Practical crystal engineering using halogen bonding: A hierarchy based on calculated molecular electrostatic potential surfaces

Article abstract of DOI:10.1016/j.molstruc.2014.02.022

A series of co-crystallization experiments were performed using four multi-topic N-heterocyclic acceptor molecules and nine aromatic halogen-bond donors in order to establish how effectively a ranking of bond strength based on calculated molecular electrostatic potential surfaces translates into predictable primary interactions in the solid state. A total of ten new crystal structures were obtained, and in each case, the observed interaction took place between the best acceptor (with the larger negative electrostatic potential) on the N-heterocycle and the halogen-bond donor. The supramolecular yield (number of successful co-crystallizations) is 70% for iodine-donors whereas none of the bromo-substituted donors produced a co-crystal which underscores the importance of the magnitude of the electrostatic potential and of the polarizability of the halogen-bond donor in the context of successful practical crystal engineering.

Full text of DOI:10.1016/j.molstruc.2014.02.022

Journal of Molecular Structure xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Journal of Molecular Structure
journal homepage: www.elsevier.com/locate/molstruc
Practical crystal engineering using halogen bonding: A hierarchy based
on calculated molecular electrostatic potential surfaces
⇑
Christer B. Aakeröy , Tharanga K. Wijethunga, John Desper
Department of Chemistry, Kansas State University, Manhattan, KS 66506, USA
h i g h l i g h t s
ꢀ
ꢀ
ꢀ
ꢀ
36 co-crystallizations on compounds with two different acceptors were performed.
An interaction hierarchy was established using calculated electrostatic potentials.
All 10 structures obtained display best donor/acceptor halogen bonds.
Halogen bonds with iodine are more effective than the bromo-analogues.
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online xxxx
A series of co-crystallization experiments were performed using four multi-topic N-heterocyclic acceptor
molecules and nine aromatic halogen-bond donors in order to establish how effectively a ranking of bond
strength based on calculated molecular electrostatic potential surfaces translates into predictable pri-
mary interactions in the solid state. A total of ten new crystal structures were obtained, and in each case,
the observed interaction took place between the best acceptor (with the larger negative electrostatic
potential) on the N-heterocycle and the halogen-bond donor. The supramolecular yield (number of suc-
cessful co-crystallizations) is 70% for iodine-donors whereas none of the bromo-substituted donors pro-
duced a co-crystal which underscores the importance of the magnitude of the electrostatic potential and
of the polarizability of the halogen-bond donor in the context of successful practical crystal engineering.
Ó 2014 Elsevier B.V. All rights reserved.
Keywords:
Halogen bond
Crystal engineering
Molecular recognition
Electrostatic potential
Co-crystallization
Hierarchy
1. Introduction
electrons, a convention which is adopted to align it with the gener-
ally accepted definition of the hydrogen bond [7,8] where the elec-
Intermolecular interactions are responsible for all molecular
recognition events, and as such represent the primary tools in
supramolecular chemistry and crystal engineering [1]. Conse-
quently, it is essential to have solid understanding of the funda-
mental nature of these interactions in order to successfully
design complex supramolecules in a predetermined and effective
manner [2–4]. Halogen bonds represent a relatively recent addition
to the tool box of supramolecular chemistry and still receive con-
siderable attention [5]. According to IUPAC, ‘‘A halogen bond
RAXꢁ ꢁ ꢁYAZ occurs when there is evidence of a net attractive inter-
action between an electrophilic region on a halogen atom X
belonging to a molecule or a molecular fragment RAX (where R
can be another atom, including X, or a group of atoms) and a nucle-
ophilic region of a molecule, or molecular fragment, YAZ00 [6].
According to this definition the halogen-bond donor is accepting
tropositive hydrogen atom is recognized as the hydrogen-bond
donor, accepting electrons from an electronegative atom [9]. The
halogen bonding ability of donor atoms increase in the order of
F ꢂ Cl ꢂ Br < I depending on the polarizability [10], and conven-
tional halogen bonds are highly directional [11]. Generally the
bond distance between acceptor and donor atoms in a halogen
bond is significantly shorter than the sum of van der Waals radii
[12]. In addition to conventional stabilizing halogen bonds, two
additional halogenꢁ ꢁ ꢁhalogen contacts, classified as type I and type
II (depending on the geometry) are frequently encountered in crys-
tal structures of halogen-substituted molecules, Scheme 1 [13].
Factors that influence the strength of a halogen-bond include the
presence of electron-withdrawing substituents that serve to
‘activate’ the halogen-bond donor atom by depleting it of electron
density thereby increasing its partial positive potential [14,15].
In order to develop reliable and transferable synthetic
guidelines for the assembly of molecular solids using multiple
intermolecular interactions, Etter [16] and others [17–19] have
⇑
Corresponding author. Tel.: +1 785 532 6096; fax: +1 785 532 6666.
E-mail address: aakeroy@ksu.edu (C.B. Aakeröy).
http://dx.doi.org/10.1016/j.molstruc.2014.02.022
0022-2860/Ó 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: C.B. Aakeröy et al., J. Mol. Struct. (2014), http://dx.doi.org/10.1016/j.molstruc.2014.02.022

2
C.B. Aakeröy et al. / Journal of Molecular Structure xxx (2014) xxx–xxx
(a)
(c)
(b)
Scheme 1. (a) Conventional halogen bond (b) Type I halogenꢁ ꢁ ꢁhalogen contact (c) Type II halogenꢁ ꢁ ꢁhalogen contact; (D & Y-connected atoms, X-halogen bond donor atom,
Y-halogen bond acceptor).
demonstrated that in hydrogen-bond based systems, there is a ten-
dency for the best-donors to bind to the best-acceptors, and the
second-best donor to bind to the second-best acceptor. If the
molecules carry the same functionality, a relative ranking can be
established reasonably well using pKa/pKb values [20–22] but if
different functionalities are employed, an approach to ranking
based on calculated molecular electrostatic potential surfaces is re-
quired [23–25].
trile (20 mL) was added to the mixture and refluxed for 24 h at 50–
60 °C. The reaction was monitored with TLC and after completion
the solvent was removed by rotary evaporation. The residue was
dissolved in water (50 mL) and extracted with methylene chloride
(30 mL ꢃ 3). Organic layers were combined, dried over anhydrous
MgSO4 and rotary evaporated to obtain the dark brown color pow-
der as the product. Yield: 0.35 g (56%); mp 157–160 °C; 1H NMR
(dH; CDCl3, 400 MHz):8.49 (d, 4H), 7.11 (d, 2H), 6.94 (d, 2H), 6.91
(d, 4H), 5.84 (s, 4H).
At this point, it is not clear how effectively an electrostatic po-
tential-based ranking of halogen-bond acceptors can be translated
directly into strategies for practical crystal engineering, although a
2.2. Synthesis of 1,10-bis(pyridin-3-ylmethyl)-2,2’-biimidazole, A2
halogen bond donor based hierarchy [26], a basicity scale [27], 19
F
NMR based studies [28], theoretical electrostatic based studies
[5,29] and solution based models [30] have been used to explain
halogen-bond interactions. In order to establish to what extent hal-
ogen bonds follow best-donor/best-acceptor guidelines, we
decided to carry out systematic co-crystallizations on a series of
N-heterocyclic halogen-bond acceptors, each with two binding
sites with different electrostatic potential, Scheme 2, A1–A3. In
addition, to ensure that the imidazole nitrogen atom was not inac-
cessible due to some steric hindrance, we also included ligand A4,
with essentially the same shape, but with only one type of acceptor
site. These four compounds have been co-crystallized with six
iodo-substituted halogen-bond donors, D5–D10, (all but one of
them ‘activated’ with fluorine groups) and three bromo-substi-
tuted halogen-bond donors, D11–D13, all of which are activated
by a fluorinated aromatic backbone, Scheme 2. The study is under-
taken in response to two hypotheses; (i) if a halogen-bond donor
has a choice of two accessible halogen-bond acceptors, it will pref-
erentially select the best-acceptor as determined by molecular
electrostatic potentials surfaces (MEPS) and (ii) a bromo-substi-
tuted halogen-bond donor will be less successful at forming co-
crystals than the corresponding iodo-analogues. The reactants
were combined using solvent-assisted grinding, and the products
were characterized using infrared spectroscopy. Successful experi-
ments were then subjected to a variety of crystal-growth experi-
ments and a total of ten samples produced crystals suitable for
single-crystal diffraction.
2,20-Biimidazole (0.27 g, 2 mmol) and NaOH (0.32 g, 8 mmol)
were placed in a 100 mL round bottomed flask with 20 mL of ace-
tonitrile. The mixture was stirred at room temperature for two
hours. 3-Picolyl chloride hydrochloride (0.65 g, 4 mmol) in acetoni-
trile (20 mL) was added to the mixture and refluxed for 24 h at 50–
60 °C. The reaction was monitored with TLC and upon completion
the solvent was removed by rotary evaporation. The residue was
dissolved in water (50 mL) and extracted with methylene chloride
(30 mL ꢃ 3). Organic layers were combined, dried over anhydrous
MgSO4 and rotary evaporated to obtain the brown color powder
as the product. Yield: 0.45 g (71%); mp 112–115 °C; 1H NMR (dH;
CDCl3, 400 MHz):8.46 (d, 2H), 8.45 (s, 2H), 7.39 (d, 2H), 7.17 (m,
2H), 7.11 (d, 2H), 6.95 (d, 2H), 5.78 (s, 4H).
2.3. Synthesis of 1,10-bis(pyridin-2-ylmethyl)-2,2’-biimidazole, A3
2,20-Biimidazole (0.27 g, 2 mmol) and NaOH (0.32 g, 8 mmol)
were placed in a 100 mL round bottomed flask with 20 mL of ace-
tonitrile. The mixture was stirred at room temperature for two
hours. 2-Picolyl chloride hydrochloride (0.65 g, 4 mmol) in acetoni-
trile (20 mL) was added to the mixture and refluxed for 24 h at 50–
60 °C. The reaction was monitored with TLC and upon completion
the solvent was removed by rotary evaporation. The residue was
dissolved in water (50 mL) and extracted with methylene chloride
(30 mL ꢃ 3). Organic layers were combined, dried over anhydrous
MgSO4 and rotary evaporated to obtain the pale brown color pow-
der as the product. Yield: 0.25 g (40%); mp 180–183 °C; 1H NMR
(dH; CDCl3, 400 MHz): 8.53 (d, 2H), 7.53 (t, 2H), 7.15 (t, 2H), 7.12
(s, 2H), 7.07 (s, 2H), 7.05 (d, 2H), 5.87 (s, 4H).
2. Experimental
All the reagents, solvents, and donors D1–D9 and D11–D13
were purchased from commercial sources and used as received.
2,20-Biimidazole was synthesized according to previously reported
methods [31]. Donor D10 was prepared according to the synthetic
methods reported [32]. A Fisher-Johns melting point apparatus
was used to determine melting points. Infrared spectra were re-
corded with a Nicolet 380 FT-IR. 1H NMR spectra were recorded
using a Varian Unity plus 400 MHz spectrometer.
2.4. Synthesis of 1,10-dibenzyl-2,20-biimidazole, A4
2,20-Biimidazole (0.33 g, 2.48 mmol) and NaOH (0.39 g,
9.92 mmol) were placed in a 100 mL round bottomed flask with
20 mL of acetonitrile. The mixture was stirred at room temperature
for two hours. Benzyl bromide (0.63 g, 5 mmol) in acetonitrile
(20 mL) was added to the mixture and refluxed for 24 h at 50–
60 °C. The reaction was monitored with TLC and after completion
the solvent was removed by rotary evaporation. The residue was
dissolved in water (50 mL) and extracted with methylene chloride
(30 mL ꢃ 3). Organic layers were combined, dried over anhydrous
MgSO4 and rotary evaporated to obtain the yellow color powder
as the product. Yield: 0.69 g (89%); mp 144–146 °C; 1H NMR (dH;
2.1. Synthesis of 1,10-bis(pyridin-4-ylmethyl)-2,20-biimidazole, A1
2,20-Biimidazole (0.27 g, 2 mmol) and NaOH (0.32 g, 8 mmol)
were placed in a 100 mL round bottomed flask with 20 mL of ace-
tonitrile. The mixture was stirred at room temperature for two
hours. 4-Picolyl chloride hydrochloride (0.65 g, 4 mmol) in acetoni-
Please cite this article in press as: C.B. Aakeröy et al., J. Mol. Struct. (2014), http://dx.doi.org/10.1016/j.molstruc.2014.02.022

C.B. Aakeröy et al. / Journal of Molecular Structure xxx (2014) xxx–xxx
3
I
I
I
I
I
I
F
F
F
F
I
F
F
F
F
F
I
F
I
F
F
F
F
F
F
F
F
F
F
F
I
F
F
F
I
D5
D6
D7
D8
D9
D10
Br
Br
Br
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
Br
F
F
Br
D13
D11
D12
Scheme 2. Halogen-bond acceptors (A1–A4) and halogen-bond donors (D5–D13) in this study.
Table 1
Values of electrostatic potentials of acceptors and donors.
Molecule
Atom
(kJ/mol)
Acceptors
1,10-Bis(pyridin-4-ylmethyl)-2,20-biimidazole A1
Pyridine N
Imidazole N
Pyridine N
Imidazole N
Pyridine N
Imidazole N
Imidazole N
ꢄ187
ꢄ132
ꢄ182
ꢄ128
ꢄ153
ꢄ125
ꢄ149
1,10-Bis(pyridin-3-ylmethyl)-2,20-biimidazole A2
1,10-Bis(pyridin-2-ylmethyl)-2,20-biimidazole A3
1,10-Bibenzyl-2,20-biimidazole A4
Donors
Iodobenzene D5
Iodopentafluorobenzene D6
Iodine
Iodine
Iodine
Iodine
Iodine
Iodine
Bromine
Bromine
Bromine
+103
+166
+162
+169
+158
n/a
+143
+139
n/a
1,2-Diiodotetrafluorobenzene D7
1,4-Diiotetrafluorobenzene D8
1,3,5-Triiodotrifluorobenzene D9
4,40-Diiodoperfluorobiphenyl D10
Bromopentafluorobenzene D11
1,4-Diiodotetrafluorobenzene D12
4,40-Dibromoperfluorobiphenyl D13
CDCl3, 400 MHz):5.70 (s, 4H), 6.93 (d, 2H), 7.03 (m, 4H), 7.12 (d,
2H), 7.24 (m, 6H).
using 6-311++G** basis set in vacuum. Calculations for D10 and
D13 could not be done using this level of theory. All calculations
were carried out using Spartan 8 software.
2.5. Molecular electrostatic potential calculations
2.6. Grinding experiments
Electrostatic potentials on the acceptor molecules were
calculated with density functional B3LYP level of theory with 6-
31G* basis set in vacuum. Electrostatic potentials on donor mole-
cules were calculated with density functional B3LYP level of theory
The initial screening was carried out using solvent-assisted
grinding (methanol). Each combination of acceptors and donors
were mixed in specific stoichiometries (see Supplementary
Please cite this article in press as: C.B. Aakeröy et al., J. Mol. Struct. (2014), http://dx.doi.org/10.1016/j.molstruc.2014.02.022

4
C.B. Aakeröy et al. / Journal of Molecular Structure xxx (2014) xxx–xxx
Table 2
Halogen-bond geometries.
Co-crystal
A1D7
CAIꢁ ꢁ ꢁN
Nꢁ ꢁ ꢁI (Å)
Nꢁ ꢁ ꢁIAC (0)
177.06(6)
175.98(5)
173.45(5)
C(31)AI(1)ꢁ ꢁ ꢁN(21)
C(31)AI(1)ꢁ ꢁ ꢁN(11)
C(31)AI(1)ꢁ ꢁ ꢁN(21)
C(31)AI(1)ꢁ ꢁ ꢁN(21)
C(32)AI(2)ꢁ ꢁ ꢁN(21)
2.7777(18)
2.7913(14)
2.7747(13)
A1D8
A1D10
A2D7
2.848(5)
3.255(5)
174.39(19)
165.16(18)
A2D8
A2D9
C(31)AI(1)ꢁ ꢁ ꢁN(21)
2.8598(13)
175.13(5)
C(312)AI(42)ꢁ ꢁ ꢁN(212)
C(311)AI(11)ꢁ ꢁ ꢁN(211)
C(332)AI(52)ꢁ ꢁ ꢁN(132)
2.757(2)
2.830(2)
3.123(2)
176.06(8)
176.70(8)
164.81(7)
A2D10
C(31)AI(1)ꢁ ꢁ ꢁN(21)
C(41)AI(2)ꢁ ꢁ ꢁN(13)
2.824(2)
3.1345(19)
175.93(7)
166.07(7)
A3D9
C(33)AI(2)ꢁ ꢁ ꢁN(21)
2.887(3)
178.09(10)
A3D10
C(31)AI(1)ꢁ ꢁ ꢁN(21)
C(41)AI(2)ꢁ ꢁ ꢁN(13)
2.8813(16)
3.3015(17)
175.87(6)
164.27(6)
A4D9
C(511)AI(11)ꢁ ꢁ ꢁN(131)
C(512)AI(42)ꢁ ꢁ ꢁN(132)
2.897(4)
3.028(4)
174.13(16)
171.17(19)
Fig. 1. Primary halogen bond interactions in the co-crystal A1D7.
crystal X-ray diffraction. Experimental details for grinding and
slow evaporations with specific IR shifts are recorded in the ESI.
2.8. X-ray crystallography
Datasets were collected at 120 K on a Bruker Kappa APEX II sys-
tem with Cu radiation (A2D7) or a Bruker SMART APEX II system
with Mo radiation at 120 K (the remainder) using APEX2 software.
An Oxford Croystream 700 low-temperature device was used to
control temperature. MoK radiation was used. Initial cell constants
were found by small widely separated ‘‘matrix’’ runs. Data collec-
tion strategies were determined using COSMO. Scan speeds and
scan widths were chosen based on scattering power and peak rock-
ing curves. Unit cell constants and orientation matrices were im-
proved by least-squares refinement of reflections thresholded
from the entire dataset. Integrations were performed with SAINT,
using these improved unit cells as a starting point. Precise unit cell
constants were calculated in SAINT from the final merged datasets.
Lorenz and polarization corrections were applied. Absorption cor-
rections were applied using SADABS. Datasets were reduced with
SHELXTL. The structures were solved by direct methods without
incident. All hydrogen atoms were assigned to idealized positions
and were allowed to ride. Isotropic thermal parameters for the
hydrogen atoms were constrained to be 1.5ꢃ (methyl)/1.2ꢃ (all
other) that of the connected atom. CCDC 966785–966794 contains
the supplementary crystallographic data for this paper. These data
can be obtained free of charge via http://www.ccdc.cam.ac.uk/con-
ts/retrieving.html (or from the Cambridge Crystallographic Data
Centre, 12, Union Road, Cambridge CB2 1EZ, UK; fax:+44 1223
336033).
Fig. 2. Secondary interactions in co-crystal A1D7.
Information for details) and the solid resulting from each reaction
was characterized by IR spectroscopy. Successful interactions be-
tween the acceptor and donor were identified using the specific
shifts of the peaks of the mixture compared to starting compounds.
2.7. Synthesis of co-crystals
The resulting mixtures used in the grinding experiments were
dissolved in a minimum amount of methanol and kept in a vial
for slow evaporation in order to obtain co-crystals suitable for sin-
gle crystal X-ray diffraction (see Supplementary Information for
details). Once the crystals were obtained they were again analyzed
using IR spectroscopy and melting point and subjected to single-
Please cite this article in press as: C.B. Aakeröy et al., J. Mol. Struct. (2014), http://dx.doi.org/10.1016/j.molstruc.2014.02.022

C.B. Aakeröy et al. / Journal of Molecular Structure xxx (2014) xxx–xxx
5
atom is involved in a Type II close contact with an iodine atom
from a neighboring donor molecule, Fig. 2.
3. Results
The ranking of acceptor sites was based on the calculated
molecular electrostatic surface potentials, Table 1. Once the ini-
tial screen was completed using IR spectroscopy, ten co-crystals
produced crystals suitable for single crystal X-ray diffraction.
The experimental crystallographic data are found in the ESI
(Table S3) and the halogen-bond geometries are listed in
Table 2.
Ten crystal structures were obtained (A1D7, A1D8, A1D10,
A2D7, A2D8, A2D9, A2D10, A3D9, A3D10 and A4D9) and subse-
quently analyzed to determine binding preferences between do-
nors and ditopic acceptors.
The crystal structure of A1D7 displays the expected 1:2 stoichi-
ometry where one acceptor molecule interacts with two donor
molecules forming a discrete trimer, Fig. 1. The primary halogen
bonds take place between one of the two iodine atoms and the
N-pyridine moiety (the best acceptor) whereas the second iodine
The crystal structure of A1D8 (with a 1:1 stoichiometry) con-
tains infinite 1-D chains constructed from Iꢁ ꢁ ꢁN(py) halogen bonds,
Fig. 3. The imidazole nitrogen atoms do not compete for halogen
bonds and instead form short contacts with CAH moieties of a
neighboring acceptor molecule, Fig. 4, resulting in a stacking of
adjacent chains, Fig. 4.
The co-crystal A1D10 has 1:2 stoichiometry (donor:acceptor)
and contains discrete trimers as the second iodine atom does not
participate in a halogen bond. Again, the best acceptor, N(py) is
the preferred binding site, Fig. 5. Unlike in the crystal structure
of A1D8, imidazole nitrogen atoms not participating in any notable
secondary interactions, while the second iodine of each donor mol-
ecule is not participating in any considerable interactions as well.
The crystal structure of A2D7 reveals a 1:1 stoichiometry. This
time, the best-acceptor, N(py) forms a bifurcated halogen bond,
Fig. 6, which in turn results in chains of tetrameric-rectangles.
Fig. 3. Formation of infinite chains via halogen bonds in A1D8.
Fig. 4. Secondary interactions in A1D8.
Fig. 5. A trimer in co-crystal A1D10.
Fig. 6. Primary halogen bond interaction in A2D7.
Please cite this article in press as: C.B. Aakeröy et al., J. Mol. Struct. (2014), http://dx.doi.org/10.1016/j.molstruc.2014.02.022

6
C.B. Aakeröy et al. / Journal of Molecular Structure xxx (2014) xxx–xxx
Fig. 7. Primary and secondary interactions in A2D8.
‘chains of hexamers’ where all four nitrogen atoms form halogen
bonds, Fig. 9. The third iodine atom of the donor molecule is inac-
tive (which is not uncommon) [33,34].
The crystal structure determination of A2D10 shows that the
donor and acceptor molecules appear in a 1:2 stoichiometry,
where all four nitrogen atoms are participating in halogen bonds
leading to infinite ribbons, Fig. 10.
The co-crystal A3D9 has a stoichiometry of 1:1 and contains 1-
D chains assembled using CAIꢁ ꢁ ꢁN(py) halogen bonds, Fig. 11. The
imidazole nitrogen atoms and third iodine atom do not have any
significant short contacts.
The structure determination of A3D10 shows that the acceptor
and donor are present in a 1:2 stoichiometry and, similarly to the
structure A2D10, contains chains of rectangles using all four nitro-
gen atoms on the acceptor, Fig. 12.
In the 1:1 co-crystal of A4D9, the ‘arms’ of the acceptor are,
unexpectedly, both appearing on the same face of the aromatic
core of A4. In addition, one of the N(im) nitrogen atoms form a con-
ventional, near-linear halogen bond, whereas the other nitrogen
atom forms a bifurcated halogen bond with two adjacent donor
molecules. Interestingly, all three iodine atoms of the donor partic-
ipate in halogen bonds which is unique in this series of crystal
structures, Fig. 13.
Fig. 8. A halogen-bond based trimer and secondary CAHꢁ ꢁ ꢁN(im) interactions in
A2D9.
No notable secondary interactions are found with the imiadozle
nitrogen atoms.
The crystal structure of A2D8 has a 1:1 stoichiometry and con-
tains 1-D chains and CAHꢁ ꢁ ꢁN(im) contacts similarly to what was
observed in the structure of A1D8, Fig. 7.
4. Discussion
The crystal structure of A2D9 also displays a 1:1 stoichiometry
and two distinct halogen-bond motifs are present. First, a trimer
assembled via a Iꢁ ꢁ ꢁN(py) (best-acceptor) halogen bond, Fig. 8,
and second, intermolecular CAHꢁ ꢁ ꢁN(im) hydrogen bonds that con-
nect neighboring acceptor molecules. The latter event leads to
Thirty-six co-crystallizations involving four bis-pyridine/bis-
imidazole based acceptor molecules and nine aromatic halogen-
bond donors were performed in order to determine to what extent
co-crystallizations driven by halogen-bond interactions follow the
same best-donor/best-acceptor rules (where the ranking is based
Fig. 9. A chain of hexagons in the crystal structure of A2D9.
Fig. 10. An infinite ribbon in A2D10.
Please cite this article in press as: C.B. Aakeröy et al., J. Mol. Struct. (2014), http://dx.doi.org/10.1016/j.molstruc.2014.02.022

C.B. Aakeröy et al. / Journal of Molecular Structure xxx (2014) xxx–xxx
7
Fig. 11. Primary halogen bond interactions in co-crystal A3D9.
Fig. 12. Chain of rectangles in the structure of A3D10.
ber of experiments with positive results (co-crystal formation),
while supramolecular yields are the success percentage relative
to the total number of experiments in each category.
The electrostatic potentials associated with the N(py) atoms of
A1 and A2 are close, and their success rates are also similar. Upon
moving from A1 to A4 the magnitudes of the negative EPs on the
nitrogen atoms decrease and their success rates are also decreas-
ing. Although the number of data points is rather small, the trend
is consistent with previously reported results, [29] and show that
the magnitude of the electrostatic potential plays a key role when
determining the likelihood that a particular combination of donor
and acceptor sites will interact strongly enough to drive the forma-
tion of a co-crystal. There is a dramatic difference in halogen-bond
ability between iodo- and bromo substituted compounds which re-
flect the polarizability, MEPS and size/depth of the respective
r-
hole; the former have a 70% supramolecular yield, whereas none
of the bromo-substituted donors produced co-crystals. On the
other hand the control experiments with iodobenzene showed that
the ‘activation’ offered by electron-withdrawing fluorine substitu-
ents is crucial to the ability of a halogen atom to form halogen
bonds of real supramolecular relevance, again consistent with the-
ory [5].
Fig. 13. Halogen bond interactions in the structure of A4D9.
on electrostatics) as are hydrogen bonds. The three ditopic accep-
tors A1–A3 have two acceptor sites that differ primarily by their
electrostatic potentials; N(py) occupies a more negative site than
N(im) making it the better acceptor. In order to ensure that any ob-
served binding preferences were not simply the result of steric hin-
drance, we also included an acceptor A4 which only contains
imidazole sites in the same steric environment as in A1-A3. The
ranking and hierarchy of the acceptor sites was established using
DFT calculated MEPS, Table 1. The calculations indicate that the
pyridine nitrogen atom carries a higher partial negative electro-
static potential (EP) than the imidazole nitrogen atom and is thus
ranked as the ‘best’ acceptor site.
The grinding experiments showed a pronounced positive corre-
lation between the EP on the acceptor sites and the supramolecular
yield (the percentage of successful reactions) again indicating that
the magnitude of the electrostatic potential plays a major role in
ranking the relative strength and efficiency of competing halo-
gen-bond sites, Table 3. In Table 3, the success rates are the num-
An examination of all ten crystal structures clearly demon-
strate that, given a choice, a strong halogen-bond donor preferen-
tially interacts with an acceptor atom with a more negative
electrostatic potential, which shows that practical halogen-bond
selectivity follows a best-donor/best acceptor guideline. In the
three structures were the imidazole nitrogen also participated in
a halogen bond, A3D10, A2D10 and A2D9, the N(im)ꢁ ꢁ ꢁI distances
were substantially longer than the N(py)ꢁ ꢁ ꢁI length which gives an
indication that the former are less favorable than the latter. Fur-
thermore, an N(im)ꢁ ꢁ ꢁI halogen bond did not appear in structures
with A1, wherein the N(py) atom has a substantially more nega-
tive electrostatic potential, relative to the N(im) atom, than in
A2 and A3.
Acceptor A4 was used as a control experiment to ensure that
the imidazole nitrogen atoms are in fact sterically accessible for
halogen bond formation and, consequently, not less competitive
due to some form of steric hindrance. In cases when all four nitro-
gen atoms participate in halogen bonds, it has resulted in the
Please cite this article in press as: C.B. Aakeröy et al., J. Mol. Struct. (2014), http://dx.doi.org/10.1016/j.molstruc.2014.02.022

8
C.B. Aakeröy et al. / Journal of Molecular Structure xxx (2014) xxx–xxx
Table 3
Supramolecular yields.
A1
A2
A3
A4
EP (in kJ/mol)
Npyridine (ꢄ187)
Npyridine (ꢄ182)
Npyridine (ꢄ153)
–
Nimidazole (ꢄ132)
Nimidazole (ꢄ128)
Nimidazole (ꢄ125)
Nimidazole (ꢄ149)
Success rate – grinding
Success rate – slow evaporation
Total supramolecular yield
Success rate with iodine donors – grinding
Supramolecular yield with iodine donors
Total supramolecular yield = 47%
5/9
4/9
55%
5/6
83%
5/9
4/9
55%
5/6
83%
4/9
3/9
45%
4/6
67%
3/9
2/9
33%
3/6
50%
Total supamolecular yield with iodine donors = 70%
Total suparmolecular yield with bromine donors = 0%
formation architectures comprising polygonal architectures such
as rectangles (A3D10 and A2D10) and hexagonal motifs (A2D9).
In A2D7 and A4D9 the nitrogen atoms are participating in dis-
symmetric bifurcated halogen bonds, where one of the contacts
is shorter than the other. Such halogen bonds are rare, as nitrogen
atoms, with only one lone pair, tend to be satisfied with a single
electron-pair acceptor forming one non-covalent interaction only.
If we consider the donor molecules in these ten co-crystal struc-
tures various behaviors can be recognized. In A1D7, one of the io-
dine atoms of halogen bond donor is participating in forming a
Type II halogen bond where the electropositive tip of the iodine
atom is interacting with the electronegative region of the another
iodine atom. In A1D10, even though the two iodine atoms in donor
D10 should have the same initial strength (due to the symmetry of
the molecule), the second iodine atom is not participating in any
interactions leading to a trimer formation in this case. It should
be pointed out that it is not possible to fully rationalize or predict
every structural aspect of the solid state through an examination
that focuses on relatively strong and directional intermolecular
interactions such a hydrogen- and halogen bonds. Although these
forces are clearly capable of, and responsible for, the bringing to-
gether of different neutral molecular within the same crystalline
solid, but the detailed outcome of nucleation and crystallization
is also influenced by weaker non-covalent interactions.
References
[1] J.M. Lehn, Chem. Soc. Rev. 36 (2007) 151.
[2] J.B. Benedict, Recent Adv. Crystallogr. (2012). InTech.
[3] F. Meyer, P. Dubois, CrystEngComm 15 (2013) 3058.
ˇ ´
ˇ ˇ
[4] D. Cincic, T. Frišcic, W. Jones, Chem. Eur. J. 14 (2008) 747.
[5] (a) P. Metrangolo, F. Meyer, T. Pilati, G. Resnati, G. Terraneo, Angew. Chem. Int.
Ed. 47 (2008) 6114;
(b) M. Erdelyi, Chem. Soc. Rev. 41 (2012) 3547;
(c) T.M. Beale, M.G. Chudzinski, M.G. Sarwar, M.S. Taylor, Chem. Soc. Rev. 42
(2013) 1667;
(d) P. Politzer, J.S. Murray, ChemPhysChem 14 (2013) 278;
(e) P. Politzer, J.S. Murray, T. Clark, Phys. Chem. Chem. Phys. 15 (2013) 11178.
[6] G.R. Desiraju, S. Ho, L. Kloo, A.C. Legon, R. Marquardt, P. Metrangolo, P.A.
Politzer, G. Resnati, K. Rissanen, IUPAC Provisional Recommendation.
[7] A. Priimagi, G. Cavallo, P. Metrangolo, G. Resnati, Acc. Chem. Res. 46 (2013)
2686.
[8] S. Tothadi, G.R. Desiraju, Chem. Commun. 49 (2013) 7791.
[9] T. Steiner, Angew. Chem. Int. Ed. 41 (2002) 48.
[10] J. Lieffrig, O. Jeannin, A. Fra˛ckowiak, I. Olejniczak, R. Swietlik, S. Dahaoui, E.
Aubert, E. Espinosa, P.A. Senzier, M. Fourmigu, Chem. Eur. J. 19 (2013) 14804.
[11] P. Politzer, J.S. Murray, T. Clark, Phys. Chem. Chem. Phys. 12 (2010) 7748.
[12] P. Metrangolo, H. Neukirch, T. Pilati, G. Resnati, Acc. Chem. Res. 38 (2005) 386.
[13] G.R. Desiraju, R. Parthasarathy, J. Am. Chem. Soc. 111 (1989) 8725.
[14] P. Metrangolo, G. Resnati, Chem. Eur. J. 7 (2001) 2511.
[15] A. Bauzá, D. Quin~onero, A. Frontera, P.M. Deyà, Phys. Chem. Chem. Phys. 13
(2011) 20371.
[16] M.C. Etter, Acc. Chem. Res. 23 (1990) 120.
[17] J.R. Bowers, G.W. Hopkins, G.P.A. Yap, K.A. Wheeler, Cryst. Growth Des. 5
(2005) 727.
´
[18] C.B. Aakeröy, A.M. Beatty, B.A. Helfrich, Angew. Chem. Int. Ed. 40 (2001) 17.
[19] G.R Desiraju, J.J Vittal, A. Ramanan, Crystal Engineering: A Textbook, World
Scientific, 2011.
[20] D.E. Lynch, P. Sandhu, S. Parsons, Aust. J. Chem. 53 (2000) 383.
[21] C.B. Aakeröy, A.M. Beatty, B.A. Helfrich, Angew. Chem. Int. Ed. 17 (2001) 3240.
[22] K.S. Huang, D. Britton, M.C. Etter, S.R. Byrn, J. Mater. Chem. 7 (1997) 713.
[23] C.A. Hunter, Angew. Chem. Int. Ed. 43 (2004) 5310.
5. Conclusion
This study highlights the importance of using calculated molec-
ular electrostatic potentials as a way of ranking the relative
efficiency of specific halogen-bond interactions in solid-state
architectures where multiple binding options are available
between the reactants. In fact, halogen bonds seem to reflect the
previously established propensity of hydrogen-bonds for following
a best-donor best-acceptor interaction hierarchy in the solid state.
In all ten crystal structures that were obtained in this study, halo-
gen bonds involving the ‘‘better’’ acceptor are favored.
[24] P.W. Kenny, J. Chem. Inf. Model. 49 (2009) 1234.
[25] C.B. Aakeröy, K.N. Epa, F. Forbes, J. Desper, CrystEngComm 15 (2013) 5946.
[26] C.B. Aakeröy, M. Baldrighi, J. Desper, P. Metrangolo, G. Resnati, Chem. – A Eur. J.
19 (2013) 16240.
[27] J.Y.L. Questel, C. Laurence, J. Graton, CrystEngComm 15 (2013) 3212.
[28] M.G. Sarwar, B. Dragisic, L.J. Salsberg, C. Gouliaras, M.S. Taylor, J. Am. Chem.
Soc. 132 (2010) 1646.
[29] K.E. Riley, P. Hobza, Phys. Chem. Chem. Phys. 15 (2013) 17742.
[30] Y. Lu, H. Li, X. Zhu, W. Zhu, H. Liu, J. Phys. Chem. A 115 (2011) 4467.
[31] J. Xiao, J.M. Shreeve, J. Org. Chem. 8 (2005) 3072.
[32] G.M. Espallargas, A. Recuenco, F.M. Romero, L. Brammer, S. Libri,
CrystEngComm 14 (2012) 6381.
[33] C.B. Aakeröy, T.K. Wijethunga, J. Desper, CrystEngComm 16 (2014) 28–31.
[34] A.C.B. Lucassen, A. Karton, G. Leitus, L.J.W. Shimon, J.M.L. Martin, M.E. van der
Boom, Cryst. Growth Des. 7 (2007) 386.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.molstruc.2014.02.
022.
Please cite this article in press as: C.B. Aakeröy et al., J. Mol. Struct. (2014), http://dx.doi.org/10.1016/j.molstruc.2014.02.022