Revealing molecular self-assembly and geometry of non-covalent halogen bonding by solid-state NMR spectroscopy

Article abstract of DOI:10.1039/b813237b

We report a new spectroscopic fingerprint of intermolecular contacts in halogen bond-driven self-assembling aggregates and a precise determination of intermolecular N...I distances in microcrystalline samples. The Royal Society of Chemistry.

Full text of DOI:10.1039/b813237b

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Revealing molecular self-assembly and geometry of non-covalent halogen
bonding by solid-state NMR spectroscopyw
Markus Weingarth,a Noureddine Raouafi,b Benjamin Jouvelet,a Luminita Duma,a
b
b
a
a
¨
Geoffrey Bodenhausen, Khaled Boujlel, Bernd Schollhorn* and Piotr Tekely*
Received (in Cambridge) 31st July 2008, Accepted 9th September 2008
First published as an Advance Article on the web 15th October 2008
DOI: 10.1039/b813237b
We report a new spectroscopic fingerprint of intermolecular
contacts in halogen bond-driven self-assembling aggregates
and a precise determination of intermolecular NÁ Á ÁI distances
in microcrystalline samples.
Heteronuclear dipolar couplings provide precise information
about internuclear distances. Magic-angle spinning (MAS),
which is mandatory to achieve sufficient resolution, causes
weak heteronuclear dipolar interactions to be completely
averaged out even at slow spinning speeds, so that a recou-
pling procedure is required in the indirect dimension of a
suitable two-dimensional (2D) NMR experiment. In the meth-
od discussed in this work, we take advantage of the well-
resolved isotropic chemical shifts of nitrogen-15 in the direct
o2 dimension, while information about dipolar nitrogen–
iodine (15N–127I) interactions is recovered in the indirect o1
dimension. The pulse sequence consists of conventional cross-
polarization (CP) from protons to nitrogen-15 followed by a
spin-lock period during which a radio-frequency (RF) field is
applied only to the nitrogen-15 nuclei with an amplitude that
matches the spinning frequency n1(15N) = nr to fulfill the so-
called n = 1 rotary resonance condition.10 The method does
not require any irradiation of iodine-127 (S = 5/2) which
would be challenging because of the large quadrupolar inter-
actions and resulting linewidths.11 Proton decoupling is ap-
plied throughout the evolution and acquisition periods. The
evolution of the 15N magnetization under rotary resonance is
therefore determined by only two recoupled interactions: the
15N chemical shift anisotropy (CSA) and the heteronuclear
dipolar 15N–127I coupling. So far, the rotary resonance effect
has been exploited for 13C and 1H CSA measurements12,13 and
for determining intramolecular dipolar interactions for
31P–15N, 13C–14N, 13C–31P and 31P–113Cd spin pairs.10–16 In
this work, the principle is extended to intermolecular distance
measurements.
Non-covalent halogen bonding1,2 plays an important role
in crystal engineering. Molecules bearing halogen atoms
(Cl, ;Br, I) can act as electron density acceptors that are able
to form complexes with Lewis bases containing donor atoms
such as nitrogen, oxygen or sulfur. These non-covalent inter-
actions have a strength that is comparable to hydrogen
bonds.4 They can be strong enough to control the aggregation
of organic molecules in solids, liquids and gases.3 The direc-
tionality of halogen bonds contributes to determine the rela-
tive orientation of molecules in solids, thus permitting the
tailored design of novel materials. Interesting examples of
NÁ Á ÁI halogen bonding have been identified in supramolecular
chemistry, in various solid-state reactions, in cation receptors
and in functional materials such as liquid crystals and
molecular-imprinted polymers.5 Recently it has also been
recognized that halogen bonds can play a central role in ligand
binding and molecular folding of macromolecules such as
proteins and DNA.6
There is a great deal of interest in a better understanding of
halogen bonding in the solid state, especially for micro-
crystalline structures for which single crystal X-ray diffraction
data are not available. Benzyl-di(4-iodobenzyl)-amine 17 (see
Scheme 1) bearing two (iodine) acceptor sites as well as a
(nitrogen) donor site was expected to aggregate by forming
polymeric chains via intermolecular NÁ Á ÁI halogen bonds, in
analogy to 4-iodo-perfluoro benzaldehyde.8
The one-dimensional 15N CP-MAS spectrum of benzyl-di-
(4-iodobenzyl)-amine 1 (Fig. 1) reveals two well resolved
isotropic chemical shifts. This must be due to the presence of
However, the quality of our crystals did not allow structure
resolution by X-ray diffraction. This prompted us to use solid-
state NMR spectroscopy. So far the formation of halogen-
bonded complexes has been investigated only by 19F and 13C
NMR spectroscopy in solution.9 To find spectroscopic evi-
dence of halogen bonds in the solid state, we determined the
dipolar interactions between the spins of 15N and 127I nuclei.
a
´
´
Departement de Chimie, associe au CNRS, Ecole Normale
Supe´rieure, 24 rue Lhomond, 75231 Paris Cedex 05, France.
E-mail: Piotr.Tekely@ens.fr. E-mail: Bernd.Schollhorn@ens.fr
b
´
Departement de Chimie, Faculte des Science de Tunis, Universite de
Tunis, El Manar, Tunisie
´
´
w Electronic supplementary information (ESI) available: Rotary
resonance lineshapes simulations, synthesis details. See DOI:
10.1039/b813237b
Scheme 1 Synthesis of benzyl-di(4-iodobenzyl)-amine (1).
Chem. Commun., 2008, 5981–5983 | 5981
ꢀc
This journal is The Royal Society of Chemistry 2008

View Article Online
Fig. 1 Experimental 15N CP-MAS spectrum (black) and simulated
subspectra (green) of the two sites I and II of compound 1. The
experimental spectrum was recorded at 40.5 MHz with a cross-
polarization (CP) contact time of 1 ms and a spinning frequency nr
= 364 Hz. The fitted parameters of the 15N chemical shift tensor of the
two sites are given in Table 1.
two magnetically nonequivalent and crystallographically dis-
tinct 15N sites in the unit cell. These two peaks cannot result
from a residual dipolar splitting (RDS) due to a second-order
quadrupolar–dipolar cross-term. Indeed, simulations show
that for spin 5/2 an asymmetric triplet should be observed
instead of a doublet. Moreover, the carbon-13 resonances (not
shown) of the carbon atoms that are directly attached to
iodine do not feature any RDS. This corroborates reports of
typical quadrupolar coupling constants for 127I nuclei of a few
MHz in a wide range of compounds.11 Finally, as expected for
through-space contacts, the spectra recorded at different spin-
ning speeds do not reveal any (isotropic or anisotropic)
J(15N–127I) coupling which would show up as an asymmetric
splitting with spinning sidebands with non-symmetric line-
shapes.17
Fig. 2 Experimental (black lines) 15N solid-state NMR rotary reso-
nance lineshape from site I (downfield signal in Fig. 1) of compound 1
along with simulated lineshapes assuming (i) only an 15N anisotropic
chemical shift interaction (blue line) and (ii) the presence of a hetero-
nuclear dipolar 15N–127I coupling in addition to the CSA (red lines).
The simulations were carried out with the SIMPSON program.19 All
relevant parameters are given in Table 1. The experimental lineshape
was obtained by Fourier transformation of an 15N signal recorded as a
function of the spin-lock duration t1 at the n = 1 rotary resonance
condition n1 = nr = 10 kHz using a 400 MHz spectrometer equipped
with a 4 mm triple-resonance CP-MAS probe, and a 1H–15N CP
contact time of 1 ms. CW and XiX proton decoupling20 with
n1H = 140 kHz were used in the evolution and acquisition periods,
respectively. The internuclear distance is rN–I = 2.7 Æ 0.04 A.
Consequently, the spectra obtained with a slow spinning speed
give access to the magnitudes of the 15N chemical shift tensors of
the two sites by fitting the spinning sideband envelopes recorded
at different spinning frequencies.18 These CSA parameters can
then be used for simulations of rotary resonance lineshapes
recorded at high spinning speeds without further iteration.
Fig. 2 and 3 show the rotary resonance lineshapes of sites I
and II, obtained by Fourier transformation of 15N signals
recorded as a function of the spin-lock duration.
To simulate the simultaneous recoupling of the dipolar and
chemical shift interactions, one needs to take into account the
(unknown) orientation of the 15N chemical shift tensor relative
to the 15N–127I vector.
The lineshapes were simulated assuming two different sce-
narios. The blue lines were calculated while neglecting
15N–127I dipolar interactions, retaining only the experimen-
tally determined anisotropy and asymmetry of the 15N chemi-
cal shift tensor (see Table 1). Although these simulated
lineshapes appear to reproduce the main splitting, some strik-
ing differences with respect to the experimental lineshapes are
clearly visible. The incorporation of dipolar interactions be-
tween 15N and 127I nuclei leads to much better fits (red lines).
Since there is no hint of 15N–15N homonuclear recoupling at
Since the principal values of the chemical shift tensor have
been determined before, only three variables remain to be
determined by iterative fitting. The relevant Euler angles and
internuclear distances were therefore varied in an iterative
fitting process. The simulations (see ESIw) show that the
spectral features are, as expected,14a,21 mostly sensitive to
the bCS angle and, to a lesser extent, to the aCS angle. The
parameters leading to the best agreement between the experi-
mental and simulated lineshapes are given in Table 1.
The magnitudes of the dipolar 15N–127I couplings obtained by
fitting correspond to internuclear distances of 2.7 and 2.9 A for
sites I and II, respectively. Both distances are shorter than the sum
of the van der Waals radii of N and I (3.36–3.68 A)22 and fit very
well with the known range (2.6–3.3 A) of halogen bond lengths.5
The higher intensity of the outer wings of the experimental
spectra compared with the simulated spectra may be due to
1
the n = rotary resonance condition, we can exclude any
2
contribution of this type to the observed features. It is also
worth pointing out that the quadrupolar–dipolar cross-terms
as well as the anisotropic J-couplings can be safely ignored at a
static magnetic field of 9.4 T,14a considering the intermolecular
character of the halogen bond.
ꢀc
This journal is The Royal Society of Chemistry 2008
5982 | Chem. Commun., 2008, 5981–5983

View Article Online
´
SESAME Program of the Region Ile-de-France, the Integrated
Infrastructure Initiative (Contract # RII3-026145, EU-NMR)
of the 6th Framework Program of the EU, the CNRS and the
Ecole Normale Superieure for financial support.
´
Notes and references
1 (a) H. A. Bent, Chem. Rev., 1968, 68, 587; (b) A. C. Legon, Angew.
Chem., Int. Ed., 1999, 38, 2687.
2 P. Metrangolo, H. Neukirch, T. Pilati and G. Resnati, Acc. Chem.
Res., 2005, 38, 386 and references therein.
3 (a) O. Hassel, Science, 1970, 170, 497–502; (b) G. R. Desiraju,
Angew. Chem., Int. Ed. Engl., 1995, 34, 2311; (c) J. P. M.
Lommerse, A. J. Stone, R. Taylor and F. H. Allen, J. Am. Chem.
Soc., 1996, 118, 3108; (d) P. Metrangolo and G. Resnati,
Chem.–Eur. J., 2001, 7, 2511; (e) J.-L. Syssa-Magale, K. Boube-
´
keur, P. Palvadeau, A. Meerschaut and B. Schollhorn, CrystEng-
¨
Comm, 2005, 7, 302.
4 G. A. Jeffrey, An Introduction to Hydrogen Bonding, Oxford
University Press, Oxford, 1997.
5 G. Resnati, T. Pilati, R. Liantonio and F. Meyer, J. Polym. Sci.,
Part A: Polym. Chem., 2007, 45, 1 and references therein.
6 P. Auffinger, F. A. Hays, E. Westhof and P. Shing Ho, Proc. Natl.
Acad. Sci. U. S. A., 2004, 16789, and references therein.
7 Starting from potassium phthalate containing 98% of 15N the
enriched benzylamine could be obtained as described in a modified
procedure of the Gabriel Synthesis. (H. R. Ing and R. H. F.
Manske, J. Chem. Soc., 1926, 2348). The following reductive
amination of 4-iodobenzaldehyde yielded compound 1. For further
details see ESIw.
Fig. 3 Experimental and simulated o1/2p projections from site II
(upfield signal in Fig. 1) of compound 1. All experimental conditions
and simulations were the same as those given in the caption of
Fig. 2.
¨
8 J. Leroy, B. Schollhorn, J.-L. Syssa-Magale, K. Boubekeur and P.
´
Palvadeau, J. Fluorine Chem., 2004, 125, 1379.
9 (a) M. T. Messina, P. Metrangolo, W. Panzeri, E. Ragg and G.
Resnati, Tetrahedron Lett., 1998, 9069; (b) P. Metrangolo, W.
Panzeri, F. Recupero and G. Resnati, J. Fluorine Chem., 2002, 114,
27; (c) R. Glaser, N. Chen, H. Wu, N. Knotts and M. Kaupp, J.
Am. Chem. Soc., 2004, 126, 4412.
10 T. G. Oas, R. G. Griffin and M. H. Levitt, J. Chem. Phys., 1988,
89, 692.
11 D. L. Bryce and G. D. Sward, Magn. Reson. Chem., 2006, 44, 409.
12 Z. Gan, D. M. Grant and R. R. Ernst, Chem. Phys. Lett., 1996,
254, 349.
Table 1 Magnitudes and relative orientations of the 15N chemical
shift and 15N–127I dipolar coupling tensors along with corresponding
internuclear N–I distances for the two distinguishable nitrogen sites in
compound 1
Site Dd/Hza
Zb
dNÁ Á ÁI/Hz rNÁ Á ÁI/A
aCS (1) bCS (1)
13 L. Duma, D. Abergel, P. Tekely and G. Bodenhausen, Chem.
Commun., 2008, 2361.
14 (a) L. Odgaard, M. Bak, H. J. Jakobsen and N. C. Nielsen, J.
Magn. Reson., 2001, 148, 298–308; (b) Z. Gan, Chem. Commun.,
2006, 4712.
I
À475 Æ 20 0.67 Æ 0.1 127 Æ 5 2.7 Æ 0.04 0.0 Æ 15 23.0 Æ 2
II À512 Æ 20 0.72 Æ 0.1 100 Æ 5 2.9 Æ 0.05 0.0 Æ 15 25.0 Æ 2
a
b
The anisotropy is defined as Dd = d33 À diso
.
The asymmetry as
15 M. Bechmann, S. Dusold, H. Forster, U. Haeberlen, T. Lis, A.
¨
Z = (d22 À d11)/Dd with d11 Z d22 Z d33
.
Sebald and M. Stumber, Mol. Phys., 2000, 98, 605.
16 S. Dusold and A. Sebald, Mol. Phys., 1998, 95, 1237.
17 P. J. Chu, J. H. Lunsford and D. J. Zalewski, J. Magn. Reson.,
1990, 87, 68.
18 (a) C. Gardiennet-Doucet, B. Henry and P. Tekely, Prog. Nucl.
Magn. Reson. Spectrosc., 2006, 49, 129–149; (b) C. Gardiennet, B.
Henry, P. Kuad, B. Spiess and P. Tekely, Chem. Commun., 2005,
180–18; (c) B. Henry, P. Tekely and J. J. Delpuech, J. Am. Chem.
Soc., 2002, 124, 2025–2034; (d) C. Gardiennet-Doucet, X. Assfeld,
B. Henry and P. Tekely, J. Phys. Chem. A, 2006, 110, 9137.
19 M. Bak, J. T. Rasmussen and N. C. Nielsen, J. Magn. Reson.,
2000, 147, 296.
20 (a) P. Tekely, P. Palmas and D. Canet, J. Magn. Reson., 1994,
A107, 129–133; (b) A. Detken, E. H. Hardy, M. Ernst and B. H.
Meier, Chem. Phys. Lett., 2002, 356, 298.
21 (a) C. Gardiennet, F. Marica, X. Assfeld and P. Tekely, Angew.
Chem., Int. Ed., 2004, 43, 3565–3568; (b) C. Gardiennet, F. Marica,
C. A. Fyfe and P. Tekely, J. Chem. Phys., 2005, 122, 054705.
22 (a) A. Bondi, J. Phys. Chem., 1964, 68, 441; (b) S. C. Nyburg and
C. H. Faerman, Acta Crystallogr., Sect. B: Struct. Sci., 1985, 41,
274.
transient oscillations of the 15N magnetization at the begin-
ning of the rotary resonance period, after the sudden reduction
of the RF field amplitude following cross-polarization.
In summary, we have shown that solid-state NMR permits
the determination of the lengths of non-covalent halogen
bonds. The new spectroscopic fingerprint of halogen bonding
reported in this work should be useful to visualize inter-
molecular contacts in halogen bond-driven self-assembling
aggregates. The approach used is not restricted to isolated
heteronuclear spin pairs and can be extended to larger spin
systems, although these will require extensive numerical
simulations.
The authors are grateful to Fabien Ferrage for stimulating
discussions. We acknowledge the ANR-05-BLAN-0255, the
ꢀc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 5981–5983 | 5983