Weak halogen bonding in solid haloanilinium halides probed directly via chlorine-35, bromine-81, and iodine-127 NMR spectroscopy

Article abstract of DOI:10.1021/cg201683p

A series of monohaloanilinium halides exhibiting weak halogen bonding (XB) has been prepared and characterized by ³⁵Cl, ⁸¹Br, and ¹²⁷I solid-state nuclear magnetic resonance (SSNMR) spectroscopy in magnetic fields of up to 21.1 T. The quadrupolar and chemical shift (CS) tensor parameters for halide ions (Cl^-, Br^-, I^-) which act as electron density donors in the halogen bonds of these compounds are measured to provide insight into the possible relationship between halogen bonding and NMR observables. The NMR data for certain series of related compounds are strongly indicative of when such compounds pack in the same space group, thus providing practical structural information. Careful interpretation of the NMR data in the context of novel and previously available X-ray crystallographic data, and new gauge-including projector-augmented-wave density functional theory (GIPAW DFT) calculations has revealed several notable trends. When a series of related compounds pack in the same space group, it has been possible to interpret trends in the NMR data in terms of the strength of the halogen bond. For example, in isostructural series, the halide quadrupolar coupling constant was found to increase as the halogen bond weakens. In the case of a series of haloanilinium bromides, the ⁸¹Br isotropic chemical shift and CS tensor span both decrease as the bromide-halogen XB is weakened. These trends were reproduced using both GIPAW DFT and cluster-model calculations of the bromide ion magnetic shielding tensor. Such trends are particularly exciting given the well-known role that NMR has played historically in the characterization of hydrogen bonding.

Full text of DOI:10.1021/cg201683p

Article
pubs.acs.org/crystal
Weak Halogen Bonding in Solid Haloanilinium Halides Probed
Directly via Chlorine-35, Bromine-81, and Iodine-127 NMR
Spectroscopy
Published as part of the Crystal Growth & Design virtual special issue on Halogen Bonding in Crystal
Engineering: Fundamentals and Applications
Robert J. Attrell, Cory M. Widdifield, Ilia Korobkov, and David L. Bryce*
Department of Chemistry and Centre for Catalysis Research and Innovation, University of Ottawa, 10 Marie Curie Private, Ottawa,
Ontario, Canada
S
* Supporting Information
ABSTRACT: A series of monohaloanilinium halides exhibiting
weak halogen bonding (XB) has been prepared and characterized
by 35Cl, 81Br, and 127I solid-state nuclear magnetic resonance
(SSNMR) spectroscopy in magnetic fields of up to 21.1 T. The
quadrupolar and chemical shift (CS) tensor parameters for halide
ions (Cl−, Br−, I−) which act as electron density donors in the
halogen bonds of these compounds are measured to provide
insight into the possible relationship between halogen bonding and
NMR observables. The NMR data for certain series of related
compounds are strongly indicative of when such compounds pack in the same space group, thus providing practical structural
information. Careful interpretation of the NMR data in the context of novel and previously available X-ray crystallographic data,
and new gauge-including projector-augmented-wave density functional theory (GIPAW DFT) calculations has revealed several
notable trends. When a series of related compounds pack in the same space group, it has been possible to interpret trends in the
NMR data in terms of the strength of the halogen bond. For example, in isostructural series, the halide quadrupolar coupling
constant was found to increase as the halogen bond weakens. In the case of a series of haloanilinium bromides, the 81Br isotropic
chemical shift and CS tensor span both decrease as the bromide−halogen XB is weakened. These trends were reproduced using
both GIPAW DFT and cluster-model calculations of the bromide ion magnetic shielding tensor. Such trends are particularly
exciting given the well-known role that NMR has played historically in the characterization of hydrogen bonding.
tensor components in a series of crystalline hydrates are clearly
correlated to the hydrogen bond environment13 and have
demonstrated the influence of hydrogen bonding on 17O NMR
interaction tensors in uracil.14 This suggests that NMR
experiments may provide equally valuable insights into halogen
INTRODUCTION
■
Weak noncovalent interactions including hydrogen bonding,
cation-π interactions, and halogen bonding (XB) play
important roles in areas as diverse as protein and nucleic acid
structure, supramolecular chemistry, and crystal engineering.1−4
Hydrogen bonding in particular is a well-established paradigm
in chemistry.1,5 Recently, the concept of halogen bonding has
garnered increased attention.2,6−10 XB refers to a D···XY
interaction where X is a halogen, D is a donor of electron
density (e.g., a halide), and Y is another atom (e.g., carbon).1
For example, Metrangolo et al. have demonstrated that iodide−
iodine interactions are key to the selective capture of
diiodoperfluoroalkanes by solid nonporous organic iodides.7
Recent work by Taylor and co-workers reported on the
thermodynamics of XB in solution and discussed the
differences between this interaction and the hydrogen bonding
interaction.8
1
bonding in solids. Bouchmella et al. have presented H, 13C,
and 15N solid-state NMR studies of some imidazole- and
morpholine-based compounds exhibiting halogen bonding
capabilities.15 However, solid-state NMR spectroscopy of
35/37Cl, 79/81Br, and 127I is challenging because all of these
nuclides are quadrupolar (I > 1/2) and typically give rise to very
broad powder patterns in the solid state.16 While still far from
routine, there have been significant advances in the develop-
ment and application of solid-state halogen NMR spectroscopy
recently. For example, we have demonstrated the sensitivity of
35/37Cl NMR parameters to the hydrogen bonding environment
in amino acid hydrochlorides,17 and of 79/81Br NMR parameters
to changes in structure and hydration state.18 The use of
NMR spectroscopy is a well-established probe of hydrogen
bonding in solution and in solids, and is in fact one of the
premier methods for studying these interactions.11,12 For
Received: December 21, 2011
Published: January 6, 2012
1
example, Wu et al. showed that the H chemical shift (CS)
© 2012 American Chemical Society
1641
dx.doi.org/10.1021/cg201683p | Cryst. Growth Des. 2012, 12, 1641−1653

Crystal Growth & Design
Article
chlorine NMR to distinguish between pharmaceutical poly-
morphs was shown by Schurko and co-workers,19 and the local
structure around halide ions in ionic liquids was probed by
Ripmeester and co-workers.20
Here, we report on the characterization of a series of
haloanilinium halides which exhibit weak halogen bonding
(Scheme 1) via 35/37Cl, 79/81Br, and 127I solid-state NMR
bonding interaction can have on the NMR parameters (e.g., the
quadrupolar coupling constant and the chemical shift tensors)
of Cl−, Br−, and I− when they are electron donors in a halogen
bond. The solid-state NMR studies are supported with single-
crystal X-ray crystallographic studies as well as gauge-including
projector-augmented-wave density functional theory (GIPAW
DFT) calculations24−27 of the relevant NMR interaction
tensors.
Scheme 1. General Molecular Structure of
Monohaloanilinium Halides
a
EXPERIMENTAL SECTION
■
(i). Sample Preparation. All organic compounds were purchased
from Sigma-Aldrich and used without further purification. The starting
materials and products are air stable, and preparation was performed
using a procedure adapted from the literature.28 Synthesis consisted of
mixing a given haloaniline with an excess of a given hydrohalic acid.
This was followed by a period where the salt was allowed to precipitate
from an ethanol solution, and the samples were filtered through a
Buchner funnel with light suction and then completely dried in air at
room temperature. Crystals were grown from ethanol by slow
evaporation over several days, and their identities were confirmed by
single-crystal X-ray diffraction. Compound numbering is given in
Table 1. Crystals were powdered prior to solid-state NMR studies.
Before NMR experiments, all powder samples were packed tightly into
1.3, 2.5, or 4 mm o.d. zirconium oxide magic-angle spinning (MAS)
rotors.
(ii). X-ray Crystallography. Crystals of 2, 3, 12, and 13 were
mounted on thin glass fibers using paraffin oil and cooled to 200.15 K
prior to data collection. Data were collected on a Bruker AXS SMART
single crystal diffractometer equipped with a sealed Mo tube source
(wavelength = 0.71073 Å) and an APEX II CCD detector. Raw data
collection and processing were performed with the APEX II software
package from BRUKER AXS.29 Diffraction data were collected with a
sequence of 0.5° ω scans at 0, 120, and 240° in φ. Initial unit cell
parameters were determined from 60 data frames collected for
a
X is an electron density acceptor and D is a donor of electron density.
X = Cl, Br, I. D = Cl−, Br−, I−.
spectroscopy. The solid-state NMR spectroscopy of the
covalently bound halogens (i.e., the electrophiles in halogen
bonds) is generally impractical, and instead we focus our
attention on the electron density donors: Cl−, Br−, and I−.
Semin and co-workers have reported 127I nuclear quadrupole
resonance (NQR) studies of covalently bound iodine atoms
which are engaged in halogen bonding,21,22 and this prompted
us to explore the potential utility of examining the electron
density donors involved in the halogen bonds. It has been
shown recently that 77Se solid-state NMR is indeed sensitive to
halogen bonding with iodines in a series of selenocyanate/
diiodoperfluorobenzene complexes.23 One of the goals of the
present study is therefore to explore what influence a halogen
Table 1. Compound Numbering and Some Crystal Structure and Local Halogen Bonding Geometrical Information for
Monohaloanilinium Halides for Which Either Single-Crystal X-ray Diffraction or Halogen Solid-State NMR Data Are Available
a
b
c
compound number
compound name
space group
crystal structure reference
C−X (Å)
X···D (Å)
C−X···D (°)
R
1
2-chloroanilinium chloride
2-bromoanilinium chloride
Pca21
Pca21
P21/n
P21/n
28
1.740
1.899
2.101
1.745
3.412
3.314
3.330
n/a
170.6
170.3
175.9
n/a
0.96
0.91
0.88
n/a
2
this work
this work
28
d
3
2-iodoanilinium chloride
4
3-chloroanilinium chloride
3-bromoanilinium chloride
3-iodoanilinium chloride
4-chloroanilinium chloride
4-bromoanilinium chloride
4-iodoanilinium chloride
2-chloroanilinium bromide
2-bromoanilinium bromide
5
6
7
P21/c
P21/c
P21ab
Pca21
Pca21
P21/n
Pca21
28,34
35,36
36
1.742
1.892
2.102
1.742
1.893
2.101
1.742
1.895
3.635
3.587
3.405
3.589
3.443
3.465
3.838
n/a
166.6
165.9
169.8
170.0
170.2
175.7
164.7
n/a
1.02
0.98
0.90
0.97
0.90
0.88
1.03
n/a
8
9
10
11
12
13
14
16
17
18
19
20
21
24
37
28
d
2-iodoanilinium bromide
this work
this work
28
3-chloroanilinium bromide
3-bromoanilinium bromide
4-chloroanilinium bromide
4-bromoanilinium bromide
4-iodoanilinium bromide
2-chloroanilinium iodide
2-bromoanilinum iodide
2-iodoanilinium iodide
P1
̅
P1
̅
36
2.102
3.704
158.4
0.94
P21/c
28
28
2.101
2.104
3.692
3.782
170.2
174.2
0.88
0.90
3-iodoanilinium iodide
P21/m
a
Compounds 15, 22, and 23 (i.e., 3-iodoanilinium bromide, 3-chloroanilinium iodide, and 3-bromoanilinium iodide) are not listed as we do not
b
report NMR data for them and to our knowledge there is no information on their crystal structures. Distance of shortest contact. Here, “X” is the
covalently bound halogen and “D” is the anionic halide (donor of electron density); X and D may be different elements. Relative XB distance, R =
d/(rX + rD), where d is the X···D distance and rX and rD are the van der Waals’ radii of the electron density acceptor and donor, respectively. Single-
c
d
crystal X-ray diffraction revealed these compounds to be monohydrates.
1642
dx.doi.org/10.1021/cg201683p | Cryst. Growth Des. 2012, 12, 1641−1653

Crystal Growth & Design
Article
Table 2. Crystallographic Data and Selected Data Collection Parameters for 2, 3, 12, and 13
a
a
2-bromoanilinium chloride
2-iodoanilinium chloride
2-iodoanilinium bromide
3-chloroanilinium bromide
compound
empirical formula
(2)
(3)
(12)
(13)
C6H7BrClN
208.49
C6H9ClINO
273.49
C6H9BrINO
317.95
C6H7BrClN
formula weight
208.49
crystal size, mm
0.32 × 0.25 × 0.21
orthorhombic
Pca21 (No. 29)
4
0.21 × 0.18 × 0.18
monoclinic
P21/n (No. 14)
4
0.23 × 0.19 × 0.16
monoclinic
P21/n (No. 14)
4
0.22 × 0.21 × 0.18
orthorhombic
crystal system
space group
Pca21 (No. 29)
Z
4
a, Å
15.606(2)
5.4026(7)
8.7865(11)
90
7.64850(10)
6.64340(10)
17.5487(3)
90
7.8293(2)
6.7808(2)
17.7752(4)
90
16.1551(5)
b, Å
5.5958(2)
c, Å
8.4538(2)
α, °
90
β, °
γ, °
volume, Å3
90
91.3540(10)
90
92.6510(10)
90
90
90
90
740.80(17)
1.869
891.44(2)
2.038
942.66(4)
2.240
764.23(4)
calculated density, Mg/m3
1.812
−1
absorption coefficient, mm
F(000)
5.816
3.830
7.577
5.638
408
520
592
408
Θ range for data collection, °
limiting indices
2.61−28.31
2.32−28.24
2.60−30.50
h = 11, k = 9, l = 25
3.49−28.28
h = 21, k = 6, l = 11
h = 20, k = 7, l = 11 h = 9, k = 8, l = 23
reflections collected/unique
R(int)
5679/1695
0.0246
9170/2167
19713/2829
8717/1871
0.0174
0.0230
0.0378
completeness to Θ = 28.32, % 98.5
97.9
98.0
99.1
max and min transmission
data/restraints/parameters
goodness-of-fit on F2
0.3747 and 0.2576
0.5456 and 0.5002
2167/0/91
0.3769 and 0.2746
2829/0/91
0.4302 and 0.3702
1871/1/82
1695/1/82
1.084
1.084
1.050
1.017
final R indices [I > 2σ(I)]
R indices (all data)
R1 = 0.0239, wR2 = 0.0614
R1 = 0.0259, wR2 = 0.0623
0.019(14)
R1 = 0.0153, wR2 = 0.0402
R1 = 0.0162, wR2 = 0.0407
R1 = 0.0170, wR2 = 0.0410
R1 = 0.0182, wR2 = 0.0416
R1 = 0.0253, wR2 = 0.0599
R1 = 0.0267, wR2 = 0.0606
0.038(9)
absolute structure parameter
−3
largest diff. peak/hole, e·Å
0.282/−0.719
0.328/−0.839
0.622/−0.966
0.778/−0.758
a
These compounds are monohydrates.
different sections of the Ewald sphere. Semiempirical absorption
corrections based on equivalent reflections were applied.30 Systematic
absences in the diffraction data set and unit cell parameters were
consistent with the orthorhombic Pca21 space group for 2 and 13 and
the monoclinic P21/n space group for 3 and 12. Solutions for all
compounds yielded chemically reasonable and computationally stable
refinement results. Because of the non-centrosymmetric group and
presence of heavy atoms for 2 and 13, floating origin restraints were
automatically generated during the refinement. The structures were
solved by direct methods, completed with difference Fourier synthesis,
and refined with full-matrix least-squares procedures based on F2. In
the structures of both compounds, the halide anions and the anilinium
cations are situated in general positions. All non-hydrogen atoms were
refined anisotropically with satisfactory thermal parameter values.
Positions of all the hydrogen atoms were obtained from the Fourier
map analysis; however, all hydrogen atoms were treated as idealized
contributions. All scattering factors are contained in several versions of
the SHELXTL program library, with the latest version used being
6.12.31 Crystallographic data and selected data collection parameters
are reported in Table 2.
using the secondary standard, and the selective central transition pulse
scaling factors were calculated using the expression 1/(I + 1/2), where
I is the spin quantum number of the nucleus. SSNMR signals for
35/37Cl, 79/81Br, and 127I were acquired using the Solomon (π/2−π/2)
and Hahn (π/2−π) echo pulse sequences. For the chlorine
compounds at 9.4 and 21.1 T, spectral windows of between 250 and
1000 kHz were used, acquisition times ranged from 1 to 4 ms, and
proton decoupling was used when possible. Recycle delays of 2−5 s
were used, π/2 pulses were between 2.5 and 4.5 μs, and echo delays
were between 36 and 80 μs. Approximately 25k to 50k transients were
collected for these spectra.
For bromine and iodine, spectral windows of 2 MHz were used for
all compounds, with acquisition times ranging from 180 to 301 μs.
Proton decoupling was used only at 21.1 T. Recycle delays were 0.25
to 0.5 s, echo delays ranged from 20 to 40 μs, and π/2 pulse lengths
ranged from 1.15 to 1.6 μs. Approximately 3k to 10k transients were
collected for each piece for these spectra. Variable Offset Cumulative
Spectrum (VOCS)32 data acquisition was used when required for
chlorine, bromine, and iodine. Offsets were set to 250 kHz for the π/
2−π/2 and π/2−π piecewise experiments. The resulting spectra were
coadded in the frequency domain to produce the final spectrum. Fast
MAS (up to 62.5 kHz) was employed for some 81Br NMR spectra.
Note that spectra acquired for 37Cl and 79Br were used to confirm all
spectral parameters, but the data are not shown here. SSNMR
lineshapes were fit using WSolids1 simulation software that allows for
analytical modeling of chemical shift anisotropy (CSA) and
quadrupolar effects simultaneously.33 All simulated spectra include
only the effects of the central transition.
(iii). Solid-State 35/37Cl,79/81Br, and 127I NMR. Data were acquired
at the University of Ottawa using 9.4 and 11.7 T wide-bore magnets
and at the National Ultrahigh-field NMR Facility for Solids in Ottawa
at 21.1 T with a standard-bore magnet. Experiments at 9.4 and 11.7 T
used a 4 mm Bruker HXY or HX MAS probe, while experiments at
21.1 T used a 4 mm HX MAS probe for bromine and iodine, a 4 mm
static X probe for chlorine, as well as a 2.5 mm HX MAS probe for
chlorine and 1.3 mm HX MAS probe for bromine. Chlorine-35/37,
bromine-79/81, and iodine-127 NMR spectra were referenced to 0.1
mol/dm3 NaCl in D2O, 0.01 mol/dm3 NaBr in D2O, and 0.01 mol/
dm3 KI in D2O, respectively. The secondary standards used were solid
NaCl (−41.11 ppm) or KCl (8.54 ppm), solid KBr (54.51 ppm), and
solid KI (192.62 ppm). Pulse widths (π and π/2) were determined
(iv). Quantum Chemical Calculations. GIPAW DFT (PBE
exchange-correlation functional under the generalized gradient
approximation) calculations of NMR parameters were performed
using the CASTEP24−27 program for all compounds for which crystal
structures are available or which were solved in the course of this
1643
dx.doi.org/10.1021/cg201683p | Cryst. Growth Des. 2012, 12, 1641−1653

Crystal Growth & Design
Article
Figure 1. Important distances and angles around the halide ion in (a) 3-chloroanilinium bromide (13) and in (b) 2-bromoanilinium chloride (2).
Labeled are the C−X···D angle, the halogen−halide distance, and the halide−nitrogen distances.
Figure 2. (a−i) Chlorine-35 SSNMR spectra of 1, 2, and 3. See the Experimental section for further details on how the data were acquired and
processed.
study. Experimental crystal structures for compounds 1,28 2, 4,28 7,34
RESULTS AND DISCUSSION
■
8,35 9,36 10,37 11,28 13, 14,28 and 1836 were used without geometry
(i). Crystal Structures. The crystal structures of 2-
bromoanilinium chloride (2) and 3-chloroanilinium bromide
(13) (Figure 1) were solved during the course of this study,
optimization, and some additional computations were also performed
with optimization of the hydrogen atom positions. Ultrafine settings
were used for the cutoff energies and k-point grids, unless calculations
with those settings would not converge. In those cases the “fine”
settings were used (see Supporting Information for energy cutoffs in
eV and k-point grids used). Pseudopotential files were obtained
directly from Accelrys Inc. Selected calculations were also performed
on cluster models using Gaussian 0938 (B3LYP/6-311++G**).
Conversion of shielding constants to chemical shifts was carried out
using the experimental absolute shielding scale for chlorine,39 while for
bromine a GIPAW DFT-computed value for KBr was used.
and some relevant crystallographic and structural parameters
are summarized in Tables 1 and 2. Also shown in Table 1 are
the space groups and some geometrical information pertaining
to the halogen bonding environment in the larger series of
monohaloanilinium halides. The crystal structures of some of
these compounds, as well as the topic of halogen bonding in
monohaloanilinium halides, have been discussed previously,
most notably by Rissanen and co-workers,36 and by Gray and
Jones.28 We therefore focus our brief discussion of crystallo-
graphic aspects on how the data for the two new crystal
1644
dx.doi.org/10.1021/cg201683p | Cryst. Growth Des. 2012, 12, 1641−1653

Crystal Growth & Design
Article
Figure 3. (a−i) Chlorine-35 SSNMR spectra of 4, 5, and 6. See the Experimental section for further details on how the data were acquired and
processed.
structures fit into the overall trends. This discussion also
reviews relevant structural features which will be useful in the
interpretation of the halogen solid-state NMR data (vide infra).
Despite the chemical similarity within the series of
compounds shown in Table 1, they crystallize from ethanol
in a variety of space groups. The majority of these compounds
feature near-linear C−X···D angles (X = Cl, Br, I; D = Cl−, Br−,
I−). It is useful to discuss these contacts in terms of a
normalized distance parameter, R,40
halogen bonding interactions are seen for 3, 12, and 21 (R =
0.88), which is not surprising as all involve iodine as the
electron density acceptor. The C−X···D angles range from
158.4° in 18 to 175.9° in 3. The most linear contact is therefore
observed when iodine is the electron density acceptor.
However, overall it is clear that the halogen bonding interaction
in monohaloanilinium halides is rather weak compared to the
more pronounced cases observed when the electron density
acceptor is p-diiodotetrafluorobenzene or other related
perfluorinated compounds, as quantified by values of R. As
noted previously,28 two compounds in the series for which
there are X-ray crystallographic data (4 and 14) do not exhibit
any halogen bonding interaction; this is of interest in the
context of the present study to assess whether this difference is
manifested in any of the NMR parameters (vide infra).
It is important to also consider that the halide ions in these
compounds are typically involved in various hydrogen bonding
interactions with nearby NH groups, in addition to
participating in a halogen bond. As an example, we show in
Figure 1 the local halide ion environment for 2 and 13. The
interactions seen are typical of many of the compounds in
d
R =
rX + rD
where d is the observed X···D distance, and rX and rD are the
van der Waals’ radii41,42 of the halogen (X) and halide (D),
respectively. The structures of 2 and 13 exhibit R values near
the two extremes of all the values shown in Table 1: 0.91 for 2
and 1.03 for 13. The latter compound, while exhibiting an
approximately linear C−Cl···Br− angle of 164.7°, therefore does
not exhibit notable halogen bonding as shown by the large
value of R. It is well-known that iodine typically forms the
strongest halogen bonds within a given series.2 The strongest
1645
dx.doi.org/10.1021/cg201683p | Cryst. Growth Des. 2012, 12, 1641−1653

Crystal Growth & Design
Article
Figure 4. (a−i) Chlorine-35 SSNMR spectra of 7, 8, and 9. The asterisks in (b) and (i) denote impurities. See the Experimental section for further
details on how the data were acquired and processed.
a
Table 3. Experimental 35Cl Quadrupolar and Chemical Shift Tensor Parameters
compound
|CQ(35Cl)|/MHz
ηQ
δiso/ppm
Ω/ppm
κ
α (°)
β (°)
γ (°)
1
2
3
4
5
6
7
8
9
6.04(3)
6.01(1)
2.12(2)
5.30(5)
5.37(3)
5.39(5)
6.02(5)
5.60(5)
4.33(7)
0.34(1)
0.330(5)
0.76(1)
0.025(5)
0.025(5)
0.02(2)
0.56(1)
0.60(1)
0.38(1)
70(1)
79(1)
75(1)
88(1)
85(1)
77(1)
69(1)
67(1)
71(1)
95(5)
95(5)
−0.20(2)
−0.50(2)
−0.10(2)
−0.80(1)
−1.0(0.1)
−0.50(2)
−1.0(0.1)
−0.40(1)
0.0(0.1)
90(20)
90(20)
20(10)
0(5)
90(5)
90(2)
85(5)
81(2)
83(2)
85(3)
90(5)
78(2)
90(5)
0(5)
0(2)
0(5)
0(2)
0(2)
5(3)
10(5)
12(2)
0(5)
40(5)
100(5)
95(10)
90(5)
0(15)
0(0)
115(5)
90(5)
20(20)
75(10)
55(10)
120(5)
a
Parameters are defined in terms of the principal components of the quadrupolar and CS tensors as follows: CQ = eQV33/h; ηQ = (V11 − V22)/V33;
δiso = (δ11 + δ22 + δ33)/3; Ω = δ11 − δ33; κ = 3(δ22 − δiso)/Ω, where δ11 ≥ δ22 ≥ δ33 and |V11| ≤ |V22| ≤ |V33|. The angles α, β, γ express the relative
orientation of the CS and quadrupolar tensor principal axis systems. Error bounds are in brackets. Although CQ may take any real value, |CQ| is
measured using SSNMR experiments.
Table 1 in that a single nearly linear C−X···D contact angle is
present, along with three hydrogen bonds to the halide. All
relevant halide−nitrogen distances are tabulated in the
Supporting Information (Table 2S).
The crystal structures for compounds 3 and 12 were solved
at the end of this study, as a result of the conclusions reached
from our solid-state NMR studies (vide infra), and those
structures will be discussed in the appropriate context below.
(ii). Chlorine-35 Solid-State NMR Spectroscopy.
Presented in Figures 2, 3, and 4 are the 35Cl NMR spectra of
powdered samples of compounds 1−9, acquired under
stationary and MAS conditions. In each case, only the central
transition (mI = 1/2 ↔ −1/2) of the 35Cl nucleus in the chloride
anion is observed. The 35Cl quadrupolar and CS tensor
parameters determined through simulation of the spectra are
presented in Table 3. It is important to note that due to the
1646
dx.doi.org/10.1021/cg201683p | Cryst. Growth Des. 2012, 12, 1641−1653

Crystal Growth & Design
Article
number of parameters that can influence the spectral
appearance, it is advantageous to (i) acquire spectra of both
stationary and MAS (where possible) samples; (ii) acquire
spectra at more than one magnetic field strength; (iii) acquire
spectra of another isotope when possible (i.e., 37Cl in this case;
data not shown). Under fast MAS conditions, the effects of
CSA are removed and thus the values of the isotropic chemical
shift (δiso), quadrupolar coupling constant (CQ), and
quadrupolar asymmetry parameter (ηQ) may be accurately
measured (see the top row of Figures 2, 3, and 4 for 35Cl MAS
NMR spectra of 1−9). Definitions for these NMR parameters
may be found in the footnotes of Table 3. Subsequent to the
proper modeling of the MAS NMR data, analysis of the NMR
spectrum of a stationary sample then affords the remaining
parameters. Since the line shape broadening effects of the
quadrupolar interaction and the chemical shift scale differently
with the applied magnetic field, acquiring data in more than
one magnetic field strength increases the precision of the
determined parameters.
3, however, exhibits completely different 35Cl NMR spectra and
35Cl quadrupolar parameters, i.e., CQ = 2.12 MHz (vs 6.04 and
6.01 MHz for 1 and 2) and ηQ = 0.76 (vs 0.34 and 0.330 for 1
and 2). The span (Ω) of the chlorine CS tensor is also
substantially different for 3 (40 ppm) as compared to 1 and 2
(95 ppm). This strongly implies that 3 packs in a space group
other than Pca21, and indeed that the halogen and hydrogen
bonding environment about the chloride ion is different from
that found in the isostructural 1 and 2. Taken together, the data
for this series (1, 2, and 3) show that the NMR parameters
reflect crystal symmetry and packing, and also hint that the
slightly different XB environment may be reflected in the value
of δiso for 1 and 2. The value of δiso for 3 cannot be
meaningfully compared to those for 1 and 2 because the
quadrupolar parameters for the three compounds indicate that
they pack in different space groups and therefore δiso will be
affected by more than just slight differences in the XB
environment in 3.
Chlorine-35 NMR spectra for the 3-halosubstituted chloride
series (4, 5, and 6) are presented in Figure 3. The spectra, as
well as the NMR parameters determined through spectral
simulations (Table 3), strongly imply that these three
compounds pack in the same space group and exhibit similar
halogen bonding geometries. Even a cursory inspection of the
NMR spectra shows that they have very similar forms for the
three compounds. The values of CQ(35Cl) vary over a small
range, from 5.30 MHz for 4 to 5.39 MHz for 6. The
quadrupolar asymmetry parameter for these three compounds
is identical within experimental error (Table 3). The CS tensor
parameters are also very similar for the three compounds, with
small variations in δiso. Among this series, a crystal structure is
only available for 4. This compound does not exhibit halogen
bonding; rather the covalently bound chlorines on adjacent
molecules in the crystal structure are oriented somewhat
toward each other rather than toward the halide ion. The 35Cl
NMR data allow us to make two important points. First, it is
probable that compounds 5 and 6 are isostructural with 4, pack
in the P21/n space group, and do not exhibit halogen bonding.
Second, assessment of all of the data in Table 3 shows that the
values of CQ, δiso, and Ω for 4, where it is known independently
that halogen bonding is absent, do not appear to be
unequivocally sensitive to halogen bonding such that its
presence or absence may be inferred simply by the value of
one of these parameters.
The 4-halosubstituted chlorides (i.e., 7, 8, and 9) all have
known X-ray crystal structures; 7 and 8 pack in the same space
group and exhibit similar close contacts around the chloride
ion, while 9 packs in a different space group (Table 1). This is
distinctly reflected in the form of their 35Cl NMR spectra
(Figure 4) and in the values of CQ(35Cl) and ηQ. For example,
ηQ = 0.56 for 7, 0.60 for 8, and 0.38 for 9. Compound 7 does
not feature a significant halogen bond, as quantified by the R
value of 1.02, while 8 features a long, weak halogen bond (R =
0.98) and 9 features a slightly more substantial halogen bond
(R = 0.90). The small difference in the value of R for the two
isostructural compounds (7 and 8) is likely not enough to
manifest conclusively in the value of δiso.
Overall, for the two series of chloride compounds exhibiting
halogen bonding (1, 2, 3, and 7, 8, 9), it is tempting to correlate
the value of CQ(35Cl) with the degree of halogen bonding (as
quantified by R). It is seen that within these two small series,
larger values of CQ tend to be observed for weaker halogen
bonds. This is a tentative conclusion for various reasons: other
Several important points may be made concerning the 35Cl
NMR spectra. Under stationary conditions, the spectra can
span up to several thousand ppm at 9.4 T (see e.g., Figure 4c).
By increasing the applied magnetic field strength from 9.4 to
21.1 T, the spectra narrow considerably and the signal-to-noise
ratio increases. This is concomitant with the fact that the
spectra are dominated by the 35Cl quadrupolar interaction
between the quadrupole moment of the nucleus and the
electric field gradient (EFG) at the nucleus, rather than the
chlorine chemical shift anisotropy. As shown in Table 3, the
value of CQ(35Cl) ranges from 2.12 MHz in 3 to 6.04 MHz in 1,
which fits well within the range known for hydrochloride
compounds not exhibiting halogen bonding, that is, near zero
to less than 10 MHz.16 Isotropic 35Cl chemical shifts for the
chloride compounds range from 67 ppm in 8 to 88 ppm in 4,
and CS tensor spans range from 40 ppm in 3 up to 120 ppm in
9. These data also fit within the known ranges for hydro-
chloride salts found in the literature. It is evident that there is
no clear-cut NMR spectral signature for the presence of
halogen bonding in these compounds. For example, attempts to
correlate the NMR parameters with the degree of linearity of
the halogen bond, the XB distance, or the R value, were
unsuccessful. Even compound 4, which does not exhibit a
halogen bond, does not show particularly distinctive 35Cl NMR
parameters. As discussed below, however, an analysis of various
trends in the NMR parameters in the context of some of the
known crystal structure data reveals useful insights into how
these parameters are characteristic of chloride ions which
participate as halogen bond electron density donors.
The NMR spectra and data for the 2-halosubstituted
compounds 1, 2, and 3 betray some interesting information.
The crystal structures of 1 and 2 are known: both pack in the
Pca21 space group and exhibit similar carbon−halogen
distances and near identical halogen bond angles. There is a
difference of 0.1 Å in the halogen bond length and a difference
of 5% in the R value, with the stronger, shorter halogen bond
observed for 2-bromoanilinum chloride. The 35Cl NMR data
are consistent with this information: the form of the second-
order quadrupolar lineshapes for 1 and 2 (see Figure 2) and the
corresponding quadrupolar parameters (Table 3) are identical
within experimental error. The same can be said for the CS
tensor span. A small difference of 9 ppm is observed for the
value of δiso(35Cl), however, with the larger shift observed for
compound 2, which has the stronger halogen bond. Compound
1647
dx.doi.org/10.1021/cg201683p | Cryst. Growth Des. 2012, 12, 1641−1653

Crystal Growth & Design
Article
Figure 5. (a−f) Bromine-81 SSNMR spectra of 10, 11, and 12. Sharp low-frequency peaks due to copper in the NMR coil are seen in (b) and (d).
The MAS NMR spectrum shown in (g) is for compound 12. See the Experimental section for further details on how the data were acquired and
processed.
factors can influence CQ, particularly when the compounds pack
in different space groups. Perhaps the most important point to
consider is that the hydrogen bonding environment around the
chloride ions may also change from compound to compound,
in addition to the halide ion bonding environment. The
relevant N(H)···halogen distances are tabulated in the
Supporting Information (Table 2S).
The Euler angles relating the orientations of the EFG and CS
tensor principal axis systems (PAS) for these compounds
consistently show that the largest components of these tensors
are oriented at approximately 90 degrees to each other, as
shown by the values of β in Table 3.
(iii). Bromine-81 Solid-State NMR Spectroscopy. The
81Br solid-state NMR spectra of compounds 10, 11, 12, 13, 14,
16, and 17 are presented in Figures 5 and 6. Experimental
quadrupolar and CS tensor parameters determined from
spectral simulations are given in Table 4. Line shape
broadening due to the quadrupolar interaction dominates the
NMR spectra. The quadrupolar coupling constants, CQ(81Br),
range from 12.3 MHz in 12 to 45.3 MHz in 13. Several of these
values are notably larger than those obtained for other organic
hydrobromides where halogen bonding is absent (i.e., CQ(81Br)
< 20 MHz).20,43 The isotropic chemical shifts range from 100
ppm to 250 ppm and the CS tensor spans range from 75 ppm
to 340 ppm. The isotropic chemical shifts are within the known
range for bromide ions,16 while the spans represent some of the
first such data for bromide ions in organic hydrobromides.16,18a
Interestingly, compound 14, which does not exhibit halogen
bonding, has the highest chemical shift (250 ppm) and the
largest chemical shift tensor span (340 ppm).
structures are available for 10 and 11, but not for 12.
Compounds 10 and 11 are isostructural and pack in the Pca21
space group. Both exhibit weak halogen bonding as
characterized by R values of 0.97 and 0.90, respectively. The
identical values of CQ(81Br) of about 38 MHz for 10 and 11,
and very similar values of ηQ (0.31 and 0.27) reflect the fact that
these compounds pack in the same space group. The 81Br
NMR spectrum and quadrupolar parameters for 12, when
contrasted with those for 10 and 11, strongly imply that 12 is
not isostructural with 10 and 11 and that it therefore packs in a
different space group. Compound 11, which exhibits a stronger
halogen bond than 10, has larger values of δiso and Ω compared
with 10. In the case of δiso, this mirrors the trend observed for
the 35Cl chemical shift of the isostructural chloride analogues 1
and 2 (vide supra). These observations show that isomorphism
is important when attempting to investigate trends among the
various compounds; otherwise, too many additional changes in
the packing of the ions in the crystal can influence the observed
NMR parameters.
Interestingly, the 81Br NMR data for 17 clearly show the
presence of two crystallographically nonequivalent bromide
sites. This is the only compound in this study which exhibits
two halide sites. Unfortunately, attempts to grow single crystals
of 17 were unsuccessful.
(iv). Validation of Solid-State NMR Results through
Additional Single-Crystal X-ray Crystallography Studies.
In order to validate the findings obtained via chlorine and
bromine solid-state NMR spectroscopy regarding the iso-
morphic nature of certain series of compounds (vide supra), the
crystal structures of 3 and 12 were solved at the conclusion of
this study using single-crystal X-ray diffraction. As shown in
Table 1, compound 3 indeed packs in a different space group
(P21/n) than do compounds 1 and 2. This finding confirms the
Spectra (Figure 5) and data for the 2-halosubstituted
bromide series (10, 11, 12) reflect the findings for the 2-
halosubstituted chloride series (1, 2, 3), vide supra. Crystal
1648
dx.doi.org/10.1021/cg201683p | Cryst. Growth Des. 2012, 12, 1641−1653

Crystal Growth & Design
Article
Figure 6. (a−f) Bromine-81 SSNMR spectra of 13, 14, 16, and 17. See the Experimental section for further details on how the data were acquired
and processed.
a
Table 4. Experimental 81Br Quadrupolar and Chemical Shift Tensor Parameters
compound
|CQ(81Br)|/MHz
ηQ
δiso/ppm
Ω/ppm
κ
α (°)
β (°)
γ (°)
10
37.95(30)
38.0(2)
12.3(2)
45.3(1)
27.7(2)
26.2(1)
20.0(1)
26.2(1)
0.31(5)
0.27(1)
0.85(1)
0.665(5)
0.67(5)
0.66(1)
1.0(1)
135(3)
155(10)
187(5)
210(20)
240(10)
75(10)
0.2(1)
0.0(5)
90(25)
45(2)
65(3)
0(25)
90(5)
90(5)
90(5)
90(2)
85(5)
75(10)
90(10)
80(5)
0(2)
0(10)
0(5)
11
12
0.2(2)
13
130(10)
250(10)
230(10)
100(10)
150(10)
120(30)
340(10)
220(30)
−0.7(2)
−0.2(1)
0.4(1)
0(15)
0(2)
14
16
15(10)
17 (site 1)
17 (site 2)
0.92(2)
190(20)
−1.0(1)
5(3)
10(3)
12(3)
a
Error bounds are in brackets. Parameter definitions are in Table 3. Although CQ may take any real value, |CQ| is measured using SSNMR
experiments reported here.
conclusion reached by 35Cl solid-state NMR. Similarly,
compound 12 was also found to pack in the P21/n space
group, which is different from those of compounds 10 and 11.
This corroborates the conclusion reached by 81Br solid-state
NMR. 3 and 12 exhibit halogen bonding motifs analogous to 1
and 2 and to 10 and 11; however, a water of hydration is also
present in the former structures. The halide ions in 3 and 12
are therefore additionally coordinated by hydrogen bonds from
these water molecules. In summary, X-ray crystallography has
provided an independent cross-validation of the conclusions
1649
dx.doi.org/10.1021/cg201683p | Cryst. Growth Des. 2012, 12, 1641−1653

Crystal Growth & Design
Article
reached regarding the packing differences in two series of
compounds via halogen solid-state NMR.
CQ(127I) is 152.50 MHz for 20. The error inherent in the value
obtained using second-order perturbation theory to fit the
spectrum is estimated at less than 0.2%.47 The distinctly
different values of CQ and ηQ for 19 and 20 strongly suggest
that these two compounds are not isomorphous. Overall, the
data obtained for the two compounds are consistent with the
limited 127I solid-state NMR data available for iodide ions in
noncubic environments.45
(v). Iodine-127 Solid-State NMR Spectroscopy. Semin
and co-workers demonstrated the existence of a frequency shift
in the 127I NQR spectra of fluorinated diiodoalkanes upon
complexation with amines (i.e., a C−I···N halogen bonding
interaction where iodine is the electron density acceptor) and
also discussed how these frequency shifts may be explained
using a simple molecular orbital approach.21,22 Here, we discuss
briefly the feasibility of acquiring 127I solid-state NMR spectra
of iodide anions, where the iodine is the halogen bond electron
density donor rather than the acceptor. Iodine-127 NMR
spectra are difficult to acquire due to the typically extreme
spectral breadths associated with the central transition for all
but the most symmetric iodide environments.44−46 We were
able to acquire spectra for compounds 19 and 20 (Figure 7).
The results show that, while technically challenging, 127I
solid-state NMR of the iodide anions involved in halogen
bonds may be successfully applied in favorable cases to gain
some insight into local structure and symmetry. As for chlorine
and bromine, the most unambiguous interpretations of the
iodine NMR data will arise when comparing data for
compounds within an isostructural series.
(vi). Quantum Chemical Calculations. The results of
GIPAW DFT calculations of the chlorine and bromine
quadrupolar and magnetic shielding tensors are reported in
Tables 6 and 7. Graphical correlations between the computed
and experimental data are presented in Figure 8. We discuss the
results concisely, and then apply the results to study relevant
trends in the following section. The values of CQ(35Cl) are
consistently underestimated by the calculations. We speculate
that this may be due in part to the fact that it was not
computationally feasible with our resources to optimize the
hydrogen atom positions prior to the calculation of the NMR
parameters for all compounds. It has been shown previously
that chlorine quadrupolar coupling constants in organic
hydrochlorides are sensitive to the positions of nearby
hydrogen atoms.48 There is also a possibility that the halogen
bonding interaction may not be adequately described using
these computational methods. For instance, there has been
some debate over the relative roles of dispersion, electrostatics,
etc. in halogen bonds.49 Calculated quadrupolar asymmetry
parameters, CS tensor spans, and CS tensor skews are in fair to
good agreement with experiment. Calculated isotropic chemical
shifts are overestimated by approximately 40−50 ppm when
compared to experiment. This situation could possibly be
improved by using a computed absolute shielding value for the
reference compound, rather than the experimental chlorine
shielding scale;39 however, this is not our preferred method as it
can hide deficiencies with the computational method. Given
that the hydrogen atom positions modeled onto the structures
during the X-ray structure refinement process are not accurate
(usually these bond lengths are too short), we optimized the
hydrogen atoms computationally in cases where this was
feasible. As shown in Figure 8 (data represented by crosses;
values given as Supporting Information), there was no
improvement in agreement with experiment for the quad-
rupolar coupling constant or the CS tensor span (the
experimental data go from being underestimated to being
overestimated), while the isotropic chemical shifts became
more severely overestimated.
The computed quadrupolar coupling constants for 81Br also
tend to be low relative to experiment; however, the differences
between experimental and calculated values are less systematic
than for chlorine. The utility of these calculations in exploring
trends in several NMR tensor parameters is explored in an
example below.
(vii). Insight into the Halogen Bonding Environment:
An Example Using a Series of Isostructural Bromides. To
explore further whether insight into the XB environment
around the halide ions in the haloanilinium halides may be
Figure 7. (a−c) Iodine-127 SSNMR spectra of 19 and 20. See the
Experimental section for further details on how the data were acquired
and processed.
Note that the spectrum of 20 obtained at 21.1 T required 17
individual subspectra and that the total breadth of the central
transition spectrum is on the order of 30000 ppm. Spectral
parameters are presented in Table 5, where it is seen that
1650
dx.doi.org/10.1021/cg201683p | Cryst. Growth Des. 2012, 12, 1641−1653

Crystal Growth & Design
Article
a
Table 5. Experimental 127I Quadrupolar and Chemical Shift Tensor Parameters
compound
|CQ(127I)|/MHz
ηQ
δiso/ppm
Ω/ppm
κ
α (°)
β (°)
γ (°)
19
20
57.50(75)
0.775(20)
0.235(20)
340(10)
300(100)
<1000
−1.0(0.3)
n/a
0(0)
n/a
85(5)
n/a
25(5)
n/a
b
152.50(25)
500(50)
a
Error bounds are in brackets. Parameter definitions are in Table 3. Although CQ may take any real value, |CQ| is measured using SSNMR
b
experiments. Because of the large value of CQ, an exact treatment using of the Zeeman-quadrupolar Hamiltonian, rather than second-order
perturbation theory, was used for spectral simulations for 20.
Table 6. GIPAW DFT Calculated 35Cl Quadrupolar and Chemical Shift Tensor Parameters
compound
CQ(35Cl)/MHz
ηQ
δiso/ppm
Ω/ppm
κ
α (°)
β (°)
γ (°)
1
2
4
7
8
9
4.21
4.25
3.83
3.43
3.91
3.36
0.42
0.54
0.13
0.65
0.69
0.34
113.2
132.1
136.2
124.7
102.6
121.0
100.2
96.2
94.5
74.0
78.1
101.0
−0.41
−0.60
−0.65
−0.38
0.07
49.3
58.4
85.3
16.9
73.4
74.7
29.3
13.6
7.1
18.1
108.1
32.6
2.1
159.7
329.9
209.6
140.5
82.9
0.26
Table 7. GIPAW DFT Calculated 81Br Quadrupolar and Chemical Shift Tensor Parameters
compound
CQ(81Br)/MHz
ηQ
δiso/ppm
Ω/ppm
κ
α (°)
β (°)
γ (°)
10
11
13
14
18
−30.94
−35.49
−38.17
−14.48
−14.10
0.61
0.23
0.87
0.68
0.82
30.9
91.4
10.3
91.5
81.6
203.9
213.5
179.8
235.7
90.0
−0.29
−0.64
−0.71
0.22
53.5
190.6
184.0
244.5
80.8
85.6
84.9
62.6
83.6
83.1
358.8
7.7
345.1
28.7
10.7
0.08
gleaned from the CS tensor, relationships between CS tensor
parameters and various distances and angles characteristic of
the Br− coordination environment were assessed. Since halide
ions are often involved in several weak interactions
simultaneously, and as differences in crystal packing (i.e.,
space group) may also influence the NMR parameters, it is
essential to select a subset of compounds wherein the major
structural variable from compound to compound is the halide−
halogen XB distance. 2-Chloroanilinium bromide, 2-bromoani-
linium bromide, and 3-chloroanilinium bromide (10, 11, 13)
satisfy this requirement and they all pack in the Pca21 space
group. The bromide ion is involved in three hydrogen bonds
whose distances are essentially invariant in these three
compounds; that is, the three shortest unique Br−(H)N
distances vary by only 0.03, 0.02, and 0.03 Å over the three
compounds (see Supporting Information, Table 2S). The
major feature that changes is the Br−−X distance, which
increases from 3.443 Å in 2-bromoanilinium bromide to 3.589
Å in 2-chloroanilinium bromide to 3.838 Å in 3-chloroanilinium
bromide. As shown in Figure 9, an inverse correlation between
the value of Ω(81Br) and the bromide ion with the shortest
halogen−halide distance is observed. The span for 3-
chloroanilinium bromide, 120 ppm, is particularly valuable
because the Br−−Cl distance is comparable to the sum of their
van der Waals’ radii, and so this value is reflective of the limiting
case of a very weak XB interaction in this series of compounds
(R = 1.03). The isotropic bromine chemical shifts also follow a
similar trend (Figure 9), which interestingly parallels the trend
Ω(81Br) are of the correct order of magnitude and clearly
follow the experimental trend (Figure 9). Computed bromine
magnetic shielding constants also substantiate the experimental
trend in bromine chemical shifts (Table 7). Second, the unit
cell dimensions were adjusted by a few percent above or below
the experimental values for compound 10 such that intra-
molecular bond distances and angles remained fixed at their
experimental values, but the length of the intermolecular
halogen bond was varied. The goal of these calculations was to
assess as directly as possible the influence of a change in the
halogen bond distance while maintaining the impact of
including the full crystal lattice on the computed parameters.
The trends obtained thusly are consistent with those shown in
Figure 9 (see Supporting Information). As a final computa-
tional verification of the trend shown in Figure 9, a cluster
model of 2-chloroanilinium bromide, which included the
central bromide ion and the aromatic moieties which are
involved in hydrogen and/or halogen bonding with the central
ion, was built on the basis of the known atomic coordinates and
the bromide−chlorine distance was systematically varied by
displacing an intact haloanilinium group, while leaving the
remainder of the model unchanged (see Supporting Informa-
tion, Figure 3S). These B3LYP/6-311++G** results also
unequivocally show that Ω(81Br) and δiso(81Br) decrease with
decreasing XB strength (as quantified by the normalized
distance parameter, R), further substantiating the experimental
findings (see Supporting Information, Figure 3S).
Despite all of the above evidence, it is critical to point out
that small changes in hydrogen bond lengths among 10, 11,
and 13 (albeit an order of magnitude smaller than the changes
in the XB distance; see Supporting Information, Table 2S) may
also be playing an important role in the observed trends in the
NMR parameters. Indeed, additional calculations (see Support-
ing Information, Figures 1S and 2S) show that small changes in
the hydrogen bonding distances will also influence the NMR
parameters in the same manner as observed experimentally.
13
1
observed for H chemical shifts in hydrogen bonds.
Given the error bars on the experimental data, we set out to
computationally corroborate the experimental findings and to
verify that the substantial change in Ω(81Br) could be correlated
with a change in the halogen−halide XB length. First, the
experimental trend was confirmed via GIPAW DFT calcu-
lations using the experimentally known crystal structure
parameters and atomic coordinates. Computed values of
1651
dx.doi.org/10.1021/cg201683p | Cryst. Growth Des. 2012, 12, 1641−1653

Crystal Growth & Design
Article
Figure 9. Plot of experimental bromine chemical shifts (red squares),
bromine chemical shift tensor spans (blue diamonds), and GIPAW-
DFT calculated spans (blue circles) as a function of the shortest X−
Br− distance (d) in 2-chloroanilinium bromide, 2-bromoanilinium
bromide, and 3-chloroanilinium bromide. The line indicates a linear fit
to the experimental span data; Ω/ppm = −310d/(Å) + 1313;
correlation coefficient R = 0.99.
electron donor (selenium) is remarkably sensitive to the
halogen bonding interaction.23
CONCLUSIONS
■
This study has demonstrated that it is possible to directly
probe, by halogen solid-state NMR, the halide ion electron
density donors in a series of halogen-bonded haloanilinium
halides. Specifically, chlorine-35, bromine-81, and iodine-127
quadrupolar and chemical shift tensors have been measured for
a large series of compounds exhibiting weak halogen bonding as
specified by a normalized distance parameter (R values ranging
from 0.88 to 1.03). The NMR data for certain series of related
compounds are strongly indicative of when such compounds
pack in the same space group, thereby affording useful
structural information. X-ray crystallography provided an
independent cross-validation of the conclusions reached
regarding the packing differences in two series of compounds
via halogen solid-state NMR spectroscopy.
Although we found no unequivocal halide solid-state NMR
spectral signature for the presence or absence of the weak XB
interaction, careful interpretation of the NMR data in the
context of new and previously available crystallographic data,
and new GIPAW DFT calculations, has revealed several
remarkable trends. In particular, when a series of related
compounds pack in the same space group, it has been possible
to interpret trends in the NMR data in terms of the strength of
the halogen bond. For example, in isostructural series, the
halide quadrupolar coupling constant was found to increase as
the halogen bond weakens. This correlation is admittedly weak
for the compounds studied here. In the case of a series of
haloanilinium bromide compounds, it was found that the 81Br
isotropic chemical shift and chemical shift tensor span both
decrease as the bromide−halogen XB is weakened. These
trends were reproduced using both GIPAW DFT and cluster-
model calculations of the bromide ion magnetic shielding
tensor. Such trends are particularly interesting given the
important role which NMR has played historically in the
characterization of hydrogen bonding. Future work on more
strongly halogen bonded complexes will help to further
generalize our findings.
Figure 8. Comparison between (a) the experimentally determined
quadrupolar coupling constants (x axis) and the GIPAW DFT-
calculated values (y axis) for the compounds for which both data exist
(computed data are based on experimental X-ray structures without
any computational optimization). Data for six of the chloride salts are
⧫
presented as filled diamond symbols ( ) and for four of the bromide
salts, they are shown as filled squares. Analogous isotropic chemical
shift and CS tensor span data are presented in panels (b) and (c),
respectively. The lines of best fit for the bromine data are represented
by solid lines. CQ(calc) = 1.39CQ(expt) − 22.10; correlation coefficient
R = 0.949. Ω(calc) = 0.255Ω(expt) + 150.3; R = 0.998. δiso(calc) =
0.532δiso(expt) − 33.07; R = 0.714. The dashed lines represent y = x.
Selected additional GIPAW DFT data for the chlorides are also
included as crosses in each of the three panels; these values were
calculated after computational optimization of the hydrogen atom
positions.
Therefore, while it is not possible to conclusively attribute the
changes in the bromine CS tensor in the series 10, 11, and 13
solely to changes in halogen bonding, the experimental and
computational data are consistent with halogen bonding
playing a role that is at least complementary to the role of
hydrogen bonding in determining the bromide NMR
parameters in these compounds. Interestingly, however, we
have recently clearly shown in the case of selenium−iodine
halogen bonds, in systems where hydrogen bonding is
completely absent, that the chemical shift tensor span of the
1652
dx.doi.org/10.1021/cg201683p | Cryst. Growth Des. 2012, 12, 1641−1653

Crystal Growth & Design
Article
(17) (a) Chapman, R. P.; Bryce, D. L. Phys. Chem. Chem. Phys. 2007,
9, 6219. (b) Bryce, D. L.; Sward, G. D.; Adiga, S. J. Am. Chem. Soc.
2006, 128, 2121.
(18) (a) Widdifield, C. M.; Bryce, D. L. J. Phys. Chem. A 2010, 114,
2102. (b) Widdifield, C. M.; Bryce, D. L. Phys. Chem. Chem. Phys.
2009, 11, 7120.
(19) Hamaed, H.; Pawlowski, J. M.; Cooper, B. F. T.; Fu, R.;
Eichhorn, S. H.; Schurko, R. W. J. Am. Chem. Soc. 2008, 130, 11056.
(20) Gordon, P. G.; Brouwer, D. H.; Ripmeester, J. A. Chem. Phys.
Chem. 2010, 11, 260.
(21) Semin, G. K.; Babushkina, T. A.; Khrlakyan, S. P.; Pervova, E. Y.;
Shokina, V. V.; Knunyants, I. L. Theor. Expt. Chem. 1970, 4, 179.
Translated from Teor. Eksp. Khim. 1968, 4, 275.
ASSOCIATED CONTENT
■
S
* Supporting Information
Table showing additional information on the GIPAW DFT
calculations; table of relevant hydrogen bond distances;
additional computational results; crystallographic information
files for 2, 3, 12, and 13; complete reference 38. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Tel: +1 613 562 5800 ext 2018. Fax: +1 613 562 5170. E-mail:
dbryce@uottawa.ca.
(22) Maksyutin, Y. K.; Babushkina, T. A.; Guryanova, Y. N.; Semin,
G. K. Theor. Chim. Acta (Berlin) 1969, 14, 48.
(23) Viger-Gravel, J.; Korobkov, I.; Bryce, D. L. Cryst. Growth Des.
2011, 11, 4984.
(24) Pickard, C. J.; Mauri, F. Phys. Rev. B 2001, 63, 245101.
(25) Yates, J. R.; Pickard, C. J.; Mauri, F. Phys. Rev. B 2007, 76,
024401.
(26) Profeta, M.; Mauri, F.; Pickard, C. J. J. Am. Chem. Soc. 2003,
125, 541−548.
(27) Clark, S. J.; Segall, M. D.; Pickard, C. J.; Hasnip, P. J.; Probert,
M. I. J.; Refson, K.; Payne, M. C. Z. Kristallogr. 2005, 220, 567.
(28) Gray, L.; Jones, P. G. Z. Naturforsch. 2002, 57b, 61.
(29) APEX Software Suite v.2010; Bruker AXS: Madison, WI, 2005.
(30) Blessing, R. Acta Crystallogr. 1995, A51, 33.
(31) Sheldrick, G. M. Acta Crystallogr. 2008, A64, 112.
(32) Massiot, D.; Farnan, I.; Gautier, N.; Trumeau, D.; Trokiner, A.;
Coutures, J. P. Solid State Nucl. Magn. Reson. 1995, 4, 241.
(33) Eichele, K.; Wasylishen, R. E. WSOLIDS NMR Simulation
ACKNOWLEDGMENTS
■
D.L.B. thanks the Natural Sciences and Engineering Research
Council (NSERC) of Canada for funding. C.M.W. thanks
NSERC for an Alexander Graham Bell CGS D2 scholarship.
Dr. Victor Terskikh, Dr. Eric Ye, Dr. Glenn Facey, and Dr.
Andy Lo are thanked for technical support. Access to the 900
MHz NMR spectrometer was provided by the National
Ultrahigh-Field NMR Facility for Solids (Ottawa, Canada), a
national research facility funded by the Canada Foundation for
́
Innovation, the Ontario Innovation Trust, Recherche Quebec,
the National Research Council Canada, and Bruker BioSpin
and is managed by the University of Ottawa (www.nmr900.ca).
NSERC is acknowledged for a Major Resources Support grant.
Package, version 1.17.30; Universitat Tubingen: Germany, 2001.
̈
(34) Ploug-Sørenson, G.; Andersen, E. K. Acta Crystallogr. 1985, C41,
613.
REFERENCES
■
(1) Desiraju, G. R.; Steiner, T. The Weak Hydrogen Bond In
Structural Chemistry and Biology. IUCr Monographs on Crystallog-
raphy; Oxford University Press: Oxford, 1999; Vol. 9.
(2) Halogen Bonding: Fundamentals and Applications; Metrangolo, P.;
Resnati, G., Eds.; Mingos, D. M. P., Series Ed.; Springer-Verlag: Berlin,
2008; Structure and Bonding, Vol. 126.
(35) Portalone, G. Acta Crystallogr. E 2005, E61, o3083.
(36) Raatikainen, K.; Cametti, M.; Rissanen, K. Beilstein J. Org. Chem.
2010, 6 DOI: 10.3762/bjoc.6.4.
(37) Liu, Z.; Yu, W.-T.; Tao, X.-T.; Jiang, M.-H.; Yang, J.-X.; Wang,
L. Z. Kristallogr. NCS 2005, 220, 415.
(38) Frisch, M. J. et al. Gaussian 09, Revision A.02; Gaussian, Inc.:
Wallingford, CT, 2009.
(39) Gee, M.; Wasylishen, R. E.; Laaksonen, A. J. Phys. Chem. A
(3) Gokel, G. W.; De Wall, S. L.; Meadows, E. S. Eur. J. Org. Chem.
2000, 2000, 2967.
(4) Plevin, M. J.; Bryce, D. L.; Boisbouvier, J. Nat. Chem. 2010, 2,
466.
(5) Gilli, G.; Gilli, P. The Nature of the Hydrogen Bond: Outline of a
Comprehensive Hydrogen Bond Theory. IUCr Monographs on
Crystallography; Oxford University Press: Oxford, 2009, Vol. 23.
(6) Metrangolo, P.; Meyer, F.; Pilati, T.; Resnati, G.; Terraneo, G.
Angew. Chem. Intl. Ed. 2008, 47, 6114.
(7) Metrangolo, P.; Carcenac, Y.; Lahtinen, M.; Pilati, T.; Rissanen,
K.; Vij, A.; Resnati, G. Science 2009, 323, 1461.
(8) Sarwar, M. G.; Dragisic, B.; Salsberg, L. J.; Gouliaras, C.; Taylor,
M. S. J. Am. Chem. Soc. 2010, 132, 1646.
(9) Sarwar, M. G.; Dragisic, B.; Sagoo, S.; Taylor, M. S. Angew. Chem.
Intl. Ed. 2010, 49, 1674.
(10) Arman, H. D.; Gieseking, R. L.; Hanks, T. W.; Pennington, W.
T. Chem. Commun. 2010, 46, 1854.
(11) Grzesiek, S.; Cordier, F.; Jaravine, V.; Barfield, M. Prog. Nucl.
Magn. Reson. Spectrosc. 2004, 45, 275.
1999, 103, 10805.
(40) Lommerse, J. P. M.; Stone, A. J.; Taylor, R.; Allen, F. H. J. Am.
Chem. Soc. 1996, 118, 3108.
(41) Bondi, A. J. Phys. Chem. 1964, 68, 441.
(42) Shannon, R. D. Acta Crystallogr. 1976, A32, 751.
(43) Alonso, B.; Massiot, D.; Florian, P.; Paradies, H. H.; Gaveau, P.;
Mineva, T. J. Phys. Chem. B 2009, 113, 11906.
(44) Bryce, D. L.; Widdifield, C. M.; Chapman, R. P.; Attrell, R. J.
Chlorine, Bromine, and Iodine Solid-State NMR. In NMR of
Quadrupolar Nuclei in Solid Materials; Wasylishen, R. E, Ashbrook,
S. E., Wimperis, S., Eds.; John Wiley and Sons Ltd.: Chichester, UK,
2011; doi: 10.1002/9780470034590.emrstm1214
(45) Widdifield, C. M.; Bryce, D. L. J. Phys. Chem. A 2010, 114,
10810.
(46) Wu, G.; Dong, S. Solid State Nucl. Magn. Reson. 2001, 20, 100.
(47) Widdifield, C. M.; Bain, A. D.; Bryce, D. L. Phys. Chem. Chem.
Phys. 2011, 13, 12413.
(48) Bryce, D. L.; Sward, G. D. J. Phys. Chem. B 2006, 110, 26461.
(49) (a) Riley, K. E.; Hobza, P. Cryst. Growth. Des. 2011, 11, 4272.
(b) Riley, K. E.; Hobza, P. J. Chem. Theory Comput. 2008, 4, 232.
(12) Aliev, A. E.; Harris, K. D. M. Probing Hydrogen Bonding in
Solids Using Solid-State NMR Spectroscopy. In Supramolecular
Assembly via Hydrogen Bonds I, Structure and Bonding; Mingos, D. M.
P., Ed.; Springer-Verlag: Berlin, 2004; Vol. 108.
(13) Wu, G.; Freure, C. J.; Verdurand, E. J. Am. Chem. Soc. 1998, 120,
13187.
(14) Ida, R.; De Clerk, M.; Wu, G. J. Phys. Chem. A 2006, 110, 1065.
(15) Bouchmella, K.; Dutremez, S. G.; Alonso, B.; Mauri, F.; Gervais,
C. Cryst. Growth Des. 2008, 8, 3941.
(16) Chapman, R. P.; Widdifield, C. M.; Bryce, D. L. Prog. Nucl.
Magn. Reson. Spectrosc. 2009, 55, 215.
1653
dx.doi.org/10.1021/cg201683p | Cryst. Growth Des. 2012, 12, 1641−1653