Solid-state self-inclusion: The missing link

Article abstract of DOI:10.1002/anie.200601665

Molecular cannibalism: The concept of self-inclusion in the solid state can be subjective. A host-guest system is prepared in which the same compound (2,7-dimethyl-octa-3,5-diyne-2,7-diol) unequivocally plays both roles simultaneously. (Figure Presented).

Full text of DOI:10.1002/anie.200601665

Communications
Host–Guest Systems
DOI: 10.1002/anie.200601665
Solid-State Self-Inclusion: The Missing Link**
Gareth O. Lloyd, Jo Alen, Martin W. Bredenkamp,
Elise J. C. de Vries, Catharine Esterhuysen, and
Leonard J. Barbour*
Cocrystallization and polymorphism are two of the most well-
documented paradigms in the fields of solid-state inclusion
chemistry and crystal engineering. Such concepts are often
open to interpretation, as evidenced by the recent debate
regarding the differences between solvates, pseudopoly-
[
1]
morphs, and cocrystals. Self-inclusion (or, alternatively,
self-hosting)in the solid state has been suggested by several
authors as a distinct paradigm in supramolecular host–guest
[
2]
chemistry. In considering the literature regarding the
concept of self-inclusion, it immediately becomes clear that
this topic is currently highly subjective. However, in view of
the various cases that have been advanced to date as examples
of self-inclusion, it is now possible to identify different levels
of the phenomenon. Thus, to forestall future disagreement, it
is useful to now consolidate the issue in a rational fashion and
to supply the “missing piece of the puzzle”.
In order of increasing constraints on interpretative free-
dom, the levels of solid-state self-inclusion can be listed and
rationalized as follows (note that the term “molecule” is used
loosely and can be substituted by “supramolecule”):
a)Any structure with Z ’ = 1. Each molecule is surrounded by
like molecules and is therefore self-included.
[
3]
b)Any molecular crystal with Z ’ > 1. As with (a), each
molecule is surrounded by other molecules of the same
kind and the designation of a crystallographically unique
[
4]
molecule as either the host or the guest is arbitrary.
c)Interpenetration or weaving of two or more hydrogen-
[
5]
[6]
bonded or coordination frameworks. Each framework
hosts” at least one other and is thus simultaneously both
“
a host and a guest.
d)A molecular crystal in which Z ’ > 1 with the distinction
between host and guest justified according to accepted
[
*] G. O. Lloyd, Dr. M. W. Bredenkamp, Dr. E. J. C. de Vries,
Dr. C. Esterhuysen, Prof. L. J. Barbour
Department of Chemistry
University of Stellenbosch
7602 Matieland (South Africa)
Fax: (+27)21-808-3849
E-mail: ljb@sun.ac.za
J. Alen
Department of Chemistry
Katholieke Universiteit Leuven
Celestijnenlaan 200F, 3001 Heverlee (Belgium)
[
**] We acknowledge the National Research Foundation (South Africa)
for financial support and the Central Analytical Facility of the
University of Stellenbosch for access to instrumentation.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
5
354
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 5354 –5358

Angewandte
Chemie
[
7]
norms, but without a documented example of the same
host framework with another guest.
Although the structure of hydrated trimesic acid (TMA)
almost meets the criteria of category (e), the host framework
consists of a supramolecular adduct that incorporates both
e)A molecular crystal in which Z ’ > 1 and the identical host
framework is precedented by at least one other structure
with a different guest.
[
18]
TMA and water, whereas the guest consists only of TMA.
A similar TMA/H O framework exists with picric acid as the
2
[
18]
guest, but there are small differences in the host frame-
works between the two structures. Therefore, the hydrated
TMA structure can at best be described as water-assisted self-
inclusion under category (d). We now report an unequivocal
case of true self-inclusion under category (e)as the missing
link at the extreme end of the series that represents all the
possible levels.
In discussing the above categories in turn, we concede that
it is not always possible to completely eliminate subjectivity as
the assignment of “host” and “guest” is generally not immune
to the liberties of interpretation. Herein, we adhere to the
suggestion by Cram that a host can be identified as a molecule
or supramolecular motif that possesses convergent binding
sites whereas a guest binds divergently. It should be noted
that binding may occur by means of many types of intermo-
lecular interactions, including weak dispersive forces, such as
van der Waals contacts.
[
7]
2,7-Dimethyl-octa-3,5-diyne-2,7-diol (1)was synthesized
as described below. Crystals of 1 ·CCl (2)were grown by slow
3
4
Clearly (a)and (b)are cases of reductio ad absurdum and
require no further discussion (namely, because any structure
qualifies as self-inclusion under one of these two categories).
Although the interpenetrated and interweaved frame-
[
5,6]
works
of category (c)host one another, it is not possible
evaporation of a solution of 1 in carbon tetrachloride and the
structure was determined by single-crystal X-ray diffraction
analysis. The binary compound crystallizes in the space group
R3, and the asymmetric unit consists of one molecule of 1 in a
to unequivocally distinguish between the host and guest in
such systems (for example, the structure of pure selenour-
[
2c]
ea). On the other hand, identical molecules can be viewed
as fulfilling two distinct roles in the structure for (d)and (e).
This observation is well illustrated in a report by Bishop and
general position and one-third of a molecule of CCl situated
4
on a crystallographic threefold rotation axis. The solid-state
adduct can reasonably be categorized as an inclusion com-
pound, with 1 fulfilling the role of the host and the solvent
that of the guest. The CÀCꢀC-CꢀCÀC spine of the host
[
2a]
Dance of a self-included alicyclic diol, in which molecules
of the same kind can be assigned as either the host or guest
according to Cram. However, this system must be placed in
category (d)as there is no evidence that the same host
framework can support the inclusion of a different guest.
In isolated examples, it is possible that a system can
appear to possess characteristics of more than one category.
Indeed, to illustrate the vagaries of interpretation, 4,4-bis(4-
hydroxyphenyl)cyclohexanone has been described as forming
molecule is slightly curved with four bond angles of approx-
imately 173, 176, 176, and 1768, respectively. The host
molecules are arranged in a spiral running parallel to [001]
to form an infinite (convergent)channel around the (diver-
gent)solvent guest molecules (Figure 1a .) The pitch of the
spiral is 6.232(4)Š (namely, the crystallographic c axis). Each
tube is surrounded by six identical tubes in a hexagonal
arrangement and, interestingly, all of the tubes in the crystal
spiral in the same direction. The hydroxy groups of 1 are
situated on the exterior surfaces of the tubes and each tube is
bound to all six of its neighbors by means of six infinite spiral
arrangements of OÀH···O hydrogen bonds (O···O: 2.733(6)
[
2b]
a self-included structure in accordance with category (d).
However, our interpretation is that the structure consists of
interweaved two-dimensional hydrogen-bonded frameworks
that conform to category (c).
To the best of our knowledge, no structure that conforms
entirely to category (e)has been recognized as such to date.
This phenomenon may be either unusual or commonplace:
there could be many such systems whose structures were
simply not commented on when reported. Indeed, given the
obvious potential for disorder, self-inclusion may have been
overlooked on many occasions. Of course this kind of
inclusion behavior in molecular solids is only to be expected
in host molecules that form a series of inclusion compounds
that are isostructural with respect to the host packing motif
and 2.657(7)Š; Figure 1a ). The guest solvent molecules in
the channels are packed parallel to [001] within each tube
(that is, the intrinsic threefold axis of each guest molecule
coincides with that of the host tubule, as shown in Figure 1b).
Notably, the linear packing of the guest molecules along [001]
is perfectly commensurate with the spiral pitch of the host
molecules.
Crystals of 1 ·C H (3)were grown from a solution of 1 in
3
6
6
(
we suggest use of the term “isoskeletal” to describe this
benzene. X-ray diffraction analysis reveals that 3 is isoskeletal
with respect to 2 and that the guest molecules are similarly
included in the channels in a commensurate arrangement
(Figure 1c)with a packing periodicity of 6.325(5)Š. The host/
guest ratio is also 1:3, and the benzene molecules are
disordered about the threefold rotation axis parallel to [001].
Our primary research interest involves engineering
porous crystals for gas- or vapor-sorption applications.
In this regard, we have been successful in either growing
suitable crystals of metastable apohost phases by sublima-
structural phenomenon). Many isoskeletal systems are now
known and include host–guest complexes of Dianinꢀs com-
pound, cholic acid, p-tert-butylcalix[4]arene, calix[4]ar-
[
9]
[10]
[11]
[
12]
[13]
[14]
ene,
hydroquinone,
Werner clathrates,
cyclodex-
[
15]
[16]
trins, Bishopꢀs alicyclic diol molecules, and many more.
The importance of these types of host is that their structures
can generally be predicted on the basis of their demonstrated
packing preferences, which facilitates the tailoring of struc-
[
11,19]
[
17]
ture as a stated goal of crystal engineering.
Angew. Chem. Int. Ed. 2006, 45, 5354 –5358ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5355

Communications
occurs concurrently with sublimation of the host compound
onset temperature (T ) = 63.58C). In a separate experiment,
(
on
crystals were placed under vacuum for several hours, after
which their structures were again determined by single-crystal
X-ray diffraction analysis. These two studies show that no
appreciable solvent loss appears to occur under vacuum
conditions and that heating to 708C results in guest removal,
but only with concomitant disassembly and volatilization of
the host framework.
Crystals of pure 1 (4)were grown by vacuum sublimation
at 708C in the hope of forming a guest-free tubular structure.
Indeed, single-crystal X-ray diffraction analysis of 4 reveals
an isoskeletal arrangement of 1 with respect to the structures
of 2 and 3. The tubular frameworks initially appeared to be
devoid of guest molecules. However, difference maps calcu-
lated during the latter stages of refinement indicated the
presence of appreciable residual electron density within the
channels. It was ultimately possible to model a disordered
molecule of 1 with its CÀCꢀC-CꢀCÀC spine situated on the
threefold rotation axis along [001].
It is immediately apparent that the structure of 4 accords
completely with category (e)of self-inclusion as delineated
above. It is therefore useful in the following discussion to
distinguish between the two crystallographically unique cases
of 1 according to their respective roles as host 1 and guest 1 .
h
g
Space-group symmetry requires the hydroxy and methyl
groups of 1g to be disordered about a threefold axis.
Furthermore, application of threefold rotation symmetry to
the asymmetric unit yields a molecule of 1 with a total site
g
occupancy of 0.5. The model employed for 1 is longer than
g
the crystallographic c axis and translation along [001] (that is,
x, y, z + 1)yields a second half-occupancy molecule which
overlaps the first (Figure 2a). The nature of the disorder of 1g
with regard to its possible positions along [100] is interesting
and warrants further discussion. Close inspection of Figure 1b
and c shows that the channels within the host tubules possess
bulges and constrictions along [001] with a periodicity equal
to the crystallographic c axis and that each bulge is occupied
by one carbon tetrachloride or benzene guest molecule.
However, in 4 the longer guest molecule 1 spans exactly two
g
bulges and is thus packed with a periodicity of 2c along [001]
within each channel. As X-ray diffraction analysis only
reveals the average structure, it is reasonable to infer from
the disordered model that two equally likely possibilities
Figure 1. a) Projection of 2 along [00-1] showing the hexagonal
arrangement of tubular motifs that consist of spiraling host molecules.
The guest molecules are situated in the channels. b) Projection
perpendicular to [001] of a tubule in 2 showing the commensurate
arrangement of carbon tetrachloride guest molecules within a channel
mapped using a spherical probe of radius 1.2 Š. c) A corresponding
projection of 3 that shows the benzene guest molecules within a
channel (also mapped using a probe of radius 1.2 Š). Hydrogen bonds
are shown as fragmented cylinders, guest molecules are shown in a
space-filling representation, and host molecules are shown as capped
sticks.
(
Figure 2b and c)exist for the end-to-end stacking of 1 along
g
[001] and that these two arrangements are staggered relative
to one another by a distance of c.
Evidently the tubular framework persistently observed in
structures 2–4 is a highly desirable arrangement of 1, which is
most likely due to the stabilizing influence of six infinite
hydrogen-bonded spirals that bind each tubule to its nearest
neighbors. Indeed, this kind of robust system is well known
and 2–4 can be considered members of the helical tubuland
[
16]
diol family described by Bishop. Using a probe of radius
1.2 Š to map the contact surface of each channel, we calculate
[
11a]
tion
and by extracting solvent molecules from guest
3
channels without causing the collapse of the host frame-
a channel volume of approximately 205(5)Š per repeat unit
[
19]
works.
In an attempt to remove the included solvent
along the c axis. It follows that carbon tetrachloride, benzene,
molecules by gentle heating, crystals of 2 and 3 were subjected
and 1 occupy the channels with packing coefficients of 0.44,
g
to thermogravimetric analysis, which reveals that solvent loss
0.39, and 0.44, respectively. These values are significantly
5
356
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 5354 –5358

Angewandte
Chemie
distinct roles in the structure. A strong case can be made that
this distinction is not subjective as at least two other structures
can be shown to possess the same host framework with
different guests (that is, the host competes for the guest-
binding void space). In the vast field of solid-state host–guest
chemistry, this observation is unprecedented at the level of
interpretative stringency displayed herein. The concept of
self-inclusion is an important phenomenon in the field of
crystal engineering that relates to porosity as it mitigates
against the formation of open structures. The assembly of any
porous structure represents a victory by competing factors
over the natural proclivity of molecules to close pack. As
discussed above, all structures exhibit self-inclusion to a
certain extent, and it is the challenge of crystal engineers to
understand and control this phenomenon in their quest to
form low-density frameworks with minimal self-inclusion.
Experimental Section
2,7-Dimethyl-octa-3,5-diyne-2,7-diol (1)was synthesized by the
Glaser reaction, in which 2-methylbut-3-yn-2-ol (1 mol)and pyridine
0.25 mol)were heated at 35 8C for 4 h in methanol with 1.2%
(
cuprous chloride as the catalyst and oxygen being passed through.
The product was extracted from aqueous ammonium chloride with
diethyl ether and recrystallized from benzene in 90% yield (m.p. 127–
1
298C). Further purification was achieved by sublimation. Solvated
Figure 2. Three identical projections of 4 along [100] that show the two
possible arrangements (light and dark) of the included examples of 1g
crystals were grown by slow evaporation of a solution of 1 in the
relevant solvent, whereas crystals of the pure form were grown by
sublimation at 708C under vacuum. X-ray diffraction data were
collected on a Bruker SMARTApex diffractometer and corrected for
Lorentz-polarization effects. Structures were solved by direct meth-
ods using SHELXS-97 and expanded/refined on F by difference
electron density synthesis using SHELXL-97 and the X-Seed graph-
ical user interface.
within each guest-accessible channel formed by 1 . a) An overlay of the
h
guest molecules that indicates the nature of the guest disorder. The
length of the crystallographic c axis is indicated. b) One distinct linear
arrangement of the guest molecules 1 and c) the alternative arrange-
g
2
ment of 1 within the same channel. The two arrangements are
g
staggered with respect to one another by translation of 6.341(1) Š
[
21–23]
along [001]. The hydroxy groups of 1 are threefold disordered about
g
the linear axis of the molecule and are not shown. As both arrange-
Received: April 27, 2006
Published online: July 17, 2006
ments of 1 are equally likely, it is presumed that a random arrange-
g
ment is assumed within each channel and that no synchronization
occurs between neighboring channels with regard to periodicity.
Keywords: crystal engineering · host–guest systems ·
.
inclusion compounds · solid-state structures ·
supramolecular chemistry
lower than packing coefficients normally observed for organic
solids (typically 0.60–0.66)and provide further evidence for
the overall stability of the host framework.
[
1] a)J. Bernstein, Chem. Commun. 2005, 5007 – 5012; b)A.
Nangia, Cryst. Growth Des. 2006, 6, 2 – 4; c)O. Almarsson,
M. J. Zaworotko, Chem. Commun. 2004, 1889 – 1896; d)G. R.
Desiraju, Cryst. Eng. Comm. 2003, 5, 466 – 467; e)J. D. Dunitz,
Cryst. Eng. Comm. 2003, 5, 506 – 506; f)K. R. Seddon, Cryst.
Growth Des. 2004, 4, 1087 – 1087; g)G. R. Desiraju, Cryst.
Growth Des. 2004, 4, 1089 – 1090; h)J. Bernstein, Cryst.
Growth Des. 2005, 5, 1661 – 1662; i)D. Braga, F. Grepioni,
Chem. Commun. 2005, 3635 – 3645.
It is noteworthy that crystals of 4 can also be obtained by
[
20]
slow evaporation of a solution of 1 in dichloromethane.
Presumably 1 competes with the solvent molecules for
occupancy of the channels. If included, dichloromethane
would occupy the channels with a rather low packing
coefficient of 0.28, and this poor efficiency may thus be
used to rationalize why 1 is included in preference to
dichloromethane. On the other hand, benzene is included
despite possessing a slightly lower space filling efficiency than
[
2] a)K. C. Pich, R. Bishop, D. C. Craig, I. G. Dance, A. D. Rae,
M. L. Scudder, Struct. Chem. 1993, 3, 41 – 51; b)S. Aitipamula,
G. R. Desiraju, M. Jaskólski, A. Nangia, R. Thaimattam, Cryst.
Eng. Comm. 2003, 5, 447 – 450; c)F. H. Herbstein, Crystalline
Molecular Complexes, Vol. 1, Oxford University Press, Oxford,
1
. However, in the case of benzene, additional stabilization
may be derived from C ÀH···pCꢀC intermolecular interactions
Ar
between the guest and host.
2005.
We have described the solid-state structures of three
different inclusion compounds of 1, which have in common
that they possess tubular arrangements of the host that are
virtually superimposable. Remarkably, one of these structures
features a host molecule as the guest. Although 4 is composed
of a pure substance, the molecule clearly plays two very
[
3] J. Dalluhn, H.-F. Pröhl, O. Moers, A. Blaschette, P. G. Jones, Z.
Naturforsch. B 2001, 56, 1188 – 1195.
[
[
4] J. W. Steed, Cryst. Eng. Comm. 2003, 5, 169 – 179.
5] a)B. Moulton, M. J. Zaworotko, Chem. Rev. 2001, 101, 1629 –
1658; b)P. J. Langley, J. Hulliger, Chem. Soc. Rev. 1999, 28, 279 –
291.
Angew. Chem. Int. Ed. 2006, 45, 5354 –5358ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5357

Communications
[
[
6] S. R. Batten, Cryst. Eng. Comm. 2001, 3, 67 – 73.
7] D. J. Cram, Angew. Chem. 1986, 98, 1041 – 1060; Angew. Chem.
Int. Ed. Engl. 1986, 25, 1039 – 1057.
[
8] J. S. Rutherford, C. Calvo, Z. Kristallogr. Kristallgeom. Kristall-
phys. Kristallchem. 1969, 128, 229 – 258.
[
9] P. Finocchiaro, S. Failla, Comprehensive Supramolecular
Chemistry, Vol. 6 (Eds.: J. L. Atwood, J. E. D. Davies, D. D.
MacNicol, F. Vögtle), Elsevier, Oxford, 1996, pp. 618 – 627.
[
[
10] M. Miyata, K. Sada, Comprehensive Supramolecular Chemistry,
Vol. 6 (Eds.: J. L. Atwood, J. E. D. Davies, D. D. MacNicol, F.
Vögtle), Elsevier, Oxford, 1996, pp. 147 – 176.
11] a)J. L. Atwood, L. J. Barbour, A. Jerga, B. L. Schottel, Science
2
002, 298, 1000 – 1002; b)J. L. Atwood, L. J. Barbour, A. Jerga,
Angew. Chem. 2004, 116, 3008 – 3010; Angew. Chem. Int. Ed.
004, 43, 2948 – 2950; c)J. L. Atwood, L. J. Barbour, G. O. Lloyd,
2
P. K. Thallapally, Chem. Commun. 2004, 922 – 923; d)J. L.
Atwood, L. J. Barbour, P. K. Thallapally, T. B. Wirsig, Chem.
Commun. 2005, 51 – 53.
[
12] J. L. Atwood, L. J. Barbour, A. Jerga, Science 2002, 296, 2367 –
2369.
[
13] T. C. W. Mak, B. R. F. Bracke, Comprehensive Supramolecular
Chemistry, Vol. 6 (Eds.: J. L. Atwood, J. E. D. Davies, D. D.
MacNicol, F. Vögtle), Elsevier, Oxford, 1996, pp. 23 – 37.
[
[
[
14] J. Lipkowski, Comprehensive Supramolecular Chemistry, Vol. 6
(
Eds.: J. L. Atwood, J. E. D. Davies, D. D. MacNicol, F. Vögtle),
Elsevier, Oxford, 1996, pp. 691 – 714.
15] K. Harata, Comprehensive Supramolecular Chemistry, Vol. 3
(
Eds.: J. L. Atwood, J. E. D. Davies, D. D. MacNicol, F. Vögtle),
Elsevier, Oxford, 1996, pp. 279 – 304.
16] R. Bishop, Comprehensive Supramolecular Chemistry, Vol. 6
(
Eds.: J. L. Atwood, J. E. D. Davies, D. D. MacNicol, F. Vögtle),
Elsevier, Oxford, 1996, pp. 85 – 115.
[
[
[
17] For example: G. O. Lloyd, M. W. Bredenkamp, L. J. Barbour,
Chem. Commun. 2005, 4053 – 4055.
18] F. H. Herbstein, R. E. Marsh, Acta Crystallogr. Sect. B 1977, 33,
2358 – 2368.
19] a)L. J. Barbour, Chem. Commun. 2006, 1163 – 1168; b)L.
Dobrza n´ ska, G. O. Lloyd, H. G. Raubenheimer, L. J. Barbour,
J. Am. Chem. Soc. 2005, 127, 13134 – 13135; c)L. Dobrzan´ska,
G. O. Lloyd, H. G. Raubenheimer, L. J. Barbour, J. Am. Chem.
Soc. 2006, 128, 698 – 699.
[
20] D. A. Leigh, A. E. Moody, R. G. Pritchard, Acta Crystallogr.
Sect. C 1994, 50, 129 – 131. It is our assessment that this structure
was misinterpreted; we believe that the authors assumed the
electron density within the tubule to be due to the presence of
dichloromethane and they did not attempt to model the guest
further.
[
[
21] G. M. Sheldrick, SHELX-97, University of Göttingen, 1997.
22] a)L. J. Barbour, J. Supramol. Chem. 2001, 1, 189 – 191; b)J. L.
Atwood, L. J. Barbour, Cryst. Growth Des. 2003, 3, 3 – 8.
23] The crystal data, thermoanalytical, and complementary analyses
can be found in the Supporting information. CCDC 605735–
[
605737 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.
5
358
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 5354 –5358