PCCP
View Article Online
Published on 15 November 2013. Downloaded by Bharathiar University on 31/07/2014 12:05:53.
PAPER
Cite this: Phys. Chem. Chem. Phys.,
2014, 16, 2038
View Journal | View Issue
Halogen bonds in crystal TTF derivatives: an
ab initio quantum mechanical study
P. Deepa,a B. Vijaya Pandiyan,b P. Kolandaivel*b and Pavel Hobza*ac
The stabilisation energies of five ionic and neutral organic crystal structures containing various halogen bonds
(I I, Br Br, I Br, I S and Br S) were calculated using the DFT-D3 method (B97D/def2-QZVP). Besides
them, the ionic I3 I2 and neutral I2 I2, complexes (in the crystal geometries) were also studied. The nature
of the bonds was deduced from the electrostatic potential evaluated for all subsystems. In almost all the cases,
the s-hole was positive; it was negative only for the ionic I3 system (although more positive than the
respective belt value). The strongest halogen bonds were those that involved iodine as a halogen-bond donor
and acceptor. Among ionic X I3 and neutral X I2 and X Y dimers, the neutral X I2 complexes were,
surprisingly enough, the most stable; the highest stabilisation energy of 13.8 kcal mol1 was found for
the I2 1,3-dithiole-2-thione-4-carboxylic acid complex. The stabilisation energies of the ionic I3 I2
and neutral I2 1,3-dithiole-2-thione-4-carboxylic acid (20.2 and 20.42 kcal mol1, respectively)
complexes are very high, which is explained by the favourable geometrical arrangement, allowing the
Received 20th September 2013,
Accepted 14th November 2013
DOI: 10.1039/c3cp53976h
formation of a strong halogen bond. An I I halogen bond also exists in the neutral I2 I2 complex,
having only moderate stabilisation energy (3.9 kcal mol1). This stabilisation energy was, however, shown
to be close to that in the optimal gas-phase L-shaped I2 I2 complex. In all the cases, the dispersion
energy is important and comparable to electrostatic energy. Only in strong halogen bonds (e.g. I3 I2),
www.rsc.org/pccp
the electrostatic energy becomes dominant.
Introduction
The presence of halogen in supramolecules, biological molecules
and organic crystals has received the attention of scientists.1–9
The halogens appearing on molecular surfaces actively take
part in molecular recognition processes. The halogen atoms
interacting with the enzyme protein and nucleic acids play an
important role in biology.10,11 Some halogens are associated
with human diseases as well. Even though natural products
lack halogen, halogens isolated from natural sources have been
added to various molecules, and this has become an important
pharmaceutical material for drug discovery. Many of the drugs
which are available on the market are halogenated due to their
secondary interactions in molecular recognition.12
Covalently bound halogens are expected to have a negative
charge, which is the reason that the hydrogen-bonding ability
of complexes containing these halogens has been discussed
in the literature.13–17 Recently, however, also the interactions of
a
Institute of Organic Chemistry and Biochemistry, Academy of Sciences of the Czech
Republic, Flemingovo nám. 2, 166 10 Prague 6, Czech Republic.
E-mail: [email protected]
b
Department of Physics, Bharathiar University, Coimbatore – 641 046, Tamil Nadu,
India. E-mail: [email protected]
c
Regional Centre of Advanced Technologies and Materials, Department of Physical
Chemistry, Palacky University, 771 46 Olomouc, Czech Republic
2038 | Phys. Chem. Chem. Phys., 2014, 16, 2038--2047
the RX YZ (X = Cl, Br, I; Y = O, N, S; YZ being a Lewis base)
type have frequently been studied. The electrostatic potential
around a halogen covalently bound to an electronegative atom
(e.g. C) possesses a region of positive charge on the extension of
a covalent RX bond which is called a s-hole.18–20 The remainder
of the halogen has generally a negative potential, and the RX
molecule can thus form a halogen bond with a YZ molecule,
and a hydrogen bond with a HZ molecule. The size of the s-hole
depends on the halogens and chemical environment. Among
the interactions mentioned above, hydrogen bonding is predominantly stronger than halogen bonding.15 The preferred
structure for all the interactions is linear or nearly linear. The
halogen–halogen contacts are sometimes referred to as the
dihalogen bond, but we call these interactions throughout
the manuscript halogen bonding.
The halogen bonds are shorter than the sum of the Van der
Waals radii of a halogen and an electron donor or two halogens.
Halogen bonding has been observed in many organic and
inorganic crystal structures.21,22 Batsanov et al.23 have synthesised the halogen-bonded tetrathiafulvalene (TTF)-derivative
crystal and reported the distances of all possible halogen bond
lengths such as X X, X X and X S (X = F, Cl, Br, I). Zordan
and Brammer24 have reported that the halogen-bond geometries
exhibit near-linear angles (157–1731) at the organic halogen
(C–X 0 ) and much smaller angles (85–1121) at the inorganic
This journal is © the Owner Societies 2014
View Article Online
Published on 15 November 2013. Downloaded by Bharathiar University on 31/07/2014 12:05:53.
Paper
halogen (M–X) (M = Pd, Pt; X = Cl, I; X 0 = Cl, Br, I), which is
consistent with the behaviour of these halogen environments as
electrophiles and nucleophiles, respectively. The stabilisation
energy of a halogen bond is to a great extent determined by
dispersion energy. The strength of the halogen-bonding interactions
decreases in the order of I > Br > Cl. The electron-density
distribution around the fluorine atom is spherical rather than
anisotropic; however, it participates in the halogen-bonding
interactions, if fluorine is bonded with an electron-withdrawing
atom or group like e.g. other fluorine. Other halogens such as I,
Br, Cl have a pronounced anisotropy of the electron density
around the halogen nucleus.25
We have made a survey of halogen-bonded structures in the
Cambridge crystal data centre. The halogen bond formed
between a bifurcated donor and a bifurcated acceptor is very
interesting in the crystal structure.21 The halogen bond has been
detected in the ionic complexes of o-dibromotetrathiafulvalene
(abbreviated as TTF(SMe)2Br2) with I3 (Fig. 1a), and I4TTF with
I2 I3 (Fig. 1b–d), as well as in the neutral complexes of I4TTF
tetramer (Fig. 1f), cyanoethyl-bromine TTF trimer (Fig. 1g), and
in the 1,3-dithiole-2-thione-4-carboxylic acid complex with I2
(Fig. 1e). Altogether, five halogen-bonded structures have been
chosen for this study. The TTF(SMe)2Br2 I3 complex has
I Br (3.85Å) and I S (3.96 Å) close contacts. The 1,3-dithiole2-thione-4-carboxylic acid I2 complex has I S and I I close
contacts of 3.72 Å and 3.44 Å, respectively.26 The I4TTF I2
complex has I I (o4.3 Å) and I S (o4.0 Å) close contacts.
It is to be noted that the I atom interacts simultaneously with
I and S. Similarly, the dimers of the cyanoethyl-bromine of TTF
have Br Br (3.57 Å) and Br S (3.68, 3.55, 3.74 Å) contacts.
Furthermore, each I4TTF dimer having an I I contact of 3.5 Å is
surrounded by an ionic I3 with close contacts of 4.2, 4.2 Å.
Finally, I2 interacts with I3 through an I I (3.37 Å) contact.19
Besides the complexes mentioned, we have also investigated the
crystal structures of I3 I2 and I2 I2. The former complex is
known to be one of the strongest halogen-bonded complexes
with stabilisation energy approaching 43 kcal mol1.27,28
In this work, we have analysed the structure and nature of
stabilisation in the crystals of tetrathiafulvalene (TTF) derivatives
containing iodine and bromine with I3 and I2. Most of the
halogen bonds observed in the crystal structures are close to the
ideal geometric arrangement with I–I Br, I–I S, C–S I, I–I I,
C–I I, C–Br Br and C–S Br angles greater than 1601 and the
X Y (X, Y = I, Br, S, C) distance smaller than the sum of the
Van der Waals radii. This sum for I I, I S, I Br, Br Br and
Br S amounts to 3.96, 3.78, 3.88, 3.80 and 3.70 Å, respectively.
The greater the difference between the X Y distance and the
sum of Van der Waals radii, the stronger the binding expected.
This work aims to determine the binding energies of a
variety of halogen bonds in organic crystal structures. Insight
into the nature of the stabilisation of these systems is obtained
by performing quantum mechanical calculations (QM). The
QM calculations describe halogen bonding and it’s characteristics in a straightforward way contrary to molecular mechanics
which fails. As the complexes investigated are rather extended
(up to 60 atoms), the high level calculations such as MP2 or
This journal is © the Owner Societies 2014
PCCP
CCSD(T) are very difficult to perform, hence we have selected
the DFT method which has been augmented with empirical
dispersion.29,30 The DFT-D3/B97D approach has been tested
extensively in the S66 data set31 with very good results. It yields
the lowest RMSE error among the GGA functionals tested and
its results are close to these obtained with most advanced
functionals.32 When passing from the S66 to X40 dataset which
contains 40 complexes containing halogens, the respective
error slightly increases (0.39 kcal mol1, 10% vs. 0.25 kcal mol1,
5% in S66) but the error is still acceptable.
DFT-SAPT analyses of the halogen bonds will provide the
relative contributions of dispersion, electrostatics, and induction to total attraction in each of the complexes. Unfortunately,
for systems having more than 60 atoms, because of computational demand it is extremely difficult to perform the above
analyses. Another complication associated with the use of
DFT-SAPT for large atoms is that the method cannot be used
with relativistic corrections and can often yield inaccurate results
when it is used along with pseudopotentials, which inherently
account for relativistic effects. With no reliable way to account for
the relativistic effect, the treatment of very large atoms, such as
iodine, will generally be deficient.4,8 Hence we have considered
dispersion contribution along with total interaction energy and
importance of both terms will be discussed. The similarity
between total interaction energy and dispersion energy does not
necessarily mean that the second energy is dominant. This will
happen also in the case when attractive electrostatic energy and
repulsive exchange–repulsion energy compensate each other.
Finally, we are aware of the fact that calculated dimer stabilization
energies which correspond to gas phase energies differ from those
in the crystal. To account for the crystal packing effect, periodic
boundary calculations should be preferred.33
Computational methods
The interaction energies for all the halogen-bonded complexes
have been computed using DFT-D3/B97D/def2-QZVP. Grimme’s
advanced dispersion-corrected approach27 (DFT-D3) with the
Becke–Johnson damping function28 in a large basis set (def2-QZVP)
has been chosen, because it provides reliable interaction energies
for different types of non-covalently bonded complexes.30 The
def2-QZVP basis set, which contains pseudopotentials and thus
implicitly describes relativistic effects, has been used to describe
all halogens. This is important especially for iodine, as the
relativistic effect plays a decisive role in the electronic behaviour.
The structures of all the complexes used in this study have
been derived from the Cambridge Crystal Structure Database
(CCSD). The small molecules obtained from the CCSD are measured
with high resolution and without any errors. Consequently, the
positions of heavy atoms are accurate. Hence, we took the crystal
structures without any optimisation for further study. The
hydrogens were added through MERCURY software34 and their
positions were optimised using the DFT-D3/B97D/def2-QZVP
method. Brief descriptions of the structures considered are
given in the introduction. Here, each system was labelled
Phys. Chem. Chem. Phys., 2014, 16, 2038--2047 | 2039
View Article Online
Paper
Published on 15 November 2013. Downloaded by Bharathiar University on 31/07/2014 12:05:53.
PCCP
Fig. 1 Organic crystal structures with halogen bond lengths (Å). For the sake of clarity the non-bonded distances (in parentheses) are also visualized.
(The crystal motifs have been shown for all the complexes, but dimers were considered for further study.)
according to the name of the crystal structure. In order to obtain
deeper insight into the nature of the stabilisation of halogen
2040 | Phys. Chem. Chem. Phys., 2014, 16, 2038--2047
bonding, we sometimes replaced heavier halogens with
lighter halogens or hydrogen. The hydrogen positions and the
This journal is © the Owner Societies 2014
View Article Online
Paper
PCCP
lighter-halogen positions along with the adjacent heavy atom
were optimised at the B97D/def2-QZVP level of theory. All interaction energies (DE(A,B); eqn (1)) were corrected for the basis set
superposition error (BSSE):
Published on 15 November 2013. Downloaded by Bharathiar University on 31/07/2014 12:05:53.
DE(A,B) = E(A,B) {E(A) + E(B)},
(1)
where E(AB) is the total energy of a dimer and E(A) and E(B) are
the total energies of all the monomers corrected for the
energies of the ghost monomers.
The electrostatic potential map has been generated for all the
monomers in order to gain insight into the nature and directionality of the halogen-bond interactions being considered here. The
electrostatic potentials have been computed on the molecular
surfaces, which were considered to be an outer contour of the
electron density, generally the 0.001 au (electrons per bohr3)
surface, as proposed by Bader et al. The most positive value of
the potentials (the local maximum) is referred to as VS,max. There
may be several of each on a given molecular surface. Here, the
electrostatic potentials were computed at the B97D/DGDZVP
level of theory.
Results and discussion
Monomers
Fig. 2 shows the electrostatic potential of all the monomers; the
VS,max values for halogens calculated at B97D/def2-QZVP are
given in Table 1 and the I3 system will be discussed first. The
s hole (negative in the I3 anion) on the peripheral iodine is less
negative (more positive), with the VS,max values of atoms 1 and 3
amounting to 0.109, 0.109, when compared to the cusp point
of atoms 1 and 3, whose VS,max values are 0.134 and 0.134.
The VS,max of the central iodine (the belt point) is slightly more
negative, namely 0.147. To acquire in-depth knowledge of
electrostatic potential in I3, we replaced the central iodine with
bromine (I2Br) or chlorine (I2Cl). The peripheral iodine atoms
are more negative than the central bromine or chlorine. The size
of the s hole (again negative) on the peripheral iodine slightly
increased, with the VS,max of atoms 1 and 3 amounting to 0.113
and 0.113 and to 0.112 and 0.112 for I2Br and I2Cl,
respectively, but those smaller than the cusp point of the same
atoms had the VS,max of 0.143, 0.143, and 0.149, 0.149,
respectively. The replacement of the central iodine by bromine
or chlorine influenced the size of the s hole. The peripheral
iodine atoms were also replaced by bromine (Br2I). The size of
the s hole increases, with the VS,max of atoms 1 and 3 amounting
to 0.129 and 0.129. It arises from a comparison of I3, I2Br,
I2Cl and Br2I that the halogen’s chemical environment
strongly affects the size of the halogen s hole.
Analysing the neutral system I2, atoms 1 and 2 have the
VS,max value of 0.053 and 0.053, where, as expected, the s hole
is positive. It should be mentioned here that the s holes on the
peripheral iodine atoms in I3 were negative.
The s hole of the bromine atoms (Br1, Br2) in the
TTF(SMe)2Br2 and cyanoethyl-bromine TTF systems are positive,
with the VS,max of atoms 1 and 2 amounting to 0.0255, 0.025 and
This journal is © the Owner Societies 2014
0.0256, 0.0256, respectively. Considering the s holes on the
iodine atoms are also positive, and the VS,max of all the atoms 1,
2, 3 and 4 amounts to 0.027.
The s holes on the iodine atoms in I4TTF and the bromine
atoms in TTF(SMe)2Br2 differ only negligibly. Upon replacing one
of the iodine atoms by bromine (Br I3TTF), the s holes on the
iodine atoms and bromine atoms remain practically unchanged
(the VS,max value of atoms 1, 2, 3 and 4 amounts to 0.0256).
Complexes
TTF(SMe)2Br2 I3. The structures of all the dimers considered are shown in Fig. 1 while their interaction energies along
with dispersion energies are depicted in Table 2. Dimer
TTF(SMe)2Br2 I3 (Fig. 1a, the AC and AB structures) will
be discussed first. The interaction energy of the AC dimer is
8.50 kcal mol1, which is comparable to the dispersion energy
(7.27 kcal mol1). A similar result holds for the dimer AB with
an interaction energy of 9.59 kcal mol1 and a dispersion
energy of 9.45 kcal mol1. This is a slightly surprising result,
which indicates that for the charged TTF(SMe)2Br2 I3 dimer
the dispersion contribution (and not the electrostatic contribution
term included in the DFT interaction energy) is similar to that of
interaction energy. An explanation is offered by a brief inspection
of Fig. 1a. The halogen bonds in the AB and AC structure do not
correspond to a typical halogen bond with attractive electrostatic
energy between an electron donor and halogen or between two
halogens. All of these electrostatic interactions in the AB and AC
structures are clearly repulsive.
The dimer AB with four halogen bonds (I S) is slightly more
stable than the dimer AC with four halogen bonds (I Br),
indicating that both the halogen bonds are equally strong.
It is to be noted that the intermolecular distances in the above
bonds are comparable.
In order to obtain deeper insight into the nature of the
stabilisation of the dimer AC, containing four halogen bonds, we
replaced both bromine atoms by hydrogen atoms. Specifically, four
halogen bonds (C41–Br37 I3, C41–Br37 I1, C31–Br26 I1,
C31–Br26 I2) were mutated to hydrogen bonds (C41–H37 I3,
C41–H37 I1, C31–H26 I1, C31–H26 I2). Upon these mutations,
stabilisation energy decreases, but not dramatically, from 8.50
to 6.57 kcal mol1. Interestingly enough, dispersion energy
decreases from 7.27 to 1.54 kcal mol1. The decrease of the
stabilisation energy (1.93 kcal mol1) is thus smaller than that of
the dispersion energy (5.73 kcal mol1). These numbers indicate
that the stabilisation energy in halogen bonds is dominated by
the dispersion energy (see above). This is not surprising, because
the heavier halogens with large polarisabilities are close to each
other, which makes the dispersion energy high. Furthermore, it
also shows that halogen bonds are more stable than the respective hydrogen bonds, but the difference is small.
I4TTF I3(I2). This system involves both charged I3 and
neutral I2 interacting with I4TTF (Fig. 1b–d). The complex
I4TTF I3 (AB in Fig. 1b) will be discussed first. The AB structure
having six halogen bonds has higher stabilisation and dispersion
energies (9.48 and 10.08 kcal mol1) than the AC structure
with only two halogen bonds (6.79 and 2.82 kcal mol1).
Phys. Chem. Chem. Phys., 2014, 16, 2038--2047 | 2041
View Article Online
Paper
Published on 15 November 2013. Downloaded by Bharathiar University on 31/07/2014 12:05:53.
PCCP
Fig. 2 Electrostatic potential for all the monomers with VS,max values at the cusp point and the belt point. Here blue indicates a positive region for
Fig. 1e–j and a negative region for Fig. 1a–d and red indicates a negative region.
Very high dispersion energy (even higher than interaction
energy) in the case of the dimer AB again implies that the
2042 | Phys. Chem. Chem. Phys., 2014, 16, 2038--2047
electrostatic interaction in halogen bonds is repulsive and
that all the attraction comes from the dispersion energy.
This journal is © the Owner Societies 2014
View Article Online
Paper
Table 1
PCCP
The electrostatic potential (VS,max) on halogens in all monomers
Monomer
I1
I2
I3
1.109
0.147
1.109
I2Br b
I1
Br2
I3
0.113
0.157
0.113
I2Cl c
I1
Cl2
I3
0.112
0.165
0.112
Br2I d
Br1
I2
Br3
0.129
0.152
0.129
I2 e
I1
I2
0.053
0.053
TTF(Sme)2Br2 f
Br1
Br2
0.0255
0.0255
I4TTFg
I1
I2
I3
I4
0.027
0.027
0.027
0.018
BrI3TTFh
Br1
I2
I3
I4
0.0256
0.0256
0.0256
0.0256
Cyanoethyl-bromine TTF dimeri
Br1
Br2
0.0256
0.0256
I3
Published on 15 November 2013. Downloaded by Bharathiar University on 31/07/2014 12:05:53.
VS,max
a
a
h
Fig. 2a. b Fig. 2b. c Fig. 2c.
Fig. 2i. i Fig. 2j.
d
Fig. 2d. e Fig. 2e. f Fig. 2f.
g
Fig. 2h.
The situation in the AC structure is different, and here the
halogen bond contains electrostatic attraction, coming from
the interaction energy between the positive s holes on the
iodine atoms in I4TTF and the negative iodine atoms in I3.
To analyse the importance of the iodine atoms in the dimer
AB, the iodine atoms in I4TTF were replaced by hydrogen
atoms, i.e. two halogen bonds (I14–I15 I18, I14–I13 I16)
were mutated to two hydrogen bonds (I14–I15 H18, I14–
I13 H16). Surprisingly enough, the stabilisation energy of
the resulting complex (9.83 kcal mol1) is slightly higher than
that of the original complex (9.48 kcal mol1). Similarly, the
replacement of heavier halogens in I3 by lighter ones, i.e. I13,
I15 - Br13, Br15 or Cl13, Cl15 and I14 - Br14 or Cl14 in I3,
resulted in a small stabilisation energy decrease (from 9.48 to
9.31, 9.11 and 8.93, 8.69 kcal mol1), respectively. This provides
evidence that the diiodine bonds in the original crystal do not
possess a favourable geometry. Indeed, the I14–I15 I18 and
I14–I13 I16 angles (116.491 and 116.491, respectively) are
far from the ideal range of 170–1801. The hydrogen bonds
(I–I H) resulting from the mutation of I to H are clearly not
as directional as the halogen bonds and bring some stabilisation. Performing the same mutation for the dimer AC (I20,
I21 - H20, H21), we obtained a significant stabilisation energy
decrease (6.79 to 3.51 kcal mol1), which unambiguously
indicates that both halogen bonds in this structure were highly
attractive. Indeed, the I14–I13 I20(I21) angles (167.401 and
150.711) are close to the ideal range.
The ionic I2 I3 I4TTF systems (Fig. 1c) consist of the
neutral dimer AB (I4TTF I2) and the ionic dimer AC (I4TTF I3).
Both dimers are surprisingly comparably stable (6.56 and
6.79 kcal mol1), but dispersion energy plays a different role
Table 2 The total interaction energy (kcal mol1) and dispersion energy (in parentheses) for all the complexes considered. Data are also presented for
the complexes where halogens are replaced by a hydrogen or other lighter halogens
I3 TTF(SMe)2Br2a
AC
AB
Br26,Br37 - H26,H37 (AC)
8.50 (7.27)
9.59 (9.45)
6.57 (1.54)
I3 I4TTF I4TTFb
BC
AB
AC
I18,I16 - H18,H16 (AB)
I20, I21 - H20,H21 (AC)
I14 - Br14 (AB)
I14 - Cl14 (AB)
I13,I15 - Br 13,Br15 (AB)
I13,I15 - Cl13,Cl15 (AB)
1.31
9.48
6.79
9.83
3.51
8.93
8.69
9.31
9.11
I2 I3 I4TTFc
A (BC)
AB
AC
BC
26.9 (9.8)
6.56 (7.16)
6.79 (2.82)
20.24 (2.68)
I4TTF I4TTF I3d
AB
BC
AC
a
Fig. 1a.
b
Fig. 1b. c Fig. 1c.
(3.50)
(10.08)
(2.82)
(7.26)
(0.39)
(9.15)
(8.55)
(9.24)
(8.26)
3.73 (6.34)
9.48 (10.08)
0.0 (0.0)
d
Fig. 1d. e Fig. 1e. f Fig. 1f.
This journal is © the Owner Societies 2014
g
I2 3-dithiole-2-thione-4,5-dicarboxylateiodinee
A (BDE)
C (BDE)
AB
CD
BC
AD
CE
AE
I4TTF I4TTFf
AD
AB
BC
CD
I43 - H43 (AD)
I23,I43,I44 - H23,H43,H44 (AB)
12.41 (5.44)
20.42 (8.82)
13.79 (5.38)
13.81 (5.38)
3.53 (5.27)
0.04 (0.09)
0.23 (0.24)
0.006 (0.03)
8.16
5.32
8.16
5.32
6.21
1.12
(9.43)
(7.97)
(9.44)
(7.97)
(7.64)
(0.82)
Cyanoethyl-bromine TTF Cyanoethyl-bromine TTFg
AB
4.90 (7.04)
Br1, Br2, Br31 - H1,H2,H31 (AB)
1.48 (1.33)
Br31 - H31 (AB)
3.89 (5.67)
BC
3.96 (5.65)
AC
2.24 (2.74)
Fig. 1g.
Phys. Chem. Chem. Phys., 2014, 16, 2038--2047 | 2043
View Article Online
Published on 15 November 2013. Downloaded by Bharathiar University on 31/07/2014 12:05:53.
PCCP
in them. The neutral dimer AB has dispersion energy
(7.16 kcal mol1) equivalent to that of interaction energy, while
in the case of the ionic system AC, the role of dispersion energy
is much smaller (2.82 kcal mol1). Evidently, the ionic system
is greatly stabilised by electrostatic energy (covered in the DFT
stabilisation energy) between the negative iodine (I9) in I3 and
the positive s hole on the iodine atoms (I12, I13) in I4TTF.
The electrostatic attraction between the iodine (I12) in I4TTF
and the iodine (I10) in I2 is apparently rather small (if any),
which is caused by the rather unfavourable geometry arrangement, +C2–I12 I10 = 156.341 and the I2 I10 distance
equals 3.5 Å.
Finally, we consider I2 and I3 as one subsystem and I4TTF
as another. As expected, the stabilisation energy of this complex
(A BC) is very high (26.9 kcal mol1), higher than the sum of
A B and A C stabilisation energies (13.35 kcal mol1). On
the other hand, the dispersion energy of the A BC dimer is
identical to the sum of the dispersion energy of both dimers
due to the definition of dispersion energy (pairwise additivity).
The difference between former and latter values (9.89 and
9.98 kcal mol1) is due to numerical errors during calculations. The increases of stabilisation energy in the A BC pair in
comparison with A B and A C could only be explained by
the polarisation of the I2 by the electric field of I3.
The trimer shown in Fig. 1d consists of the neutral dimer
AB, the ionic dimer BC and the neutral dimer AC. The first dimer
stabilised by two I I halogen bonds, has only a moderate
stabilisation energy of 3.73 kcal mol1. Clearly, the geometrical
arrangement is not favourable and the positive s hole of
one iodine does not interact with the negative belt of the
second iodine (I22 I19, I19 I22, I23 I18, I18 I23 =
3.7 Å, +C12–I23 I18, +C6–I19 I22 = 151.31, +C5–I18 I23,
+C11–I22 I19 = 126.91). The ionic dimer BC is identical with the
structure AB in Fig. 2b. Finally, the dimer AC has a negligible
stabilisation energy.
1,3-dithiole-2-thione-4-carboxylic acid I2. The system depicted
in Fig. 1e is electroneutral and consists of two dimer motifs,
which will be discussed separately. The dimers are considered
in such a way that neutral I2 molecules (B, D and E) form one
system and A or C different ones. The stabilisation energy of
the A(BDE) dimer, having two I S halogen bonds, is much
lower (12.41 kcal mol1) than that of the C (BDE) dimer
(20.42 kcal mol1), having four I S halogen bonds. The higher
stability of the C (BDE) dimer can be easily understood if we
determine the stabilisation energies of A or C with single iodine
(I2) molecules. The stabilisation energies of A B and C D are
practically identical (13.79, 13.81 kcal mol1), whereas those of
A D and C B differ dramatically. C forms two halogen bonds
with B, and the stabilisation energy of the complex is thus
rather high (3.53 kcal mol1). On the other hand, A practically
does not interact with D (DE = 0.04 kcal mol1). The C E and
A E pair stabilisation energies are low (0.23 and 0.0 kcal mol1)
and thus cannot affect the interaction between A or C and the
iodine molecules in the trimer.
I4TTF dimer. The I4TTF tetramer (Fig. 1f), which is electroneutral, contains four dimer structures, two of them being
2044 | Phys. Chem. Chem. Phys., 2014, 16, 2038--2047
Paper
identical (AD = BC, AB = CD). Among all dimers, AD (BC),
having one I I halogen bond and three I S halogen bonds,
is the most stable (8.16 kcal mol1), followed by AB (CD)
(5.32 kcal mol1). The latter structure also possesses one I I
halogen bond and three I S halogen bonds. The higher
stabilisation energy for the AD dimer (in comparison with the
AB one) could only be explained on the basis of the geometry
arrangement of the AD dimer, which allows the formation of
strong halogen bonds. The I I distance of the halogen bond in
the AD dimer is indeed much shorter (by 0.9 Å) than that of
the AB dimer, which is clearly the reason for the stronger
stabilisation of the AD dimer. The I S distances of the three
halogen bonds in the AD and AB dimers are more or less
comparable. By replacing I43 by hydrogen, we mutate the
C41–I36 I43 halogen bond to a C41–I36 H43 hydrogen
bond. Upon this mutation, the stabilisation energy decreases
by about 2 kcal mol1 from 8.16 to 6.21 kcal mol1. Evidently,
the C41–I36 I43 halogen bond contributes to the stability of
the A D motif and is responsible for the higher stability of
A D (in comparison with A B). Further evidence of the
strength of the halogen bonds in the AD dimer can be found
upon replacing three iodine atoms by hydrogen atoms (i.e. I44,
I43, I23 - H44, H43, H23). This makes all the halogen bonds
disappear and the total stabilisation drops dramatically from
8.16 to 1.12 kcal mol1.
Cyanoethyl-bromine–TTF dimer. Dimer cyanoethyl-bromine–
TTF (Fig. 1g) is electroneutral and consists of three dimers, AB,
BC and AC. Among all the dimers, the AB structure with one
halogen Br Br bond and three halogen Br S contacts is the most
stable (4.90 kcal mol1), followed by the BC (3.96 kcal mol1) and
AC structures (2.24 kcal mol1), which have one Br Br halogen
bond, two halogen bonds or three halogen bonds, respectively.
The higher stabilisation of the former dimer AB clearly arises
from the presence of the strong Br Br halogen bond, which is
geometrically close to the ideal structure (+C–Br Br = 167.311;
the Br Br distance equals 3.56 Å). When bromine was mutated
to hydrogen (i.e. the halogen bond was changed to the hydrogen
bond, C11–Br2 Br31 to C11–Br2 H31), the stabilisation
energy decreased from 4.90 to 3.89 kcal mol1. Once three
bromine atoms were mutated to hydrogen atoms (Br1, Br2,
Br31 - H1, H2, H31), i.e. the halogen bonds were replaced by
hydrogen bonds, the stabilisation energy dropped dramatically
from 4.90 to 1.48 kcal mol1, indicating the importance of the
halogen bonds in the original crystal. The distance of Br2 Br62
in the BC dimer is 0.6 Å larger than that in the AB dimer,
resulting in weaker stabilisation energy of the BC dimer.
X I3 and X I2 complexes. In this paragraph, we will
compare the stabilisation energies of the binding motifs of
I3 and I2, occurring in different crystals. From Fig. 1 and
Table 2, we can learn that there are five complexes containing
I3 : I3 TTF(SMe)2Br2 (the structures AB and AC in Fig. 1a),
I3 I4TTF (the structures AB and AC in Fig. 1b), the structure
AC in Fig. 1c; the structure BC in Fig. 1d is identical to the
structure AB in Fig. 1b with the following stabilisation energies
of 9.59, 8.50, 9.48, 6.79 and 6.79 kcal mol1, respectively. The
most stable dimers among them are AB in Fig. 1a with four
This journal is © the Owner Societies 2014
View Article Online
Published on 15 November 2013. Downloaded by Bharathiar University on 31/07/2014 12:05:53.
Paper
halogen bonds (I S), AB in Fig. 1b with two halogen bonds
(I I) and four halogen bonds (I S), and AC in Fig. 1a with
four halogen bonds.
Neither of the halogen bonds in the first two dimers are a
typical halogen bond characteristic of the attractive electrostatic interaction. Fig. 1a and b show that all the contacts
between the iodine in I3 (which are all negative) and the
sulphur atoms in TTF(SMe)2Br2 or the sulphur atoms and the
iodine atoms in I4TTF are electrostatically repulsive and
that the stabilisation of these complexes should originate in
dispersion energy. This was confirmed earlier by showing the
total stabilisation energies and their dispersion parts in the
complexes mentioned. The only attractive electrostatic interaction typical of a halogen bond could exist in the third most
stable dimer (AC in Fig. 1a), having four C–Br I halogen
bonds. All the Br I distances (3.6 Å) and C–Br I angles
(1641) are close to the ideal values for a halogen bond. The
stabilisation energies of all of these ionic dimers are moderate;
in fact, they are lower than we expected. We have tried to find in
literature any other theoretical study on the complexes of I3
with electron donors, however without success.
There are seven complexes containing neutral I2 (I2 I4TTF,
the structure AB, Fig. 2c; I2 1,3-dithiole-2-thione-4-carboxylic acid,
the structures AB, CD, AD, BC, AE, CE, Fig. 1e) with the following
stabilisation energies of 6.56, 13.79, 13.81, 0.04, 3.53, 0.20 and
0.0 kcal mol1, respectively. The very high stabilisation energies of
the second and third complexes are surprising. They are more than
4 kcal mol1 higher than those of the most stable complexes with
anionic I3. The elemental iodine (I2) is, however, known to form
very stable complexes, of which the I2 1,4-diazabicyclo[2.2.2]octane
(DABCO) complex with a stabilisation energy of 17.4 kcal mol1 and
an extremely short N I distance of 2.37 Å in the N I halogen bond
can be mentioned as an example.35
To verify the stabilisation energies given above, we repeated
the calculation for the AB dimer (Fig. 1e) at a different DFT level
(the B3LYP functional was used instead of B97). B3LYP-D/def2QZVP yields comparable stabilisation and dispersion energies
(12.35 and 5.09 kcal mol1) to the presently used B97-D/def2QZVP (13.79, 5.38 kcal mol1). Furthermore, we replaced the
sulphur S22 in the I22 S22 halogen bond with O22 (the
electronegativity of O is smaller). Both the stabilisation energy and
the dispersion energy decreased to 11.28 and 3.74 kcal mol1, but
the stabilisation energy was still very high. Evidently, the high
stabilisation energy of the AB dimer was real.
The high stabilisation energy of the AB dimer (Fig. 1e) is
connected with the surprisingly close contacts of 2.732 Å
between S28 and I18. To prove that it is not an artefact of
X-ray experiments, we recalculated the whole potential energy
curve for the dimer AB; the distance between S28 and I22
increased from 2.732 Å by 0.1 Å up to 4.332 Å. The resulting
potential energy curve, shown in Fig. 3, indicates that the
energy minimum is localised at a slighter greater distance
(2.832 Å) and is slightly deeper (14.5 kcal mol1). Evidently,
the close contact found in the X-ray crystal is correct.
We have shown in this paragraph that neutral X I2 complexes are surprisingly more stable than the ionic X I3 ones.
This journal is © the Owner Societies 2014
PCCP
Fig. 3 The halogen bond distance (S22 I18) vs. interaction energy (IE in
kcal mol1).
A possible explanation for this fact might be the existence of a
strong halogen bond of the I I type existing in the X I2
complexes on the one hand and the non-existence of the
halogen bond in the X I3 complexes on the other.
X Y dimers. Altogether, we investigated six neutral X Y
dimers (the AD and AB structures, Fig. 1f; the AB, AC and BC
structures, Fig. 1g; the AB structure, Fig. 1d) possessing I S
and Br S halogen and I I and Br Br halogen bonds. Their
stabilisation energies (8.16, 5.32, 4.90, 2.24, 3.96, 3.73 kcal mol1)
are lower than those of the X I2 complexes but are still high.
The highest stabilisation energy found for the first dimer arises
from the presence of three I S halogen and one I I halogen
bond, all having a very favourable geometry arrangement. For
the sake of comparison, it should be added that among the
various halogen-bonded complexes included in the X40 benchmark
dataset (the stabilisation energies determined at the CCSD(T)/
complete basis set limit),27 the iodobenzene trimethylamine
complex with a stabilisation energy of 5.8 kcal mol1 was found
to be the strongest.
I3 I2 and I2 I2 dimers. The dimer BC (I3 I2, Fig. 2c)
with one halogen bond (I10 I19) has a very high stabilisation
energy of 20.24 kcal mol1, which is, however, connected with
an only moderate dispersion energy (2.68 kcal mol1). The
electrostatic energy covered in the DFT stabilisation energy
evidently forms a dominant part of stabilisation. The very
favourable geometrical arrangement (+I11–I10 I19 = 169.01
and I10 I19 = 3.37 Å) allows the formation of a strong halogen
bond between the positive s-hole of one iodine (I2) and the
highly negative cusp of another iodine (I3).
Three I2 dimers (BD, DE, BE; Fig. 1e) with one or two
halogen bonds have an only moderate stabilisation energy of
1.75, 3.97, 3.29 kcal mol1, respectively, which is almost
completely formed by dispersion energy (1.94, 3.85 and
3.15 kcal mol1, respectively). The subsystems in the BD
dimer are separated by a rather large distance, and their
orientation does not allow the formation of an attractive halogen bond. On the other hand, the DE dimer has a favourable
geometrical arrangement (+I16–I15 I13 = 175.511 and
I15 I13 = 3.44 Å), allowing the formation of an attractive
I15 I13 halogen bond between the positive s-hole of one
Phys. Chem. Chem. Phys., 2014, 16, 2038--2047 | 2045
View Article Online
Published on 15 November 2013. Downloaded by Bharathiar University on 31/07/2014 12:05:53.
PCCP
iodine (I2) and the negative belt of the other iodine in (I2). We,
however, expected the dimer to have a higher stabilisation
energy, and now the question thus arises whether the crystal
structure of the DE dimer is close to the optimal gas-phase
geometry or whether the neighbouring molecules in the crystal
induce some important geometry (and thus also stabilisationenergy) change. To answer this question, we have optimised
(B97-D/def2-QZVP) the distance between I15 and I13 in the rectangular L-shape dimer (+I1–I2 I3 = 90.01 and +I16–I15 I13 =
180.01). Like in the previous case, the crystal geometry is close to
the optimal one (I15 I13 = 3.40 Å), and also the DE dimer
stabilisation energy is close to the optimal one (5.04 kcal mol1).
Evidently, the crystal environment does not induce any dramatical
changes in the optimal geometries and hence in the optimal
stabilisation energies either.
Finally, we compared the I I halogen bonds in neutral
I4TTF (AB; Fig. 1d) and I2 dimers (DE; Fig. 1e). The I4TTF dimer,
stabilised by two halogen bonds, has a lower stabilisation
energy, 3.73 kcal mol1, because of its unfavourable geometrical
arrangement (I22 I19, I19 I22, I23 I18, I18 I23 = 3.7 Å,
+C12–I23 I18, +C6–I19 I22 = 151.31, +C5–I18 I23, +C11–
I22 I19 = 126.91) compared to the DE dimer (3.85 kcal mol1)
with a favourable geometrical arrangement as discussed above.
Conclusion
The halogen bonds found in the crystal investigated are strong
interactions that play an important role in the binding of organic
crystal structures. As expected, the strongest halogen bonds are
those that involve iodine as a halogen-bond donor and acceptor.
This is caused by the large s-hole on iodine (larger than that on
bromine and chlorine) as well as by its large polarisability. Among
ionic X I3 and neutral X I2 and X Y dimers, the neutral
X I2 complexes are, surprisingly enough, the most stable. The
very high stabilisation energy of the I2 1,3-dithiole-2-thione-4carboxylic acid complex (13.8 kcal mol1), verified by performing
a calculation with a different functional as well as by mutating the
electron donor in the halogen bond (S to O), is attributable to the
existence of a very strong I I halogen bond.
The stabilisation energy of the ionic I3 I2 complex is
extremely high (20.24 kcal mol1), which is explained by the
very favourable geometrical arrangement allowing the formation of
a strong halogen bond. On the other hand, the stabilisation of the
neutral I2 I2 complex is only moderate (3.9 kcal mol1), but this
value is close to that in the optimal gas-phase L-shaped complex.
In all the cases, the dispersion energy is important and
comparable to the electrostatic energy. Only in the case of a
strong halogen bond (I3 I2), the electrostatic energy becomes
dominant.
Acknowledgements
This work was part of the Research Project RVO: 61388963 of
the Institute of Organic Chemistry and Biochemistry, Academy
of Sciences of the Czech Republic. This work was also
2046 | Phys. Chem. Chem. Phys., 2014, 16, 2038--2047
Paper
supported by the Czech Science Foundation [P208/12/G016]
and the operational program Research and Development for
Innovations of European Social Fund (CZ 1.05/2.1.00/03/0058).
References
1 M. Kolář and P. Hobza, J. Chem. Theory Comput., 2012, 8,
1325–1333.
2 G. Roy, K. P. Bhabak and G. Mugesh, Cryst. Growth Des.,
2011, 11, 2279–2286.
3 C. B. Aakeröy, M. Fasulo, N. Schultheiss, J. Desper and
C. Moore, J. Am. Chem. Soc., 2007, 129, 13772–13773.
4 K. E. Riley and P. Hobza, Cryst. Growth Des., 2011, 11,
4272–4278.
5 K. E. Riley and P. Hobza, J. Phys. Chem. B, 2008, 112,
3157–3163.
6 E. Munusamy, R. Sedlak and P. Hobza, ChemPhysChem,
2011, 12, 3253–3261.
7 K. E. Riley, J. S. Murray, P. Politzer, M. C. Concha and
P. Hobza, J. Chem. Theory Comput., 2008, 5, 155–163.
8 K. E. Riley and P. Hobza, J. Chem. Theory Comput., 2008, 4,
232–242.
9 W. Wang and P. Hobza, J. Phys. Chem. A, 2008, 112,
4114–4119.
10 V. Fritsch and E. Westhof, Structure, 2009, 17, 784–786.
11 S. D. Gilbert, F. E. Reyes, A. L. Edwards and R. T. Batey,
Structure, 2009, 17, 857–868.
12 D. B. Fox, R. Liantonio, P. Metrangolo, T. Pilati and
G. Resnati, J. Fluorine Chem., 2004, 125, 271–281.
13 P.-P. Zhou, W.-Y. Qiu, S. Liu and N.-Z. Jin, Phys. Chem.
Chem. Phys., 2011, 13, 7408–7418.
14 J. L. Alonso, S. Antolı́nez, S. Blanco, A. Lesarri, J. C. López
and W. Caminati, J. Am. Chem. Soc., 2004, 126, 3244–3249.
15 K. E. Riley, J. Řezáč and P. Hobza, J. Mol. Model, 2013, 19,
2879–2883.
16 P. Politzer, J. S. Murray and M. C. Concha, J. Mol. Model,
2008, 14, 659–665.
17 K. E. Riley, J. S. Murray, J. Fanfrlı́k, J. Řezáč, R. J. Solá,
M. C. Concha, F. M. Ramos and P. Politzer, J. Mol. Model,
2013, 19, 4651–4659.
18 T. Clark, M. Hennemann, J. S. Murray and P. Politzer, J. Mol.
Model, 2007, 13, 291–296.
19 P. Politzer, J. S. Murray and T. Clark, Phys. Chem. Chem.
Phys., 2013, 15, 11178–11189.
20 P. Politzer and J. S. Murray, ChemPhysChem, 2013, 14,
278–294.
21 C. Wang, J. Y. Becker, J. Bernstein, A. Ellern and
V. Khodorkovsky, J. Mater. Chem., 1995, 5, 1559–1562.
22 A. S. Batsanov, A. J. Moore, N. Robertson, A. Green,
M. R. Bryce, J. A. Howard and A. E. Underhill, J. Mater.
Chem., 1997, 7, 387–389.
23 A. S. Batsanov, M. R. Bryce, A. Chesney, J. A. Howard,
D. E. John, A. J. Moore, C. L. Wood, H. Gershtenman,
J. Y. Becker and V. Y. Khodorkovsky, J. Mater. Chem., 2001,
11, 2181–2191.
This journal is © the Owner Societies 2014
View Article Online
Published on 15 November 2013. Downloaded by Bharathiar University on 31/07/2014 12:05:53.
Paper
24 F. Zordan and L. Brammer, Cryst. Growth Des., 2006, 6,
1374–1379.
25 Y.-X. Lu, J.-W. Zou, Q.-S. Yu, Y.-J. Jiang and W.-N. Zhao,
Chem. Phys. Lett., 2007, 449, 6–10.
26 P. J. Skabara, R. Berridge, N. Bricklebank, H. Lath, S. J. Coles
and P. N. Horton, Polyhedron, 2006, 25, 989–995.
27 K. E. Riley and K. M. Merz, J. Phys. Chem. A, 2007, 111,
1688–1694.
28 P. Metrangolo, H. Neukirch, T. Pilati and G. Resnati, Acc.
Chem. Res., 2005, 38, 386–395.
29 S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys.,
2010, 132, 154104.
This journal is © the Owner Societies 2014
PCCP
30 S. Grimme, S. Ehrlich and L. Goerigk, J. Comput. Chem.,
2011, 32, 1456–1465.
31 L. Goerigk, H. Kruse and S. Grimme, ChemPhysChem, 2011,
12, 3421–3433.
32 J. Rezáč, K. E. Riley and P. Hobza, J. Chem. Theory Comput.,
2012, 8, 4285–4292.
33 G. J. Beran and K. Nanda, J. Phys. Chem. Lett., 2010, 1,
3480–3487.
34 J. Chambers and F. Migliorini, Bull. Am. Astron. Soc., 1997,
29, 1024.
35 A. Peuronen, A. Valkonen, M. Kortelainen, K. Rissanen and
M. Lahtinen, Crys. Growth Des., 2012, 12, 4157–4169.
Phys. Chem. Chem. Phys., 2014, 16, 2038--2047 | 2047