Schmidtmann, M.; Mu¨ller, A.;
Mattay, J. New J. Chem. 2004, 28, 1335. Shimizu, L. S.; Hughes, A. D.;
Smith, M. D.; Davis, M. J.; Zhang, B. P.; zur Loye, H.-C.; Shimizu, K. D.
J. Am. Chem. Soc. 2003, 125, 14972. Ouyang, X.; Fowler, F. W.; Lauher,
J. W. J. Am. Chem. Soc. 2003, 125, 12400. Kobayashi, K.; Sato, A.;
Sakamoto, S.; Yamaguchi, K. J. Am. Chem. Soc. 2003, 125, 3035.
Rakotondradany, F.; Whitehead, M. A.; Lebuis, A.-M.; Sleiman, H. F.
Chem. Eur. J. 2003, 9, 4771. Jagadish, B.; Carducci, M. D.; Bosshard, C.;
Gu¨nter, P.; Margolis, J. I.; Williams, L. J.; Mash, E. A. Cryst. Growth
Des. 2003, 3, 811. Bhogala, B. R.; Nangia, A. Cryst. Growth Des. 2003,
3, 547. Olenik, B.; Smolka, T.; Boese, R.; Sustmann, R. Cryst. Growth
Des. 2003, 3, 183. Kitamura, M.; Nakano, K. Cryst. Growth Des. 2003, 3,
25. Ma, B.-Q.; Coppens, P. Chem. Commun. 2003, 2290. Yagai, S.; Karatsu,
T.; Kitamura, A. Chem. Commun. 2003, 1844. Walsh, R. D. B.; Bradner,
M. W.; Fleischman, S.; Morales, L. A.; Moulton, B.; Rodr´ıguez-Hornedo,
R.; Zaworotko, M. J. Chem. Commun. 2003, 186. Blay, G.; Ferna´ndez, I.;
Pedro, J. R.; Ruiz-Garc´ıa, R.; Mun˜oz, M. C.; Cano, J.; Carrasco, R. Eur.
J. Org. Chem. 2003, 1627. Braga, D.; Polito, M.; Bracaccini, M.; D’Addario,
D.; Tagliavini, E.; Sturba, L. Organometallics 2003, 22, 2142. Hosseini,
M. W. Coord. Chem. ReV. 2003, 240, 157. Roesky, H. W.; Andruh, M.
Coord. Chem. ReV. 2003, 236, 91. Ermer, O.; Kusch, A.; Ro¨bke, C. HelV.
Chim. Acta 2003, 86, 922. Arora, K. K.; Pedireddi, V. R. J. Org. Chem.
2003, 68, 9177. Wen, J.-F.; Hong, W.; Yuan, K.; Mak, T. C. W.; Wong,
H. N. C. J. Org. Chem. 2003, 68, 8918. Moon, K.; Chen, W.-Z.; Ren, T.;
Kaifer, A. E. CrystEngComm 2003, 5, 451. Burke, N. J.; Burrows, A. D.;
Mahon, M. F.; Pritchard, L. S. CrystEngComm 2003, 5, 355. Pe´rez, C.;
Rodr´ıguez, M. L.; Foces-Foces, C.; Pe´rez-Herna´ndez, N.; Pe´rez, R.; Mart´ın,
J. D. Org. Lett. 2003, 5, 641. Shan, N.; Bond, A. D.; Jones, W. New J.
Chem. 2003, 27, 365. Kobayashi, N.; Naito, T.; Inabe, T. Bull. Chem. Soc.
Jpn. 2003, 76, 1351.
+
Figure 1. View of the structure of crystals of tetraphenylborate 2·PPh4
grown from DMSO/toluene, showing a central tecton (red) surrounded by
its six hydrogen-bonded neighbors (yellow, green, and blue). Hydrogen
bonds appear as broken lines, and all guests are omitted for clarity. The
four diaminotriazine groups of the central tecton form a total of 16 hydrogen
bonds according to all three standard motifs I-III. Two of the neighbors
(yellow) each form two hydrogen bonds with the central tecton according
to motif I, and two other neighbors (green) each form two hydrogen bonds
according to motif III. Two diaminotriazine groups of each of the remaining
two neighbors (blue) interact simultaneously with two arms of the central
tecton according to motif II.
found that its salts crystallize to generate an anionic hydrogen-
bonded network with special properties of porosity and selective
ion exchange. Moreover, the behavior of tecton 2 suggests new
strategies of potentially general value for creating molecular
crystals by design.
Results and Discussion
Synthesis, Crystallization, and Structure of the Tetra-
phenylphosphonium Salt of Tetraphenylborate 2. Salts of
anionic tecton 2 were made in the following way. Monolithiation
of 1,4-diiodobenzene (BuLi, 1 equiv, -10 °C), followed by the
addition of BF3·O(C2H5)2 (0.2 equiv, 25 °C), provided lithium
tetrakis(4-iodophenyl)borate (3) in 85% yield. Subsequent Pd-
catalyzed cyanation (KCN, Pd(OOCCH3)2)19 gave an 85% yield
of potassium tetrakis(4-cyanophenyl)borate (4). Intermediate 4
was then converted in 87% yield into the sodium salt of anionic
tecton 2 by reaction with dicyandiamide under standard condi-
tions,20 and the tetraphenylphosphonium salt and other salts were
prepared in quantitative yield from the sodium salt by cation
exchange.
Colorless single crystals of tecton 2·PPh4+ suitable for X-ray
diffraction could be grown from DMSO/toluene. The crystals
proved to belong to the monoclinic space group P2/n and to
correspond to an inclusion compound of approximate composi-
tion 2·PPh4+·5DMSO.21 Views of the structure appear in Figures
(14) Hosseini, M. W. CrystEngComm 2004, 6, 318.
(15) Kobayashi, N.; Naito, T.; Inabe, T. CrystEngComm 2004, 6, 189. Lam,
C.-K.; Mak, T. C. W. Chem. Commun. 2003, 2660. Ma, B.-Q.; Sun, H.-L.;
Gao, S. Chem. Commun. 2003, 2164. Gale, P. A.; Navakhun, K.; Camiolo,
S.; Light, M. E.; Hursthouse, M. B. J. Am. Chem. Soc. 2002, 124, 11228.
Zhu, H.-F.; Fan, J.; Okamura, T.; Sun, W.-Y.; Ueyama, N. Chem. Lett.
2002, 898. Braga, D.; Maini, L.; Grepioni, F.; Elschenbroich, C.; Paganelli,
F.; Schiemann, O. Organometallics 2001, 20, 1875. Ermer, O.; Lindenberg,
L. Chem. Ber. 1990, 123, 1111.
(16) For discussions of interpenetration in networks, see: Batten, S. R.
CrystEngComm 2001, 18, 1. Batten, S. R.; Robson, R. Angew. Chem., Int.
Ed. 1998, 37, 1460.
(17) An updated list of examples of interpenetration is available on the web
site of Dr. Stuart R. Batten at Monash University (www.chem.mo-
nash.edu.au).
(19) Takagi, K.; Okamoto, T.; Sakakibara, Y.; Ohno, A.; Oka, S.; Hayama, N.
Bull. Chem. Soc. Jpn. 1976, 49, 3177.
(20) Simons, J. K.; Saxton, M. R. Organic Syntheses; Wiley: New York, 1963;
Collect. Vol. IV, p 78.
(21) The composition was estimated by 1H NMR spectroscopy of dissolved
samples. The amount of any H2O included could not be determined
accurately.
(18) For a review of ionic self-assembly, see: Faul, C. F. J.; Antonietti, M.
AdV. Mater. 2003, 15, 673.
9
J. AM. CHEM. SOC. VOL. 127, NO. 16, 2005 5911

A R T I C L E S
Malek et al.
series of neutral compounds MPh4, each molecule has 14
neighbors in van der Waals contact. Each ion in crystals of the
salt BPh4-PPh4+ also has 14 neighbors, but only four have the
same charge, thereby confirming the importance of ionic
repulsion as a structural determinant in molecular crystal
engineering. For similar reasons, neutral tectons and their
charged analogues are unlikely to crystallize isostructurally;
however, the disadvantage of losing a degree of structural
predictability is offset by gaining significantly higher porosity,
at least in the case of tecton 2.
As shown in Figure 1, the four diaminotriazine groups of
each anionic tecton 2 interact with eight diaminotriazine groups
provided by six neighboring tectons according to all three normal
hydrogen-bonding motifs I-III.6-9 Figure 5 reveals that the
Figure 2. Representation of the structure of crystals of anionic tecton
2·PPh4+ grown from DMSO/toluene showing the three-dimensional 6-con-
nected network defined by joining the central boron atom of each tecton
with the centers of the set of six neighbors shown in Figure 1. For
clarification, one boron center and its six connections are highlighted in
blue.
1-4. Anionic tecton 2 and its neutral analogue 1 have the same
hydrogen-bonding sites, oriented tetrahedrally by cores with
nearly identical molecular structures, so the emerging rules of
molecular tectonics suggest that the resulting networks should
share key features and perhaps even be isostructural. In fact,
the networks show extensive homology, and both are porous,
three-dimensional, and connected by multiple hydrogen bonds,
as planned. Moreover, each tecton in both networks partici-
pates in a total of 16 hydrogen bonds, and in neither net-
work is interpenetration observed. It is noteworthy that ionic
repulsion and the need to incorporate counterions do not con-
spire to prevent tecton 2 from self-associating in the normal
manner to form a porous hydrogen-bonded three-dimensional
network similar in many ways to the one formed by neutral
analogue 1.
four diaminotriazine groups of each neutral tecton 1 form
hydrogen bonds of type II only, using eight diaminotriazine
groups provided by eight neighbors. Lines joining the central
boron atom of each anionic tecton 2 with the centers of its
hydrogen-bonded neighbors define the complex three-dimen-
sional 6-connected network shown in Figure 2. In crystals of
salt 2·PPh4+, the C-B-C angles at the core of each anionic
tecton vary from 100° to 113°, whereas the central C-C-C
angles in crystals of neutral analogue 1 lie in a narrower range
(107°-111°). This shows that deviations from an ideal tetra-
hedral geometry are more pronounced in the network derived
from tetraphenylborate 2, possibly because it is inherently more
difficult to optimize hydrogen bonding, ionic interactions, and
molecular packing at the same.
Despite these fundamental similarities, however, the archi-
tectures of the two networks differ in important details. In
particular, each tecton 1 in the neutral network has eight
hydrogen-bonded neighbors (Figure 5), whereas each tecton 2
in the anionic network has only six (Figure 1). The C‚‚‚C
distance between the central carbon atom of each neutral tecton
1 and those of its hydrogen-bonded neighbors is 14.91 Å,
whereas the corresponding average B‚‚‚B distance in crystals
The anionic network produced by crystallization of tecton
+
of anionic tecton 2·PPh4 is 15.64 Å. Reducing the number of
+
2·PPh4 is exceptionally porous. Only 26% of the volume of
neighbors and increasing their separation, as observed in crystals
of tecton 2, presumably decreases overall ionic repulsion and
creates space for the inclusion of counterions. This suggests
that modifying tectons by introducing charge may be a generally
effective strategy for making molecular networks less compact
and more porous.
Even when the charge/volume ratio of ions is small, as it is
in the salt 2·PPh4+, the structural consequences of incorporating
charge may be profound. Indeed, crystals of the closely related
salt BPh4-PPh4+, which lacks hydrogen-bonding sites, are not
isostructural with those of neutral analogues MPh4, where M
) C, Si, Ge, Sn, Pb, and Os.22 In crystals of the isostructural
the crystals is occupied by the ordered tectons themselves,23,24
and the remaining 74% is available for including cations and
guests, which are disordered. The observed porosity is greater
than that of all other networks built from small molecules, with
(23) The percentage of volume accessible to cations and guests was estimated
by the PLATON program.24 PLATON calculates the accessible volume
by allowing a spherical probe of variable radius to roll over the internal
van der Waals surface of the crystal structure. PLATON uses a default
value of 1.20 Å for the radius of the probe, which is an appropriate model
for small guests such as water. The van der Waals radii used to define
surfaces for these calculations are as follows: C: 1.70 Å, H: 1.20 Å, N:
1.55 Å, and B: 1.63 Å. If V is the volume of the unit cell and Vg is the
guest-accessible volume as calculated by PLATON, then the porosity P in
% is given by 100Vg/V.
(24) Spek, A. L. PLATON, A Multipurpose Crystallographic Tool; Utrecht
University: Utrecht, The Netherlands, 2001. van der Sluis, P.; Spek, A. L.
Acta Crystallogr. 1990, A46, 194.
(22) Lloyd, M. A.; Brock, C. P. Acta Crystallogr. 1997, B53, 773.
9
5912 J. AM. CHEM. SOC. VOL. 127, NO. 16, 2005

Molecular Tectonics
A R T I C L E S
Figure 3. Porosity of the network constructed from anionic tecton 2·PPh4+, viewed along the a axis (Figure 3a), the c axis (Figure 3b), the ab diagonal
(Figure 3c), and the ac diagonal (Figure 3d). All views show a 3 × 2 × 4 array of unit cells, with guests omitted and atoms represented by spheres of van
der Waals radii to reveal the cross sections of the channels. Atoms of hydrogen appear in light gray, atoms of boron in yellow, atoms of carbon in dark gray,
and atoms of nitrogen in blue.
Figure 4. Stereoscopic representation of interconnected channels within the network constructed from anionic tecton 2·PPh4+. The image shows a 3 × 3
× 2 array of unit cells viewed along the c axis. The outsides of the channels appear in light gray, and dark gray is used to show where the channels are cut
by the boundaries of the array. The surface of the channels is defined by the possible loci of the center of a sphere of diameter 5 Å as it rolls over the surface
of the ordered network.27
only a single exception.10 Moreover, the porosity exceeds the
percentage of volume occupied by water and other guests in
all but the most highly hydrated crystals of proteins,25 even
though tecton 2 has a simple shape that can in principle be
packed efficiently, whereas proteins have complex topologies
that disfavor close packing.
Cations and guests included in the anionic network built from
tecton 2 occupy large interconnected channels (Figures 3 and
(25) Rangarajan, S. E.; Tocilj, A.; Li, Y.; Iannuzzi, P.; Matte, A.; Cygler, M.
Acta Crystallogr. 2003, D59, 2348. Quillen, M. L.; Matthews, B. W. Acta
Crystallogr. 2000, D56, 791. Andersson, K. M.; Hovmo¨ller, S. Acta
Crystallogr. 2000, D56, 789.
9
J. AM. CHEM. SOC. VOL. 127, NO. 16, 2005 5913

A R T I C L E S
Malek et al.
Scheme 1
Anionic tecton 2 provides a special opportunity to test these
hypotheses. Many tectons that have been studied so far are polar
and rigid molecules, with low solubility in most solvents. As a
result, conditions for crystallization can normally be varied only
within narrow limits. In attempts to crystallize tecton 1, for
example, HCOOH can be replaced by similar polar solvents,
or precipitants other than dioxane can be used. For these reasons,
the notion that tectons resist polymorphism cannot normally be
tested rigorously by crystallizing them under widely different
conditions. In contrast, the hypothesis that anionic tecton 2
favors a particular network can easily be subjected to a demand-
ing test, simply by crystallizing salts with many different cations.
Inducing salts of tecton 2 to form single crystals suitable for
X-ray diffraction was difficult, but eventually we obtained
structural data for salts with the diverse group of mono- and
dications shown in Scheme 1. In all seven cases, crystallization
from DMSO/toluene or DMF/toluene produces the same anionic
network, and the unit cell parameters are closely similar (6%)
to those of the PPh4+ salt. Again, the included cations and guests
are disordered. We were also able to crystallize the Li+, Na+,
K+, and [CH3(CH2)3]4N+ salts of tecton 2 from DMSO/toluene.
In this group of salts, the unit cells appear to be different from
those of salts of the cations shown in Scheme 1; however, the
quality of the crystals was poor, and we could not collect full
sets of data suitable for complete determination of the structures.
Salts of tetraphenylborate 2 with different cations are dis-
tinctly different compounds. Moreover, the size and structural
complexity of some of the cations in Scheme 1 rival those of
anionic tecton 2 itself. For these reasons, the consistent prefer-
ence for the formation of the same hydrogen-bonded network
is a significant observation. It validates the notion that tecton 2
and other compounds with oriented arrays of sticky sites are
programmed to form specific molecular networks and to disfavor
alternatives, in agreement with the emerging principles of
molecular tectonics.14
Figure 5. View of the structure of crystals of tetraphenylmethane 1 grown
from HCOOH/dioxane, showing a central tecton (light gray) surrounded
by four hydrogen-bonded neighbors (dark gray).9 Four other symmetry-
equivalent neighbors are omitted for clarity. Hydrogen bonds appear as
broken lines, and guests are not shown. The four diaminotriazine groups
of the central tecton form a total of 16 hydrogen bonds, all of type II.
4). The most impressive channels are aligned with the c axis
and have triangular cross sections measuring approximately 8
× 12 Å2 at the narrowest points (Figure 3b).26 The channels
themselves are represented by the surface shown in Figure 4.27
Those in crystals of neutral tecton 1 are separate and parallel,
but those in crystals of anionic analogue 2 are highly intercon-
nected. In principle, guests that diffuse inside crystals of tecton
2 can reach any point within the channels by many redundant
paths, which should facilitate diffusion when channels are
blocked or other defects are present.28
Crystallization and Structure of Other Salts of Tetra-
phenylborate 2. Tectons 1 and 2 incorporate multiple sticky
sites that have reliable patterns of association, oriented tetra-
hedrally by attachment to rigid cores with well-defined struc-
tures. Such compounds are valuable tools for probing the
generality of polymorphism in molecular crystals and for testing
the widespread notion that the number of polymorphic forms
of any compound is limited only by the effort needed to find
them.29,30 In contrast, the particular structural features of tectons
are designed to disfavor crystallization in many polymorphic
forms; instead, tectons are programmed to form one particular
network, or a restricted number of related alternatives, even
under markedly different conditions of crystallization.
(26) The dimensions of a channel in a particular direction correspond to
those of an imaginary cylinder that could be passed through the hypo-
thetical open network in the given direction in contact with the van der
Waals surface. Such values are inherently conservative because (1) they
measure the cross section at the narrowest constriction and (2) they
systematically underestimate the sizes of channels that are not uniform and
linear.
(27) Representations of channels were generated by the Cavities option in the
program ATOMS (ATOMS, Version 5.1; Shape Software: 521 Hidden
Valley Road, Kingsport, TN 37663; www.shapesoftware.com).
(28) For a discussion of pathways for diffusion in microporous materials, see:
Venuto, P. B. Microporous Mater. 1994, 2, 297.
(29) Bernstein, J. Polymorphism in Molecular Crystals; Oxford University
Press: New York, 2002.
Exchange of Ions in Single Crystals of Salts of Tetra-
phenylborate 2. The strategy of using ionic tectons to build
porous charged networks offers unique advantages and may help
solve long-standing problems in the chemistry of inclusion
compounds. For example, including a specific desired guest in
a neutral network is often difficult, but networks and guests
can be made to have stronger mutual affinity by giving them
opposite charges. The diversity of the cations shown in Scheme
1 is consistent with the notion that ionic attraction provides an
(30) McCrone, W. C. In Physics and Chemistry of the Organic Solid State;
Fox, D., Labes, M. M., Weissberger, A., Eds.; Wiley-Interscience: New
York, 1965; Vol. 2, pp 725-767.
9
5914 J. AM. CHEM. SOC. VOL. 127, NO. 16, 2005

Molecular Tectonics
A R T I C L E S
additional driving force for inclusion in molecular crystals,
thereby ensuring that guests with an impressively wide range
of sizes and shapes can be induced to enter, at least within
geometric limits imposed by the channels.
In crystalline salts of anionic tecton 2, the cations and guests
are disordered, potentially mobile, and located in large intercon-
nected channels that in principle provide multiple paths of
escape. Furthermore, the structure is maintained by an extensive
network of hydrogen bonds, with each tecton held in place by
16 hydrogen bonds. Although the resulting network is robust,
it is not strong enough to resist decomposition when crystals
are removed from their mother liquors for extended periods or
when volatile guests are removed under vacuum. Nevertheless,
single crystals show high structural integrity, and they undergo
exchange of cations without loss of crystallinity. For example,
red single crystals of the salt of tecton 2 with phenazinium cation
8, with approximate dimensions 0.5 mm × 0.5 mm × 0.5 mm,
were exposed to a solution of excess PPh4+Br- in DMSO/
toluene for 24 h at 25 °C. Complete exchange of cations (g98%)
Figure 6. Selective exchange of tetraalkylammonium cations [CH3-
(CH2)n]4N+ in single crystals of tetraphenylborate 2, showing relative
preferences as a function of chain length.
by exposing single crystals of the (R)-R-methylbenzylammo-
nium salt of tecton 2 to solutions in DMF/toluene containing
excess amounts of competing pairs of tetraalkylammonium
bromides, present in equal amounts. The following relative
preferences for [CH3(CH2)n]4N+ were observed at equilibrium
at 25 °C: 0.18, 0.75, 1.0, 0.64, and 0.20 for n ) 1-5,
respectively (Figure 6). Similar degrees of selectivity have been
observed for the inclusion of alkylammonium cations in natural
zeolites and synthetic analogues, including chabazite, faujasite,
and Linde A.33 However, the largest cations accommodated in
the anionic network derived from tecton 2 are much too big to
enter classical natural and synthetic zeolites. These observations
underscore the high degree of porosity that can be attained in
molecular networks, and they show that the performance of such
molecular materials can equal or even surpass that of class-
ical inorganic analogues. Preference for the inclusion of
[CH3(CH2)3]4N+ in networks built from tecton 2 appears to
reflect various competing factors, including the size of the
cation, the interaction of its substituents with the walls of the
channels, the ability of the cation to approach the anionic boron
centers of the network, and the entropically favorable effect of
replacing a small cation with a large one, which presumably
frees neutral guests from inside the network and places them
in a more disordered state in the surrounding solution.
occurred under these conditions to give single crystals of the
+ 31
colorless salt 2·PPh4
.
The recovered crystals (1) remained
transparent and morphologically unchanged, (2) continued to
diffract and to exhibit uniform extinction between crossed
polarizers, and (3) showed unit cell parameters similar to those
of the starting salt ((2%). Similar single-crystal to single-crystal
exchanges could be carried out with other pairs of cations
selected from the group in Scheme 1, as well as with many
other cations that do not exceed geometric limits imposed by
the channels.
Of particular interest are exchanges that replace a larger cation
by a smaller cation. Such exchanges liberate space within the
crystals that can then be occupied by additional neutral guests.
Reducing the volume of the cation allows the network to use
more of its high porosity (74%) to include neutral guests, with
less space needed for the counterions that maintain overall
electroneutrality. For this reason, it is proper to claim that ionic
tecton 2 forms one of the most porous molecular networks
observed so far, even though counterions are a necessary
component of the structure and must occupy part of the available
volume. It is also noteworthy that the exchange of cations is a
predictable source of single crystals of new salts of tecton 2
with the specific structure shown in Figures 1-4, even when
direct crystallization of the new salt is impossible or favors a
different structure. For example, single crystals of the
[CH3(CH2)3]4N+ salt can be prepared by exchange, starting from
single crystals of the salts shown in Scheme 1, but direct
crystallization of the [CH3(CH2)3]4N+ salt from DMSO/toluene
gives crystals that belong to a different space group. Such
examples demonstrate that (1) structural reorganization of the
network is slower than the exchange of cations, which presum-
ably involves simple diffusion, and (2) exchange does not
require substantial movement of the individual tectons, which
is expected to permit recrystallization and lead to the formation
of alternative structures.32
Conclusions
Our study of crystals formed by salts of tetraphenylborate 2
is significant because it subjects the emerging principles of
molecular tectonics to critical tests.14 In addition, the results
provide new insights that can be used to guide future attempts
to engineer crystalline materials with predictable properties. The
simple expedient of replacing carbon by boron in the core of
tetraphenylmethane 1 creates anionic tecton 2 and logically
transforms the robust porous hydrogen-bonded network built
from tecton 1 into an anionic analogue with closely similar
structural features. The resulting charged network is among the
most porous ever observed in molecular crystals, in part because
ionic repulsion disfavors close packing. Crystallization of salts
of tecton 2 with diverse cations consistently yields the same
network, supporting the hypothesis that compounds with
oriented arrays of sticky sites are intrinsically disposed to form
specific molecular networks and to disfavor polymorphs. Cations
can be exchanged in single crystals of salts of tecton 2 with
retention of crystallinity and with selectivities similar to those
observed in typical zeolites. Such exchanges can produce new
As in zeolites, the exchange of cations in the anionic network
built from tecton 2 shows useful selectivity. This was established
(31) The extent of exchange was estimated by 1H NMR spectroscopy of dissolved
samples.
(32) For an analysis of potential mechanisms of ion exchange in porous charged
networks, see: Khlobystov, A. N.; Champness, N. R.; Roberts, C. J.;
Tendler, S. J. B.; Thompson, C.; Schro¨der, M. CrystEngComm 2002, 4,
426.
(33) Breck, D. W. Zeolite Molecular SieVes; Wiley-Interscience: New York,
1974, Chapter 7.
9
J. AM. CHEM. SOC. VOL. 127, NO. 16, 2005 5915

A R T I C L E S
Malek et al.
cation. The product was filtered from the resulting suspension, washed
with deionized H2O, and dried in vacuo.
salts in a predictable crystalline form, even when direct
crystallization proves impossible or favors a different structure.
Together, these observations underscore the potential of mo-
lecular tectonics as a strategy for making new ordered molecular
materials by design, and they reveal that constructing such
materials from charged tectons has special benefits.
Crystallization of Salts of Tecton 2. In a typical procedure, the
salt (∼5 mg) was dissolved in a small amount of DMSO or DMF (∼1
mL), toluene was added dropwise until cloudiness appeared, more
DMSO or DMF was added until the cloudiness disappeared, and the
resulting solution was kept undisturbed at 25 °C. Crystals appeared
within 24 h in the form of complex plates. Alternatively, crystallization
could be effected by diffusion of vapors of toluene into solutions of
salts of tecton 2 in DMSO or DMF. For analysis by 1H NMR
spectroscopy, the crystals were rinsed with pentane and allowed to dry
briefly in a current of air.
Experimental Section
Ether was dried by distillation from the sodium ketyl of benzophe-
none. All other reagents were commercial products that were used
without further purification.
Ion Exchange in Single Crystals of Salts of Tecton 2. Single
crystals of salts of tecton 2 were obtained by the procedures described
above, and the mother liquors were removed by pipet. The crystals
were immediately treated with a saturated solution of the chloride or
bromide salt of a new cation in a 1:1 mixture of DMSO (or DMF) and
toluene, in which the crystals are essentially insoluble. The crystals
were kept in contact with this solution at 25 °C until they were analyzed
Lithium Tetrakis(4-iodophenyl)borate (3). A solution of 1,4-
diiodobenzene (10.0 g, 30.3 mmol) in ether (300 mL) was stirred at
-10 °C under N2 and treated dropwise with a solution of butyllithium
(12.1 mL, 2.5 M in hexane, 30 mmol). The resulting mixture was kept
at -10 °C for 10 min, and BF3·O(C2H5)2 (0.796 g, 5.61 mmol) was
then added. The temperature was allowed to rise to 25 °C, and after
18 h, the mixture was filtered to remove a solid. The solid was washed
with ether (200 mL), and the desired product was extracted with acetone
(20 mL). Acetone was removed from the filtered extracts by evapora-
tion under reduced pressure, leaving a residue of lithium tetrakis(4-
iodophenyl)borate (3) as a colorless solid (3.95 g, 4.76 mmol, 85%).
An analytically pure sample was prepared by recrystallization from
CH2Cl2: mp 289-290 °C; 1H NMR (300 MHz, DMSO-d6) δ 6.87 (m,
1
by X-ray crystallography and H NMR spectroscopy.
X-ray Crystallographic Studies. Data were collected using (1) an
Enraf-Nonius CAD4 diffractometer with Cu KR radiation at 205 K or
(2) a Bruker SMART 2000 diffractometer with Cu KR radiation at
223 K. Structures were solved by direct methods using SHELXS-96
or SHELXS-97 and refined with SHELXL-96 or SHELXL-97.34 All
non-hydrogen atoms were refined anisotropically, whereas hydrogen
atoms were placed in ideal positions and refined as riding atoms.
Structure of the Salt of Tetraphenylborate 2 with PPh4+. Crystals
of the salt belong to the monoclinic space group P2/n with a ) 13.045-
(6) Å, b ) 17.151(8) Å, c ) 24.023(8) Å, â ) 92.21(3)°, V ) 5371(4)
Å3, Dcalcd ) 1.2122 g/cm3, and Z ) 2. Full-matrix least-squares
refinements on F2 led to final residuals Rf ) 0.0546, Rw ) 0.1391, and
GoF ) 0.579 for 5519 reflections with I > 2σ(I).
3
8H), 7.29 (d, 8H, J ) 8.0 Hz); 11B NMR (96.0 MHz, DMSO-d6) δ
-6.35; 13C NMR (100.6 MHz, DMSO-d6) δ 88.5, 134.5, 137.9, 161.1
1
(q, JC-B ) 49 Hz); HRMS (FAB, 3-nitrobenzyl alcohol) calcd for
C24H1611BI4 m/e 822.7524, found 822.7503. Anal. Calcd for
C24H16BI4Li: C, 34.74; H, 1.94. Found: C, 34.80; H, 1.76.
Potassium Tetrakis(4-cyanophenyl)borate (4). A mixture of
lithium tetrakis(4-iodophenyl)borate (3; 4.50 g, 5.42 mmol), KCN (3.90
g, 59.9 mmol), KOH (0.054 g, 0.96 mmol), and Pd(OOCCH3)2 (0.018
g, 0.080 mmol) in tetramethylurea (9 mL) was heated at reflux under
N2. After 18 h, volatiles were removed by evaporation under reduced
pressure; the residual solid was washed with H2O, saturated aqueous
KCl, and CH2Cl2; and the product was then dried in vacuo. This
provided potassium tetrakis(4-cyanophenyl)borate (4) as a beige solid
(2.03 g, 4.43 mmol, 82%). An analytically pure sample was prepared
Structure of the Salt of Tetraphenylborate 2 with Phenazinium
Cation 8. Crystals of the salt belong to the monoclinic space group
P2/n with a ) 12.845(8) Å, b ) 17.134(10) Å, c ) 23.896(16) Å, â
) 92.11(5)°, V ) 5256(6) Å3, Dcalcd ) 1.305 g/cm3, and Z ) 2. Full-
matrix least-squares refinements on F2 led to final residuals Rf ) 0.0480,
Rw ) 0.1084, and GoF ) 0.520 for 5435 reflections with I > 2σ(I).
Structure of the Salt of Tetraphenylborate 2 with Cinchonidin-
ium Cation 9. Crystals of the salt belong to the monoclinic space group
P2/n with a ) 12.6010(13) Å, b ) 18.1778(17) Å, c ) 23.825(2) Å,
by recrystallization from CH3CN: mp >300 °C; IR (KBr) 2222 cm-1
;
1H NMR (400 MHz, DMSO-d6) δ 7.23 (m, 8H), 7.44 (d, 8H, 3J ) 7.8
Hz); 11B NMR (96.0 MHz, DMSO-d6) δ -5.92; 13C NMR (100.6 MHz,
1
â ) 97.436(5)°, V ) 5411.4(9) Å3, Dcalcd ) 0.647 g/cm3 (calculated
without the contribution of included solvent), and Z ) 2. Full-matrix
least-squares refinements on F2 led to final residuals Rf ) 0.0668, Rw
) 0.1703, and GoF ) 0.955 for 6741 reflections with I > 2σ(I).
DMSO-d6) δ 105.6, 120.2, 129.7, 135.6, 167.6 (q, JC-B ) 49 Hz);
HRMS (FAB, 3-nitrobenzyl alcohol) calcd for C28H1611BN4 m/e
419.1468, found 419.1454.
Sodium Salt of Tetrakis[4-[2-(4,6-diamino-1,3,5-triazinyl)]phenyl]-
borate (2). A mixture of potassium tetrakis(4-cyanophenyl)borate (4;
1.50 g, 3.27 mmol), dicyandiamide (1.36 g, 16.2 mmol), and KOH
(0.441 g, 7.86 mmol) in 2-methoxyethanol (9 mL) was heated at reflux
under N2. After 18 h, volatiles were removed by evaporation under
reduced pressure, and the residual solid was washed with CH3OH and
dried. The solid was dissolved in boiling deionized H2O (20 mL), and
saturated aqueous NaCl was added. The resulting suspension was
filtered, and the solid was washed with deionized H2O and dried in
vacuo. This provided the sodium salt of tetrakis[4-[2-(4,6-diamino-
1,3,5-triazinyl)]phenyl]borate (2) as a colorless solid (2.21 g, 2.84 mmol,
87%). A purified sample was prepared by recrystallization from H2O:
Acknowledgment. We are grateful to the Natural Sciences
and Engineering Research Council of Canada, the Ministe`re de
l’EÄ ducation du Que´bec, the Canada Foundation for Innovation,
the Canada Research Chairs Program, and Merck Frosst for
financial support. In addition, acknowledgment is made to the
donors of the Petroleum Research Fund, administered by the
American Chemical Society, for support of this research.
Supporting Information Available: Further crystallographic
details, including ORTEP drawings and tables of crystal-
lographic data, atomic coordinates, anisotropic thermal param-
eters, and bond lengths and angles for salts of tetraphenylborate
2 with PPh4+, phenazinium cation 8, and cinchonidinium cation
9. This material is available free of charge via the Internet at
http://pubs.acs.org.
1
mp >300 °C; H NMR (300 MHz, DMSO-d6) δ 6.56 (bs, 16H), 7.28
(m, 8H), 7.90 (d, 8H, 3J ) 7.7 Hz); 11B NMR (96.0 MHz, DMSO-d6)
δ -5.98; 13C NMR (75.5 MHz, DMSO-d6) δ 125.4, 130.8, 135.1, 167.3
(2C), 171.4.
Other Salts of Tetrakis[4-[2-(4,6-diamino-1,3,5-triazinyl)]phenyl]-
borate (2). The sodium salt of tetrakis[4-[2-(4,6-diamino-1,3,5-triazi-
nyl)]phenyl]borate (2) was dissolved in boiling deionized H2O, and
the resulting solution was treated with a concentrated aqueous solution
containing a large excess of the chloride or bromide salt of the desired
JA042233M
(34) Sheldrick, G. M. SHELXS-97, Program for the Solution of Crystal Structures
and SHELXL-97, Program for the Refinement of Crystal Structures;
Universita¨t Go¨ttingen: Germany, 1997.
9
5916 J. AM. CHEM. SOC. VOL. 127, NO. 16, 2005