Rational design of supramolecular chirality in porphyrin assemblies: The halogen bond case

Article abstract of DOI:10.1039/b719625c

Asymmetric functionalization of the tetraarylporphyrin scaffold, combined with directional supramolecular halogen bonding, yields chiral architectures. The Royal Society of Chemistry.

Full text of DOI:10.1039/b719625c

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Rational design of supramolecular chirality in porphyrin assemblies: the
halogen bond casewz
Sankar Muniappan, Sophia Lipstman and Israel Goldberg*
Received (in Cambridge, UK) 20th December 2007, Accepted 23rd January 2008
First published as an Advance Article on the web 14th February 2008
DOI: 10.1039/b719625c
Asymmetric functionalization of the tetraarylporphyrin scaffold,
combined with directional supramolecular halogen bonding,
yields chiral architectures.
directing the construction of a chiral structure from a non-
chiral module, has been reported recently.3h
All three crystal structures can best be described as being
composed of flat supramolecular layers of porphyrin units
(irrespective of their metalation state, molecular conformation
and unit cell content), interconnected to one another by head-
to-tail Nꢀ ꢀ ꢀI bonds within 2.87–3.06 A between adjacent mo-
lecules, perfectly aligned along straight lines (Fig. 2). The trans-
related pyridyl and iodophenyl groups of the porphyrin moiety
are involved in these interactions at distances that are consider-
ably shorter than the sum of the corresponding van der Waals
radii of N and I atoms (ca. 3.5–3.8 A).3a,4 Such neighboring
chains further interact on both sides via secondary Iꢀ ꢀ ꢀp inter-
actions (at iodine-ring distances within 3.4–3.6 A, with the
sideways-directed C–I bonds of one chain being oriented nearly
perpendicular to the iodophenyl and pyridyl aryl groups of
adjacent Nꢀ ꢀ ꢀI-bonded chains).5 In 2, attractive inter-chain
Iꢀ ꢀ ꢀI contacts (within 3.68–3.73 A) are also apparent.
The tailored induction of supramolecular chirality in non-
covalent assemblies in solution, as well as in the solid state,
without resorting to chiral reagents has been intensively
studied in recent years, as such supramolecular materials
may be useful in the processes of enantioselective synthesis,
separation and catalysis, and the fabrication of polar devices.1
While seeking protocols for the deliberate synthesis of homo-
chiral supramolecular crystals from typically square planar
tetraarylporphyrin components, our earlier studies revealed
that introduction of a polar axis into the porphyrin unit
(whether in a lateral or an axial direction) via asymmetric
substitution of the molecular recognition groups (hydrogen
bond donors and acceptors), in combination with specific
directionality of the intermolecular interaction synthons that
govern the self-assembly process, represents a useful strategy
for the induction of non-centrosymmetric architectures.2
Exploring further the above guidelines, we report here the
deliberate synthesis of non-centrosymmetric crystals based on
the pre-designed compound 5-(40-pyridyl)-10,15,20-tris(40-
iodophenyl)porphyrin (PyTIPP; Fig. 1) in different reaction
environments and metalation states.y PyTIPP bears a single
pyridyl function and three diverging iodophenyl groups that
can be involved in halogen bonds of the Npyꢀ ꢀ ꢀI type, as well
as related Iꢀ ꢀ ꢀI and Iꢀ ꢀ ꢀp interactions,3 and its molecular
framework has a polar axis due to the asymmetric substitu-
tion. The resulting structures, involving the nitrobenzene
solvate of PyTIPP (1), the DMF solvate of [Zn(py)-PyTIPP]
(2) and the pyridine solvate of [Cu-PyTIPP] (3) were mon-
itored by single crystal X-ray diffraction.z Together, they
demonstrate a consistent assembly of the PyTIPP scaffolds
into chiral architectures. This study provides a unique example
of the induction of supramolecular chirality by halogen bonds
(the tendency of halogen atoms to be involved in directional
interactions with lone pair-possessing atoms) and confirms
that such bonding can serve as an important tool in the
patterning of supramolecular assemblies.3 A unique example,
showing that strong Iꢀ ꢀ ꢀI halogen bonding can be used for
In the three compounds, the halogen-bonded layers offset-
stack along the normal direction at an interlayer distance of
approximately 4.5 A, as has been commonly observed in
numerous porphyrin structures.6 More interestingly, however,
all the layers are aligned in the same direction, imparting
chirality upon the entire 3D assembly. The latter is character-
ized by space group symmetry P1 in 1 and 2, and C2 in 3. The
chirality of the layered structure is preserved, even though
there are four crystallographically-independent molecules (and
layers) in 2 and two independent porphyrin moieties (and
layers) in 3 (Fig. 3). The interlayer stacking is stabilized by
dispersion forces between the porphyrin assemblies, as well as
Fig. 1 (a) Structural formula of the porphyrin building blocks
[M-PyTIPP]. (b) Space-filling illustration of the free base PyTIPP
derivative. The arrow indicates the main direction of the molecular
asymmetry. Note the lack of inversion or strict mirror symmetry in the
observed molecular conformation in the crystal (see text).
School of Chemistry, Sackler Faculty of Exact Sciences, Tel Aviv
University, Tel Aviv, Israel. E-mail: Goldberg@post.tau.ac.il;
Fax: +972 3 6409293; Tel: +972 3 6409965
w Electronic supplementary information (ESI) available: Details of the
synthetic procedures. See DOI: 10.1039/b719625c
ꢁc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 1777–1779 | 1777

View Article Online
given Nꢀ ꢀ ꢀI halogen-bonded chain preferentially effect the
assembly of chains of similar chirality into layers of specific
directionality, while optimising the inter-chain arrangement by
Iꢀ ꢀ ꢀI and Iꢀ ꢀ ꢀp secondary interactions that involve the iodo-
phenyl groups perpendicular to the chains. The observed
2.87–3.06 A Nꢀ ꢀ ꢀI contacts represent halogen bonds of med-
ium strength. They are only slightly more extended that the
shortest Nꢀ ꢀ ꢀI bonds, of 2.69 A, recorded thus far,3h yet they
appear to be adequately effective in directing the multi-
porphyrin self-assembly process in 1–3.
Structures 1–3 reveal characteristic offset stacking of the
supramolecular porphyrin layers, stabilized by dispersion
forces6 and supplemented by a minor coulombic contribution.
For a quantitative estimate of the involved energies,7 the
atom–atom potential level8 yields interactions of ꢂ45 to
ꢂ53 kJ molꢂ1 between pairs of closely-stacked porphyrin
molecules, and PIXEL analysis9,10 yields an interaction of
ꢂ11 kJ molꢂ1 between the model molecular pair iodobenzene–
pyridine. Dispersive interactions are obviously orientation-
insensitive, and the minor but quite effective halogen bond
contribution could equally be well satisfied in a chiral or
racemic arrangement. Thus, enthalpic contributions alone do
not seem to be adequate for a discussion of crystal stereo-
chemistry, which could be highly sensitive to entropic or even
kinetic factors at the nucleation and growth stages. So how
does nature account for the negative entropy associated with
the chiral stacking in 1–3? The answer to this question is
provided by the crystallographic evaluations. As the structural
models involved in this study incorporate heavy electron-rich
iodine atoms, analyses of the anomalous dispersion effects in
the crystallographic refinements indicate that all the examined
crystals are twinned to a varying degree.z Thus, they represent
conglomerates of solid domains of opposed chirality. The
Fig. 2 (a) The supramolecular layered assembly of PyTIPP in 1. The
dashed Iꢀ ꢀ ꢀN contact indicates the directional halogen bond, while the
I0ꢀ ꢀ ꢀN and I00ꢀ ꢀ ꢀC dashed lines represent the Iꢀ ꢀ ꢀp interactions. (b)
Uni-directional arrangement of two successive layers; only the Nꢀ ꢀ ꢀI
halogen bonds are indicated by dashed lines. (The crystal packing
diagrams in all figures exclude the crystallization solvent.)
by cooperative dipolar attractions between iodophenyl groups
(those not involved in the Nꢀ ꢀ ꢀI bonds) of neighbouring layers,
which are aligned in an anti-parallel fashion.
It is clearly apparent from the above examples that the
consistent occurrence of supramolecular chirality in crystals of
the [M-PyTIPP] scaffold (1–3) is associated with the asym-
metric substitution of the pyridyl and iodophenyl functional
groups on the porphyrin core. Moreover, the molecular
structure cannot sustain a planar macrocyclic core, with the
aryl substituents perfectly perpendicular to it, due to steric
hindrance between the proxime protons on the peripheral C-
atoms of the pyrrole groups and the ortho positions of the aryl
rings (Fig. 1(b)). Thus, the [M-PyTIPP] unit, irrespective of its
actual conformation (planar or distorted) and whether it is in
its free base or metalated form, lacks inversion or mirror plane
symmetry elements and is chiral (the chirality of the individual
molecular units is particularly apparent in structure 3 due to
their marked saddle-type deformation). This asymmetry, when
combined with effective directional Nꢀ ꢀ ꢀI halogen bonding
between the porphyrin units (in spite of the low energy
associated with the iodoꢀ ꢀ ꢀpyridine bonds),3c,d is transmitted
to the layered supramolecular assemblies that form, and
subsequently to the entire crystalline architecture of the
stacked layers. In a recent publication, Miyata and co-workers
proposed an elegant empirical explanation of the generation of
chiral layered crystals from dynamically achiral (in solution)
molecular units, which as a result of the supramolecular
interactions, freeze in a chiral form.1c The observation of
supramolecular chirality in the PyTIPP-based structures can
be interpreted in similar terms. As in Miyata’s case, the
induced asymmetry of the porphyrin and the polarity of a
Fig. 3 (a) Uni-directional arrangement of the four crystallographi-
cally-independent layers (viewed edge-on) of 2 in space group P1. (b)
Supramolecular layered assembly of [Cu-PyTIPP] units and the uni-
directional arrangement of three successive layers in 3. The Nꢀ ꢀ ꢀI
halogen bonds are marked by dashed lines. Note that in 1, the
porphyrin core in flat, in 2 the Zn(py)-porphyrin core has a domed
topology, while in 3 the Cu-porphyrin core reveals a saddle-type
deformation into a chiral form.
ꢁc
This journal is The Royal Society of Chemistry 2008
1778 | Chem. Commun., 2008, 1777–1779

View Article Online
J = 5.9 Hz), 8.80 (m, 8H), 8.10 (asymmetric t, 8H, J = 6 Hz, J = 8.6
occurrence of twinned (conglomerate), rather than racemic
crystals, provides another indication of the high propensity of
the [M-PyTIPP] system to exhibit, from a microscopic struc-
tural point of view, supramolecular chirality in the crystalline
state.
Hz), 7.89 (d, 6H, J = 8.2 Hz), ꢂ2.93 (s, 2H). UV-vis in CH2Cl2: lmax
/
nm (log e) 419 (5.96), 515 (4.60), 549 (4.18), 589 (4.01), 645 (3.72). FAB
mass spectrum (m/z) for C43H26N5I3: found 994, calc. 993.43.
Preparation of metal derivatives: The free base porphyrin (0.05 g,
5.03 mmol) was dissolved in 25 ml of CHCl3. To this, M(OAc)2
hydrate (where M = Cu(II) and Zn(II)) (0.09–0.125 g, 0.05 mol) in 2 ml
of methanol was added, and the mixture heated for 10 min on a water
bath at 75 1C and then evaporated to dryness. The crude porphyrin
was next washed with water. The metal complexes were dried under
vacuum and their yields were found to be almost quantitative. [Zn-
PyTIPP]: 1H NMR (DMSO-d6): 8.98 (d, 2H, J = 5.8 Hz), 8.79 (m,
8H), 8.16 (asymmetric t, 8H, J = 5.76 Hz, J = 8.2 Hz), 7.94 (d, 6H, J
= 8.2 Hz). UV-vis in a CH2Cl2/pyridine mixture: lmax/nm (log e) 429
(5.69), 562 (4.15), 602 (3.69). FAB mass spectrum (m/z) for
C43H24N5I3Zn: found 1057, calc. 1056.8. [Cu-PyTIPP]: UV-vis in a
CH2Cl2/pyridine mixture: lmax/nm (log e) 416 (5.82), 540 (4.37), 589
(4.18). FAB mass spectrum (m/z) for C43H24N5I3Cu: found 1055, calc.
1054.96.
However, this tendency cannot be harnessed at the present
time in practical applications due to the conglomerate appear-
ance of these materials. Yet, this study (along with our earlier
results),2 provides a useful concept and a reliable protocol for
the supramolecular synthesis of polar porphyrin frameworks,
which will be applied in future attempts to crystal-engineer
chiral microporous solids.11 A rigorous thermodynamic inter-
pretation of the observed supramolecular chirality in crystals
provides a considerable challenge for future investigations.
This research was supported by the Israel Science
Foundation.
Supramolecular synthesis: 4 mg (0.004 mmol) of PyTIPP was
dissolved in 2 ml of chloroform, followed by a few drops of either
nitrobenzene or ethyl benzoate, and allowed to evaporate slowly. X-
Ray quality crystals of the nitrobenzene solvate (structure 1) were
obtained after 7 d. When B4 mg (0.004 mmol) of [Zn-PyTIPP] and
[Cu-PyTIPP] were dissolved in 1.5 ml of a 2 : 1 mixture of DMF/
pyridine (v/v) and allowed to evaporate slowly, X-ray quality crystals
of a DMF solvate of [Zn(py)-PyTIPP] (structure 2) and a pyridine
solvate of [Cu-PyTIPP] (structure 3) were obtained after 10 d. The
uniform identity of the crystal lattices formed from different reactions
was confirmed in each case by repeated measurements of the unit cell
dimensions of different single crystallites. All crystals were
coloured red/purple.
Notes and references
z Crystal data:
1: C43H26I3N5ꢀ2C6H5NO2, Mr = 1239.61, triclinic, space group P1,
a = 10.6814(3), b = 11.1396(3), c = 11.4818(4) A, a = 103.635(1), b
= 93.441(1), g = 110.096(2)1, V = 1231.87(6) A3, T = 110 K, Z = 1,
m
Mo-Ka= 19.57 cmꢂ1, rcalc = 1.671 g cmꢂ3, prism 0.35 ꢃ 0.20 ꢃ 0.15
mm, 13 768 reflections measured, of which 9381 were unique (Rint
=
0.046) and 7890 with I 4 2s(I). Final R1 = 0.042 and wR2 = 0.084
for the 7890 data above the intensity threshold. CCDC 662252.
2: C48H29I3N6ZnꢀC3H7NO, Mr = 1208.94, triclinic, space group
P1, a = 15.9721(4), b = 17.4838(4), c = 18.8968(5) A, a = 66.985(1),
b = 86.963(1), g = 75.270(1)1, V = 4690.9(2) A3, T = 110 K, Z = 4,
1 (a) L. J. Prins, J. Huskens, F. de Jong, P. Timmerman and D. N.
Reinhoudt, Nature, 1999, 398, 498–502; (b) R. Purrello, Nat.
Mater., 2003, 2, 216; (c) A. Tanaka, K. Inoue, I. Hisaki, N. Tohnai,
M. Miyata and A. Matsumoto, Angew. Chem., Int. Ed., 2006, 45,
4142; (d) G. Seeber, B. E. F. Tiedemann and K. N. Raymond, Top.
Curr. Chem., 2006, 265, 147; (e) T. Imakubo, T. Maruyama, H.
Sawa and K. Kobayashi, Chem. Commun., 1998, 2001.
2 The induction of supramolecular chirality in porphyrin crystals
directed by O–Hꢀ ꢀ ꢀCl and O–Hꢀ ꢀ ꢀN hydrogen bonds has been
demonstrated: (a) S. George and I. Goldberg, Cryst. Growth Des.,
2006, 6, 755; (b) S. George, S. Lipstman, S. Muniappan and
I. Goldberg, CrystEngComm, 2006, 8, 417; (c) M. Vinodu and
I. Goldberg, CrystEngComm, 2005, 7, 133.
3 (a) F. F. Awwadi, R. D. Willett, K. A. Peterson and B. Twamley,
Chem.–Eur. J., 2006, 12, 8952; (b) P. Metrangolo, H. Neukirch, T.
Pilati and G. Resnati, Acc. Chem. Res., 2005, 38, 386; (c) B. K.
Saha, A. Nangia and M. Jaskulski, CrystEngComm, 2005, 7, 355;
(d) R. B. Walsh, C. W. Padgett, P. Metrangolo, G. Resnati, T. W.
Hanks and W. T. Pennington, Cryst. Growth Des., 2001, 1, 165; (e)
P. Metrangolo and G. Resnatti, Chem.–Eur. J., 2001, 7, 2511; (f) E.
Bosch and C. L. Barnes, Cryst. Growth Des., 2002, 2, 299; (g) A.
Casnati, R. Liantonio, P. Metrangolo, G. Resnati, R. Ungaro and
F. Ugozzoli, Angew. Chem., Int. Ed., 2006, 45, 1915; (h) M. Amati,
F. Lelj, R. Liantonio, P. Metrangolo, S. Luzzati, T. Pilati and G.
Resnati, J. Fluorine Chem., 2004, 125, 629.
m
Mo-Ka = 25.44 cmꢂ1, rcalc = 1.712 g cmꢂ3, needle 0.60 ꢃ 0.40 ꢃ 0.20
mm, 39 597 reflections measured, of which 29 624 were unique (Rint
=
0.066) and 20 744 with I 4 2s(I). Final R1 = 0.067 and wR2 = 0.159
for the 20 744 data above the intensity threshold. CCDC 662254.
3: C43H24CuI3N5 (excluding disordered solvent), Mr = 1054.91,
monoclinic, space group C2, a = 24.2489(13), b = 20.5698(11), c =
9.1014(5) A, b = 94.477(4)1, V = 4525.9(4) A3, T = 110 K, Z = 4,
m
Mo-Ka = 25.61 cmꢂ1, rcalc = 1.548 g cmꢂ3, plate 0.30 ꢃ 0.20 ꢃ 0.05
mm, 20 122 reflections measured, of which 10 024 were unique (Rint
=
0.066) and 7222 with I 4 2s(I). Final R1 = 0.060 and wR2 = 0.148
for the 7222 data above the intensity threshold. CCDC 662255. In 3,
the contribution of the disordered solvent, which couldn’t be
modelled, was subtracted from the diffraction pattern.
The three chiral layered crystal structures (with all the layers
oriented in the same direction) are pseudo-centrosymmetric, and the
molecular frameworks within them arrange in a nearly centrosym-
metric manner. It is the asymmetric disposition of the iodine atom
between the halogen-bonded pyridyl and iodophenyl groups that
makes the major difference; the iodine is 2.87–3.07 A distant from
the N-site it points at, but only B2.1 A from the Cphenyl atom it
covalently binds to. Alternative structural models, representing race-
mic intermolecular organization, were found to be highly inconsistent
with the diffraction data, yielding high R-factors (the diffraction
pattern is dominated by the strong I-scatterers, and the R-factors
are very sensitive to their accurate positioning). All of the crystals were
found to be twins of varying composition of inter-grown domains of
opposing chirality: 1 : 1 (1), B1 : 9 (2) and B1 : 2.5 (3).
4 A. Bondi, J. Phys. Chem., 1964, 68, 441.
5 Similar Iꢀ ꢀ ꢀp and Iꢀ ꢀ ꢀI interactions characterize the crystal struc-
ture of tetrakis(iodophenyl)porphyrin: S. Lipstman, S. Muniappan
and I. Goldberg, Acta Crystallogr., Sect. C: Cryst. Struct.
Commun., 2007, 63, m300–m303.
For crystallographic data in CIF or other electronic format
6 (a) M. P. Byrn, C. J. Curtis, Y. Hsiou, S. I. Khan, P. A. Sawin, S.
K. Tendick, A. Terzis and C. E. Strouse, J. Am. Chem. Soc., 1993,
115, 9480; (b) R. Krishna-Kumar, S. Balasubramanian and I.
Goldberg, Inorg. Chem., 1998, 37, 541.
see DOI: 10.1039/b719625c
y Synthesis of PyTIPP: 1.25 g (5.4 mmol) of 4-iodobenzaldehyde and
0.34 ml (3.6 mmol) of 4-pyridinecarboxaldehyde were added to 35 ml
of hot propionic acid at 120 1C. Then, 0.5 ml of distilled pyrrole (7.21
mmol) was added, and the resulting solution heated to reflux for 90
min. Next, the solvent was removed by vacuum distillation, and the
crude product digested with hot water to remove traces of propionic
acid, then filtered, washed with water and dried. The crude porphyrin
mixture was dissolved in chloroform and loaded onto a silica column
using chloroform. The desired porphyrin, PyTIPP, was eluted with 3%
acetone in chloroform as the second fraction. The yield was found to
be 0.165 g (9%) based on pyrrole. 1H NMR (CDCl3): 9.00 (dd, 2H,
7 A. Gavezzotti, personal communication.
8 A. Gavezzotti and G. Filippini, J. Phys. Chem., 1994, 98, 4831.
9 (a) A. Gavezzotti, J. Chem. Theory Comput., 2005, 1, 834; (b) J. D.
Dunitz and A. Gavezzotti, Cryst. Growth Des., 2005, 5, 2180.
10 A. Gavezzotti, in Molecular Aggregation: Structure Analysis and
Molecular Simulation of Crystals and Liquids, Oxford University
Press, Oxford, 2007, pp. 304–325.
11 I. Goldberg, Chem. Commun., 2005, 1243.
ꢁc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 1777–1779 | 1779