Polyfluorinated versus cationic multidentate halogen-bond donors: A direct comparison

Article abstract of DOI:10.1016/j.jfluchem.2013.02.027

Two polyfluorinated azo-linked halogen-bond donors were synthesized and their relative Lewis acidities were compared to a known dicationic azo-linked halogen-bond donor with almost identical geometrical arrangement of the iodine substituents. This structu

Full text of DOI:10.1016/j.jfluchem.2013.02.027

Accepted Manuscript
PII:
S0022-1139(13)00108-5
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.jfluchem.2013.02.027
FLUOR 8066
To appear in:
FLUOR
Received date:
Revised date:
Accepted date:
14-1-2013
22-2-2013
28-2-2013
Please cite this article as: S.M. Walter, S.H. Jungbauer, F. Kniep, S. Schindler,
E. Herdtweck, S.M. Huber, Polyfluorinated versus Cationic Multidentate Halogen-
Bond Donors: A Direct Comparison, Journal of Fluorine Chemistry (2013),
http://dx.doi.org/10.1016/j.jfluchem.2013.02.027
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Highlights:
• Synthesis of polyfluorinated multidentate halogen-bond donors with azo
linkage
• Geometry of iodine centers almost identical to pyridinium-based halogen-bond
donors
• Much weaker activation of test substrate compared to cationic variants
• Comparison of Lewis acidity by DFT calculations supports experiments
Page 1 of 13

Polyfluorinated versus Cationic Multidentate Halogen-Bond
Donors: A Direct Comparison
Sebastian M. Walter, Stefan H. Jungbauer, Florian Kniep, Severin Schindler, Eberhardt Herdtweck,
Stefan M. Huber*
* Department Chemie, Technische Universität München, Lichtenbergstraße 4, D-85747 Garching, Germany; Email:
stefan.m.huber@tum.de.
ABSTRACT. Two polyfluorinated azo-linked halogen-bond donors were synthesized and their relative Lewis
acidities were compared to a known dicationic azo-linked halogen-bond donor with almost identical geometrical
arrangement of the iodine substituents. This structural similarity was confirmed by the results of x-ray structural
analyses. In a benchmark reaction to gauge the activation potential of these halogen-based Lewis acids in halide
abstraction reactions, we found that the polyfluorinated compounds were markedly less active than the cationic
ones and did, in fact, show very little activation at all. Repeated measurements seem to indicate however, that
there is a small effect of the iodinated versus the non-iodinated polyfluorinated compound in their role as
potential activators. Quantum-chemical calculations confirmed the experimentally observed trend.
GRAPHICAL ABSTRACT.
1. Introduction
Halogen-bonds, i.e. non-covalent interactions between Lewis bases and compounds with electrophilic terminal
halogen substituents, have seen renewed interest in the last two decades.1,2 Traditionally, most of the applications
of halogen-bonds concern the solid phase,3 although several studies on halogen-bonds in solution have also been
published lately.4 Halogen-bonds share many similarities with hydrogen-bonds, but feature a markedly higher
directionality: the R-X -- LB angle (R = backbone; X = Cl, Br, I; LB = Lewis base) is always found to be close to
180° in halogen-bond-based complexes. A second prerequisite for the formation of a reasonably strong halogen-
bond adduct is the presence of electronegative substituents R in the halogen-bond donor R-X (i.e., the halogen-
based Lewis acid). In crystal engineering, this has mostly been achieved by the use of polyfluorinated backbones,
preferentially polyfluoroiodoarenes.1;5 Based on these building blocks, liquid crystalline and conductive
materials have e.g. been designed.6 Somewhat less prominent are cationic backbones,4c,7 which have e.g. been
used for the construction of catenanes and anion receptors.4g;5b,c;8 Recently, we could show that bidentate
imidazolium-,9a pyridinium-,9b or triazolium-based9c halogen-bond donors may be employed as activators for the
cleavage of a carbon-bromine bond (see Figure 1 for the iodopyridinium-based Lewis acid 1). In these reactions,
we found that neutral (polyfluorinated) halogen-bond donors like 1,4-diiodotetrafluorobenzene were not active in
the same timeframe.9a This could be due to a reduced electrophilicity of the C-I bonds of these compounds
(compared to cationic ones) or due to the monodentate binding of these Lewis acids (or both).
1
Page 2 of 13

Figure 1: Iodopyridinium-based bidentate halogen-bond donor 1.9b
In general, although both polyfluorinated and cationic multidentate halogen-bond donors have been used in
various applications (see above), it is difficult to directly compare the influence of the two types of backbones on
the electrophilicity of the corresponding halogen-bond donors. Because of the high directionality of the
interaction, a fair comparison would need to be based on topologically identical halogen-bond donors. Herein we
present, to the best of our knowledge, the first comparison of a bidentate polyfluorinated with a bidentate
dicationic halogen-bond donor, where both feature the same geometric arrangement of the electrophilic iodine
substituents.
2. Results and discussion
2.1 Synthesis of neutral multidentate halogen-bond donors 4a and 4b
Previously, we could show that halogen-bond donor 1 is a potent electrophile.9b Since compound 1 was
obtained by azo coupling of the corresponding aminopyridine precursors (and subsequent methylation), we
reasoned that a structurally very similar neutral halogen-bond donor should be accessible by azo coupling of two
polyfluorinated iodoaniline derivatives 2a (see Scheme 1). The latter ones could be synthesized by iodination of
3,4,5-trifluoroaniline with potassium iodide and potassium iodate in methanol, in analogy to a literature-known
procedure, in 82 % yield.10 Oxidative coupling of two anilines 2a with the powerful hypervalent iodine reagent
PhIL2(OTf)2 311 (L = 4-dimethylaminopyridine) gave the target structure 4a in 44 % yield after several days. This
rather moderate yield, which could not be improved, is likely due to the very electron-poor nature of aniline 2a.
Although it is disappointing that the synthesis of 4a could not be realized in higher yield, the amount of product
obtained was still sufficient for the intended benchmarking studies (see below). For comparison reasons, we also
synthesized the analogous tetrabromo-substituted halogen-bond donor 4b. The required precursor, 2,5-dibromo-
3,4,5-trifluoroaniline 2b, was obtained by bromination of 3,4,5-trifluoroaniline with N-bromosuccinimide in 95
% yield. Azo coupling with oxidant 3 again gave the target compound 4b in low to moderate yield (38 %),
despite a reaction time of several days at room temperature (decomposition was observed at higher
temperatures). Finally, the non-iodinated/brominated reference compound 4c could by synthesized by direct azo
coupling of 3,4,5-trifluoroaniline (32 % yield).
Scheme 1: Synthesis of polyfluorinated halogen-bond donors 4a and 4b; i) KI, KIO3, HCl, MeOH (X = I) or
N-bromosuccinimide, CH2Cl2 (X = Br); ii) PhIL2(OTf)2 (L = 4-dimethylaminopyridine), CH2Cl2.
All azo compounds could be obtained in pure form, as evidenced by elemental analysis and 19F NMR spectra.
In the latter, the chemical shift of the triplett (or multiplett) corresponding to the fluorine substituents at the para-
positions does only change by few ppm from 4a (-151 ppm) to 4c (-154 ppm). The doublet (or multiplett) of the
fluorine substituents at the meta-position, however, shifts markedly high-field from 4a (-107 ppm) via 4b (-123
ppm) to 4c (-132 ppm).
2
Page 3 of 13

2.2 X-ray structural analysis: Comparison of the orientations of the iodine substituents in compounds 4a and 1
Single crystals of compound 4a that were suitable for an X-ray structural analysis were obtained by slow
evaporation of the solvent from a solution of 4a in chloroform.12 Since the structural analysis of dicationic
halogen-bond donor 1 in the solid state has also been published,9b a direct comparison between these two
halogen-based Lewis acids is now possible (see Figures 2a and 2b).
Figure 2: X-ray structural analyses of halogen-bond donors 4a (a) and 19b (b) (ellipsoids at 50 % probability).
‒
In the case of 1, the anions have been omitted for clarity. Selected bond lengths [Å] and angles [°]: (a) N1 N2 =
‒
‒
‒
‒
‒
‒
‒
‒
‒
‒
1.250(3), C8 I3 = 2.088(2); C8 C7 N2 N1 = 132.1(2), C2 C1 N1 N2 = 129.6(2), C2 C1 C7 C8 = 61.8(3);
‒
‒
‒
‒
‒
‒
‒
‒
(b) N1 N1 = 1.242(3), C5 I2 = 2.090(3); C5 C1 N1 N1 = -58.8(3); C5 C1 C1 C5 =75.4(3); symmetry code
for equivalent atoms in 1: -x, y, 0.5 - z.
In both cases, the orientation of the two aryl moieties towards each other is very roughly perpendicular, with
an angle of about 60° between the two planes in the case of 4a (compared to 75° for compound 1). The azo
bridge forms a torsion angle of approx. 130° with an attached aryl group, also similar in both cases. The bond
length of the azo bridge itself is marginally higher for 4a compared to 1 (1.250 Å vs. 1.242 Å),13 and the C I
‒
bond lengths are also almost identical in both structures (2.088 Å vs. 2.090 Å). These data illustrate that the
structural environment of the Lewis acidic iodine substituents is indeed very similar for both halogen-bond
donors.
There is one marked difference in the solid-phase structures, however, namely the crystal packing
arrangement: in contrast to the dicationic compound 1, there are no further Lewis basic counterparts (i.e., anions)
to interact with for the neutral halogen-bond donor 4a. Thus, the solid-phase structure of 4a is built up by
interactions between different molecules of 4a (Figure 3).
…
‒
Figure 3: Intermolecular halogen-bonds formed in the crystal packing of 4a; selected angles [°]: C12 I4 I1 =
…
…
…
‒
‒
‒
68.67(6), C6 I1 I4 = 146.91(7), C8 I3 I2 = 76.13(6), C2 I2 I3 = 166.18(6). In the crystal structure of
3
Page 4 of 13

compound 4a, each of the four iodine substituents is connected to neighbouring molecules by one of the two
kinds of halogen-bonds shown in this figure.
In principle, there are two electron-rich kinds of substituents in 4a: the azo nitrogen atoms and the iodine
substituents. The azo group, however, does not engage in any short contact with other molecules of 4a, most
likely for steric reasons (including the shielding by the four surrounding iodine atoms). This leaves the iodine
atoms as the only remaining groups which could act as Lewis basic centers. A peculiar feature of such terminal
halogen substituents attached to highly electron-withdrawing groups is the anisotropy of their electron density:
‒
there is comparably little electron density in the extension of the C X bond (which is the reason for the
‒
electrophilicity of these atoms), but also increased electron density in a belt perpendicular to the C X bond
(which is due to the filled p-orbitals on the halogen atom). Thus, for an ideal intermolecular iodine-iodine
contact, one would expect the C‒I--I angles to be 180° at the iodine atom acting as the electrophile and to be 90°
at the iodine substituent acting as the nucleophile. In the solid-state structure of 4a, two short I-I contacts are
14
‒
found. The shorter of those, with an I I distance of 3.790 Å (see Figure 3; sum of van-der-Waals radii: 3.96 Å)
indeed features structural parameters close to the ones just discussed (with C-I--I angles of 166° and 76°). In the
‒
second intermolecular contact, the I I distance is 3.884 Å, and thus this interaction is apparently weaker than the
other contact. In line with this, the geometric parameters deviate even more strongly from the ideal values (with
‒
‒
C I--I angles of 147° and 69°). In fact, the rather strong deviation of the C I--I angle from 180° in the second
contact is in some contrast to the high directionality of halogen-bonds, and thus this interaction may not
constitute a “classical” halogen-bond, but rather a weak van-der-Waals contact. In the crystal packing of
compound 4a, each of the four iodine substituents is connected to neighbouring molecules of 4a by one of the
two types of halogen bonds described above. This situation is thus somewhat similar to the crystal packing of the
cationic halogen-bond donor 1, in which also all four halogen substituents of 1 form halogen bonds (to the
triflate anions, in this case).
‒
‒
Similar intermolecular C I--I C halogen-bonds as just described for the crystal structure of 4a have already
been reported in the literature, e.g. in the solid state structures of 1,3,5-trifluoro-2,4,6-triiodibenzene15 and
1,1,2,2-tetrafluoro-1,2-diiodoethane.16 In these cases, short I-I distances of 3.74 Å and 3.89 Å have been
observed, respectively.
2.3 Performance in a benchmark carbon-halogen bond activation reaction
As the structural similarity of the core part of compounds 4a and 1 was also confirmed by the x-ray analyses
presented above, we next turned our attention towards the activation potential of the neutral halogen-bond donors
4a and 4b in an actual chemical reaction. Previously, we had established the solvolysis of benzhydryl bromide
(5) in acetonitrile as a suitable benchmark reaction to test the activation of a carbon-bromine bond by halogen-
bond donors (Scheme 2).9a After heterolytic cleavage of the C-Br bond, the resulting cationic intermediate is
transformed into the amide 6 by attack of acetonitrile and subsequent hydrolysis in a Ritter-type reaction.
Importantly, it is possible to rule out traces of acid as the actual activating reagent by addition of a
substoichiometric amount of pyridine.9a For all previously examined halogen-bond donors, the corresponding
non-iodinated reference compounds were not similarly active in this reaction. The dicationic halogen-bond donor
1 had converted more than 90% of benzhydryl bromide to the amide after 36 hours at room temperature.9b,17
Scheme 2: Benchmark carbon-halogen bond activation reaction.
Most likely, the cationic backbone in compound 1 is more electronegative than the polyfluorinated one in 4a
(see also chapter 2.4) and thus the electrophilicity of the iodine substituents in halogen-bond donor 1 might be
expected to be higher than in 4a. The dicationic Lewis acid 1, however, invariably is accompanied by two
4
Page 5 of 13

triflate anions, which will compete with the substrate (5) and the liberated bromide for the electrophilic iodine
centers, an effect that will reduce the activation potential of the dicationic halogen-bond donor.
Figure 4 shows the formation of amide 6 in a solution of benzhydryl bromide in deuterated acetonitrile, in
which equimolar amounts of the neutral azo compounds 4a, 4b or 4c were present with otherwise identical
conditions. We have also added 10 mol-% of pyridine to each solution to quench any acid impurities. The yield
of amide 6 over time could be conveniently determined by 1H-NMR spectroscopy. For comparison, we have also
added the analogous graph for the formation of 6 in the presence of one equivalent of activator 1 from our
previous measurement.9b
Figure 4: Yield-versus-time profile of the solvolysis of benzhydryl bromide (see Figure 2) in the presence of
one equivalent of various potential activating reagents.
It is immediately obvious that the neutral halogen-bond donors are markedly less active in this benchmark
reaction than the cationic variant 1: while the solvolysis of benzhydryl bromide is almost quantitative in the
presence of 1 after 36 hours, only ca. 2% of amide were formed in the presence of halogen-bond donor 4a in the
same time. In fact, even after 25 days of reaction time, only ca. 18% of benzhydryl bromide were converted to
amide 6 in the presence of one equivalent of 4a. As this slow solvolysis reaction is only marginally more
pronounced than in the background reaction (see Figure 4), we repeated the measurement several times and
found the same trend in all measurements. While this does not completely rule out the possibility that the
observed difference is due to the margin of error of the measurements, it provides at least a weak indication that
compound 4a is indeed capable of activating the solvolysis of 5, even if marginally. With the same caution in the
interpretation of the results, it seems that the brominated compound 4b and the reference compound 4c are both
not capable of activating the substrate. This provides an, albeit weak, indication, that the electrophilicity of the
iodine centers is indeed responsible for the activation by 4a.18 The most important finding, however, remains the
fact that the polyfluorinated halogen-bond donors presented in this study are much less potent than analogous
cationic ones, where all structural parameters are virtually identical. This fact constitutes an important
orientation for the future design of potent halogen-bond based activators and catalysts.
2.4 Quantum-chemical calculations
In order to complement the experimental comparison between compounds 1 and 4a, we also performed
quantum chemical calculations on the complexes of 1 and 4a with bromide (which is liberated during the
benchmark reaction presented above). Due to the size of the complexes, we used density functional theory with
the M06-2X functional19 that has been recommended for non-covalent interactions and a triple-zeta basis set with
polarization functions (including the corresponding pseudopotential for iodine).20 All calculations were
performed with Gaussian09.21 The counteranions were not included in the case of compound 1. Since the
complex formation involves charges species, we added an intrinsic solvation model, namely the PCM method22
that is available in Gaussian09, using parameters for acetonitrile.
5
Page 6 of 13

First we were interested to see whether bidentate coordination of the halogen-bond donors to the halide is
geometrically possible.23 In both cases, optimization of the bidentate adducts resulted in true energetic minima,
as was evident from the absence of imaginary frequencies. In the case of Lewis acid 4a, the planes of the two
phenyl moieties form an angle of approx. 30°, to enable best-possible overlap of the electrophilic part of the
iodine atoms with the bromide. A very similar angle of the corresponding planes is observed in compound 1.9b
The anion is bound more closely in the dicationic halogen-bond donor than in the polyfluorinated one (see
‒ ‒
distances in Figure 5), indicating stronger halogen-bonds for the former. In both adducts, the C I Br angles
deviate somewhat from the ideal of 180° (156° and 162° for the complex with 4a, 159° for both bonds in the
adduct with 1). This indicates that although the backbones feature some flexibility, it is not possible to align both
iodine substituents for perfectly linear halogen-bonds, and thus a compromise is realized.
Figure 5: Adducts of the halogen-bond donors with bromide, as calculated with M06-
2X/TZVPP/PCM(CH3CN); a) complex of 4a with bromide; b) complex of 1 with bromide. The graphic
representations have been prepared with CYLview.24 C-I--Br halogen-bond angles: a) 156° (halogen-bond shown
left) and 162° (halogen-bond shown right); b) 159° (both halogen-bonds).
Additionally, we sought to obtain a rough estimate of the relative complexation energies of both complexes.25
We note that the calculation of absolute complexation energies (and Gibbs free energies) poses several
challenges due to the presence of counterions and solvation, and we are aware that these computations will have
only orientating character. Consequently, we were primarily interested in the energies of both adducts relative to
each other, as this should give a coarse indication of the relative Lewis acidities of the two halogen-bond donors.
We computed the complexation energy as the difference between the adduct and the isolated Lewis acid and
bromide in the intrinsic solvation environment. Adduct formation is predicted to be favourable in both cases,
namely by approx. 3 kcal/mol for the complex with 4a and by approx. 8 kcal/mol for the complex with 1.26 The
markedly higher complexation energy for the adduct with 1 indicates a higher Lewis acidity for this compound
and is in agreement with the experimental findings presented above.
3. Conclusions
In summary, we have synthesized azo-linked polyfluorinated halogen-bond donors 4a and 4b in order to
compare their relative Lewis acidities to the known halopyridinium-based azo compound 1, which features a
virtually identical geometric arrangement of the iodine substituents. Thus, for the first time it was possible to
directly study the influence of a cationic vs. a polyfluorinated backbone on the Lewis acidity of the
corresponding halogen-bond donor. A comparison of the results of x-ray structural analyses of halogen-bond
donors 4a and 1 illustrated the high geometrical similarity of the core part of these Lewis acids. Using a
previously established test reaction to gauge the activation potential of halogen-bond-based Lewis acids in a
solvolysis reaction, we found that the polyfluorinated compound 4a was markedly less active than 1 and did
indeed show very little activation of the substrate at all. Repeated measurements seem to indicate however, that
there is a small effect of the iodinated (4a) versus the non-iodinated (4c) compound in their role as potential
activators. Quantum-chemical calculations confirmed that 4a might act as a bidentate halogen-bond donor, but
6
also showed that its Lewis acidity is much lower than that of compound 1 (in agreement with experiment).
Page 7 of 13

These findings should be interesting for the future design of strong multidentate halogen-bond donors. In cases
were both a cationic and a polyfluorinated backbone are available for the synthesis of such a halogen-based
Lewis acid, it seems that the cationic core structure will lead to the stronger halogen-bond donor, in spite of the
fact that its counteranions will compete with the respective substrate for the Lewis-acidic halogen substituent.
Still, further comparisons with other multidentate binding motives will be necessary to gain a better
understanding towards the rational design of highly potent halogen-bond donors.
4. Experimental
General
Commercially available reagents and starting materials were used without further purification. Solvents were
used after single distillation. Nuclear magnetic resonance spectra (NMR-spectra) were obtained on instruments
of the type AV-500c from Bruker. Chemical shifts (δ) are given as parts per million (ppm) and refer to the shift of
the hydrogen, carbon or fluorine atoms in CDCl3. Reference (1H-NMR): residual CHCl3 (δ = 7.26 ppm);
reference (13C-NMR): residual CDCl3 (δ
=
77.0 ppm); reference (19F-NMR): external standard
hexafluorobenzene (δ = -162.0 ppm). The following abbreviations were used for the assignment of the signals
and their multiplicities: s (singulet), bs (broad singulet), ddd (doublet of doublets of doublets), dt (doublet of
triplets), q (quartet), m (multiplet). The given coupling constants J (in Hertz) are listed as the average of the
experimental findings. Mass spectrometry spectra (MS-spectra) were obtained by using electrospray ionization
(ESI) or gas chromatography (GC-MS). Assigned m/z-relations are listed in ascending order. Infrared-spectra
(IR-spectra) were measured directly from the substance via attenuated total reflectance (ATR-IR). The signals
are labelled with the following abbreviations: vs (very strong), s (strong), m (medium), w (weak), vw (very
weak).
Halide abstraction benchmark reactions
From stock solutions (in CD3CN), the respective substrate (10.0 µmol, 1.0 eq.) and, if applicable, the activating
reagent (10.0 µmol, 1.0 eq.) as well as, if applicable, pyridine (1.0 µmol, 0.1 eq.) were mixed in an NMR tube,
and the solution was filled up to an overall volume of 0.60 mL. All experiments were started simultaneously and
the yield was determined by NMR spectroscopy after approx. 1, 2, 4, 7, 10, 14, 18, 21 and 24 days of reaction
time.
4.1.1 3,4,5-Trifluoro-2,6-dibromoaniline (2b)
1.00 g (6.80 mmol) of 3,4,5-trifluoroaniline and 2.42 g (13.6 mmol, 2 eq.) of N-bromosuccinimide were
dissolved in 45 mL of CH2Cl2. The yellow solution was stirred at room temperature for 24 h, resulting in a
deeply violet solution and an off-white precipitate. The mixture was filtered over a silica pad, collecting a light-
brown and a violet fraction. The solvent was removed from the light-brown fraction to yield 2b as 1.95 g
(6.44 mmol, 95%) of an off-white solid.
1
Analytical data for 2b: H-NMR (500 MHz, CDCl3): δ (ppm) = 4.64 (bs); 13C-NMR (126 MHz, CDCl3): δ
(ppm) = 148.36 (ddd, 1JCF = 246.2, 2JCF = 12.1, 3JCF = 5.5 Hz), 139.17 – 138.56 (m), 133.03 (dt, 1JCF = 246.7, 2JCF
= 17.3 Hz), 91.33 – 91.00 (m); 19F-NMR (471 MHz, CDCl3): δ (ppm) = -125.5 (d, 3JFF = 22.5 Hz), -169.1 (t, 3JFF
= 22.6 Hz); Elemental Analysis: calculated C: 23.6, H: 0.66, N: 4.59, found C: 24.2, H:0.58, N: 4.46; MS (GC):
calculated for C6H2Br2F3N: 305, found: 305 [M], 226 [M - Br], 144 [M - 2·Br]; IR: 3485 (m), 3386 (m), 1625
(m), 1585 (s), 1480 (s), 1445 (vs), 1287 (m), 1115 (m).
4.1.2 3,4,5-Trifluoro-2,6-diiodoaniline (2a)
3.97 g (27.0 mmol) of 3,4,5-trifluoroaniline and 6.13 g (36.9 mmol) of potassium iodide were dissolved in a
mixture of 65 mL of methanol and 350 mL of water. 4.12 g (19.3 mmol) of potassium iodate were added to the
clear solution, resulting in a slowly clearing suspension. 36 mL of hydrochloric acid (1.5 M) were slowly added
over a period of 45 min. The mixture was stirred at room temperature under exclusion of light for 21 h. The off-
white precipitate was filtered off and washed twice with water. Upon dissolving the solid, an orange solution
7
Page 8 of 13

resulted. It was filtered over a silica pad, dried over MgSO4 and the solvent removed under reduced pressure to
yield a reddish-brown oil. From this oil, 8.80 g (22.1 mmol, 82%) of 2a were isolated by sublimation under high
vacuum as a white solid.
1
Analytical data for 2a: H-NMR (500 MHz, CDCl3): δ (ppm) = 4.78 (bs); 13C-NMR (126 MHz, CDCl3): δ
(ppm) = 151.37 (ddd, 1JCF = 243.3, 2JCF = 11.2, 3JCF = 6.0 Hz), 143.34 – 142.93 (m), 131.05 (dt, 1JCF = 248.6, 2JCF
= 18.7 Hz), 63.57 – 62.57 (m); 19F-NMR (471 MHz, CDCl3): δ (ppm) = -109.4 (d, 1JFF = 23.4 Hz), -167.2 (t, 2JFF
= 23.2 Hz); Elemental Analysis: calculated C: 18.1, H: 0.51, N: 3.51, found C: 18.4, H: 0.44, N: 3.49; MS (GC):
calculated for C6H2I2F3N: 399, found: 399 [M], 272 [M - I], 145 [M - 2·I]; IR: 3459 (m), 3367 (m), 1617 (m),
1592 (s), 1467 (vs), 1428 (s), 1280 (m), 1117 (s).
4.1.3 (E)-1,2-bis(3,4,5-trifluorophenyl)diazene (4c)
0.80 g (1.10 mmol) of PhIL2(OTf)2 (L = 4-dimethylaminopyridine)11 were suspended in 10 mL of absolute
CH2Cl2 under argon in a flame-dried Schlenk flask. 0.10 g (0.68 mmol) of 3,4,5-trifluoroaniline were added in
10 mL of CH2Cl2. The mixture was stirred at reflux for 3 h, and subsequently at room temperature for 24 h.
Afterwards, it was filtered over a silica pad with pentane. Solvent was removed under reduced pressure from the
first yellow fraction to yield 32 mg (0.11 mmol, 32%) of 4c as an orange solid.
Analytical data for 4c: 1H-NMR (500 MHz, CDCl3): δ (ppm) = 7.69 – 7.55 (m); 13C-NMR (126 MHz, CDCl3):
1
2
3
1
δ (ppm) = 151.75 (ddd, JCF = 252.5, JCF = 10.9, JCF = 4.2 Hz), 147.30 – 146.41 (m), 142.27 (dt, JCF = 259.1,
2JCF = 15.8 Hz), 108.19 – 107.70 (m); 19F-NMR (471 MHz, CDCl3): δ (ppm) = -132.32 – -132.12 (m), -153.9 – -
154.1 (m); Elemental Analysis: calculated C: 49.7, H: 1.39, N: 9.65, found C: 49.4, H: 1.34, N: 9.59; MS (ESI):
calculated for C12H4F6N2: 290, found: 290 [M], 159 [M - C6H3F3], 131 [M - C6H3N2F3]; IR: 3091 (w), 1624 (m),
1598 (m), 1509 (vs), 1444 (s), 1355 (vs), 1299 (m), 1230 (s), 1188 (m), 1111 (w), 1041 (vs), 990 (m), 875 (vs),
791 (vs), 706 (m), 663 (s).
4.1.4 (E)-1,2-bis(3,4,5-trifluoro-2,6-dibromophenyl)diazene (4b)
0.80 g (1.10 mmol) of PhIL2(OTf)2 (L = 4-dimethylaminopyridine)11 were suspended in 10 mL of absolute
CH2Cl2 under argon in a flame-dried Schlenk flask. 0.32 g (1.06 mmol) of 2b were added in 10 mL of CH2Cl2.
The mixture was stirred at room temperature for 3 d. Afterwards, it was filtered over a silica pad with CH2Cl2,
yielding a dark and a yellow fraction. The dark fraction was filtered over a silica pad with pentane. Solvent was
removed under reduced pressure to yield 126 mg (0.21 mmol, 38%) of 4b as a dark red solid.
1
2
3
Analytical data for 4b: 13C-NMR (126 MHz, CDCl3): δ (ppm) = 148.63 (ddd, JCF = 251.5, JCF = 12.0, JCF
=
4.0 Hz), 144.98 – 144.07 (m), 140.54 (dt, JCF = 262.4, JCF = 17.3 Hz), 101.00 – 100.45 (m); 19F-NMR
1
2
3
3
(471 MHz, CDCl3): δ (ppm) = -122.6 (d, JFF = 21.3 Hz), -151.3 (t, JFF = 21.4 Hz); Elemental Analysis:
calculated C: 23.8, H: 0.00, N: 4.62, found C: 24.0, H: < 0.1, N: 4.56; MS (ESI): calculated for C12Br4F6N2: 606,
found: 606 [M], 319 [M - C6HBr2F3], 289 [M - C6HN2Br3F3]; IR: 1602 (m), 1581 (m), 1483 (s), 1414 (s), 1324
(m), 1224 (m), 1072 (s), 883 (s), 723 (s), 696 (m).
4.1.5 (E)-1,2-bis(3,4,5-trifluoro-2,6-diiodophenyl)diazene (4a)
1.61 g (2.16 mmol) of PhIL2(OTf)2 (L = 4-dimethylaminopyridine)11 were suspended in 22 mL of absolute
CH2Cl2 under argon in a flame-dried Schlenk flask. 0.57 g (1.43 mmol) of 2a were added in 15 mL of CH2Cl2.
The mixture was stirred at reflux for 2 h, and subsequently at room temperature for 10 d. Afterwards, it was
filtered over a silica pad with CH2Cl2 and again with pentane. Solvent was removed under reduced pressure to
yield 256 mg (0.32 mmol, 44%) of 4a as a violet solid.
1
2
3
Analytical data for 4a: 13C-NMR (126 MHz, CDCl3): δ (ppm) = 151.98 (ddd, JCF = 248.4, JCF = 11.4, JCF
=
8
4.6 Hz), 146.33 – 145.24 (m), 138.80 (dt, 1JCF = 264.6, 2JCF = 18.7 Hz), 74.60 – 74.25 (m); 19F-NMR (471 MHz,
CDCl3): δ (ppm) = -107.4 (d, JFF = 21.9 Hz), -150.8 (t, 3JFF = 2.1 Hz); Elemental Analysis: calculated C: 18.2,
3
Page 9 of 13

H: 0.00, N: 3.53, found C: 18.8, H: < 0.1, N: 3.5; MS (ESI): calculated for C12I4F6N2: 794, found: 793 [M], 667
[M - I]; IR: 2924 (w), 2853 (vw), 1591 (m), 1465 (vs), 1401 (s), 1317 (m), 1208 (m), 1066 (vs), 858 (vs), 728
(w), 704 (m), 672 (s).
Acknowledgements
We are grateful to the Fonds der Chemischen Industrie (Liebig Fellowship to S.M.H.) and the Deutsche
Forschungsgemeinschaft for financial support. We also thank Prof. Dr. Thorsten Bach and his group for their
great support.
References
[1]
Selected reviews: (a) P. Metrangolo, H. Neukirch, T. Pilati, G. Resnati, Acc. Chem. Res. 38 (2005) 386;
(b) P. Metrangolo, F. Meyer, T. Pilati, G. Resnati, G. Terraneo, Angew. Chem. Int. Ed. 47 (2008) 6114;
(c) M. Fourmigué, Curr. Opin. Solid State Mater. Sci. 13 (2009) 36;
(d) A.C. Legon, Phys. Chem. Chem. Phys. 12 (2010) 7736;
(e) R.W. Troff, T. Mäkelä, F. Topic, A. Valkonen, K. Raatikainen, K. Rissanen, Eur. J. Org. Chem.,
DOI:10.1002/ejoc.201201512.
[2]
For recent theoretical investigations on halogen bonding, see for example: (a) K.E. Riley, J.S. Murray, J.
Fanfrlík, J. Rezác, R.J. Solá, M.C. Concha, F.M. Ramos, P. Politzer, J. Mol. Mod., DOI:
10.1007/s00894-012-1428-x.
b) P. Politzer, K.E. Riley, F.A. Bulat, J.S. Murray, Comput. Theor. Chem. 998 (2012) 2.
c) M.D. Esrafili, J. Mol. Model. 18 (2012), 5005.
d) M.D. Esrafili, N.L. Hadipour, Mol. Phys. 109 (2011) 2451.
e) M.D. Esrafili, J. Mol. Mod., DOI: 10.1007/s00894-012-1691-x
f) M. Palusiak, J. Mol. Struct. (Theochem) 945 (2010) 89.
g) P. Metrangolo, J.S. Murray, T. Pilati, P. Politzer, G. Resnati, CrystEngComm 13 (2011) 6593.
h) M.G. Chudzinski, M.S. Taylor, J. Org. Chem. 77 (2012) 3483.
i) S.M. Huber, E. Jimenez-Izal, J.M. Ugalde, I. Infante, Chem. Commun. 48 (2012) 7708.
Selected reviews: (a) P. Metrangolo, G. Resnati, Chem. Eur. J. 7 (2001) 2511;
(b) K. Rissanen, CrystEngComm 10 (2008) 1107;
[3]
[4]
(c) L. Brammer, G. M. Espallargas, S. Libri, CrystEngComm 10 (2008) 1712;
(d) R. Bertani, P. Sgarbossa, A. Venzo, F. Lelj, M. Amati, G. Resnati, T. Pilati, P. Metrangolo, G.
Terraneo, Coord. Chem. Rev. 254 (2010) 677.
See, for example: (a) A. Mele, P. Metrangolo, H. Neukirch, T. Pilati, G. Resnati, J. Am. Chem. Soc. 127
(2005) 14972;
(b) E. Dimitrijevic, P. Kvak, M.S. Taylor, Chem. Commun. 46 (2010) 9025;
(c) C.J. Serpell, N.L. Kilah, P.J. Costa, V. Félix, P.D. Beer, Angew. Chem., Int. Ed. 49 (2010) 5322;
(d) A. Caballero, N.G. White, P.D. Beer, Angew. Chem., Int. Ed. 50 (2011) 1845;
(e) A.V. Jentzsch, D. Emery, J. Mareda, P. Metrangolo, G. Resnati, S. Matile, Angew. Chem. Int. Ed., 50
(2011) 11675;
(f) A. Caballero, F. Zapata, N.G. White, P.J. Costa, V. Félix, P.D. Beer, Angew. Chem. Int. Ed. 51
(2012) 1876;
9
Page 10 of 13

(g) M. Cametti, K. Raatikainen, P. Metrangolo, T. Pilati, G. Terraneo, G. Resnati, Org. Biomol. Chem.
10 (2012) 1329;
(h) M. Erdélyi, Chem. Soc. Rev. 41 (2012) 3547;
(i) S.M. Walter, F. Kniep, L. Rout, F.P. Schmidtchen, E. Herdtweck S.M. Huber, J. Am. Chem. Soc. 134
(2012) 8507;
(j) T.M. Beale, M.G. Chudzinski, M.G. Sarwar, M.S. Taylor, Chem. Soc. Rev. 42 (2013) 1667.
(k) M.G. Sarwar, B. Dragisic, E. Dimitrijevic, M.S. Taylor, Chem. Eur. J. 19 (2013) 2050.
[5]
(a) G. Cavallo, P. Metrangolo, T. Pilati, G. Resnati, M. Sansotera, G. Terraneo, Chem. Soc. Rev. 39
(2010) 3772;
(b) M.G. Sarwar, B. Dragisic, S. Sagoo, M.S. Taylor, Angew. Chem. Int. Ed. 49 (2010) 1674;
(c) M.G. Chudzinski, C.A. McClary, M.S. Taylor, J. Am. Chem. Soc. 133 (2011) 10559;
For the use of non-fluorinated neutral halogen-bond donors, see for example:
(d) H.M. Yamamoto, J.-I. Yamaura, R. Kato, J. Am. Chem. Soc. 120 (1998) 5905;
(e) H.M. Yamamoto, Y. Kosaka, R. Maeda, J.-I. Yamaura, A. Nakao, T. Nakamura, R. Kato, ACS Nano
2 (2008) 143;
(f) H. Wang, X. R. Zhao, W. J. Jin, Phys. Chem. Chem. Phys. 2013, DOI: 10.1039/c3cp43865a.
[6]
[7]
(a) H.L. Nguyen, P.N. Horton, M.B. Hursthouse, A.C. Legon, D.W. Bruce, J. Am. Chem. Soc. 126
(2004) 16;
(b) D.W. Bruce, P. Metrangolo, F. Meyer, T. Pilati, C. Praesang, G. Resnati, G. Terraneo, S.G.
Wainwright, A.C. Whitwood, Chem. Eur. J. 16 (2010) 9511.
See, for example: (a) N. Kuhn, T. Kratz, G.J. Henkel, Chem. Soc., Chem. Commun. (1993) 1778;
(b) R. Weiss, M. Rechinger, F. Hampel, Angew. Chem. Int. Ed. 33 (1994) 893;
(c) A.J. Arduengo, III, M. Tamm; C.J. Calabrese, J. Am. Chem. Soc. 116 (1994) 3625;
(d) N.L. Kilah, M.D. Wise, C.J. Serpell, A.L. Thompson, N.G. White, K.E. Christensen, P.D. Beer, J.
Am. Chem. Soc. 132 (2010) 11893 and references cited therein.
(e) S. Derossi, L. Brammer, C.A. Hunter, M.D. Ward, Inorg. Chem. 48 (2009) 1666.
(f) K. Raatikainen, M. Cametti, K. Rissanen, Beilst. J. Org. Chem. 6 (2010), no. 4.
N.L. Kilah, M.D. Wise, P.D. Beer, Cryst. Growth Des. 11 (2011) 4565.
[8]
[9]
(a) S.M. Walter, F. Kniep, E. Herdtweck, S.M. Huber, Angew. Chem. Int. Ed. 50 (2011) 7187;
(b) F. Kniep, S.M. Walter, E. Herdtweck, S.M. Huber, Chem.–Eur. J. 18 (2012) 1306;
(c) F. Kniep, L. Rout, S.M. Walter, H.K.V. Bensch, S.H. Jungbauer, E. Herdtweck, S.M. Huber, Chem.
Commun. 48 (2012) 9299.
[10]
[11]
S. Adimurthy, G. Ramachandraiah, P.K. Ghosh, A.V. Bedekar, Tetrahedron Lett. 44 (2003) 5099.
(a) R. Weiss, J. Seubert, Angew. Chem. Int. Ed. 33 (1994) 891;
(b) J. Seubert, Dissertation, University of Erlangen–Nuremberg, Germany (1995).
[12]
Crystal data for 4a: formula: C12F6I4N2; Mr=793.74; crystal color and shape: red needle, crystal
dimensions: 0.100.150.56 mm; crystal system: monoclinic; space group: C2/c (no. 15); a=20.7401(5),
b=12.7959(3), c=13.3328(3) Å, β=97.1338(11)°; V=3510.98(14) Å3 ; Z=8; (MoK)=7.156 mm-1;

calcd=3.003 g cm-3; range=1.87–25.47°; data collected: 39095; independent data [Io>2(Io)/all
data/Rint]: 3070/3248/0.030; data/restraints/parameter: 3248/0/218; R1 [Io>2(Io)/all data]:
0.0130/0.0148; wR2 [Io>2(Io)/all data]: 0.0321/0.0329; GOF=1.079; max/min: 0.46/-0.50 eÅ-3. For
detailed information see Supporting information. CCDC 919254 (4a) contains the supplementary
crystallographic data for this compound.
10
Page 11 of 13

In addition, we have also obtained single crystals of compound 4b that were suitable for X-ray structural
analysis. Similar to the case of 4a, the crystal packing of 4b is based on a complex network of Br-Br
halogen-bonds, in which bromine substituents act both as halogen-bond donor and acceptor.
Crystal data for 4b: formula: C12Br4F6N2; Mr=605.74; crystal color and shape: orange fragment, crystal
dimensions: 0.130.230.30 mm; crystal system: orthorhombic; space group: Pna21 (no. 33);
a=23.0474(9), b=11.6967(4), c=5.4749(2) Å, V=1475.92(9) Å3 ; Z=4; (MoK)=10.972 mm-1;

calcd=2.726 g cm-3; range=1.77–25.36°; data collected: 44989; independent data [Io>2(Io)/all
data/Rint]: 2669/2705/0.057; data/restraints/parameter: 2705/1/219; R1 [Io>2(Io)/all data]:
0.0174/0.0179; wR2 [Io>2(Io)/all data]: 0.0445/0.0452; GOF=1.110; max/min: 0.50/-0.36 eÅ-3. For
detailed information see Supporting information. CCDC XXX {will be inserted prior to publication}
(4b) contains the supplementary crystallographic data for this compound.
The data of 4a and 4b can be obtained free of charge from The Cambridge Crystallographic Data Centre
via www.ccdc.cam.ac.uk/data_request/cif.
[13]
The N=N bond lengths of the azo bridges are comparable to that of trans-azobenzene with 1.25-1.26 Å
(J. Harada, K. Ogawa, J. Am. Chem. Soc. 126 (2004) 3539) and of trans-perfluoroazobenzene with 1.23
Å (K. Chinnakali, H.-K. Fun, O.B. Shawkataly, S.-G. Teoh, Acta Cryst. C: Cryst. Struct. Commun. 49
(1993) 615).
[14]
[15]
A. Bondi, J. Phys. Chem. 68 (1964) 441.
C.M. Reddy, M.T. Kirchner, R.C. Gundakaram, K.A. Padmanabhan, G.R. Desiraju, Chem. Eur. J. 12
(2006) 2222.
[16]
[17]
A. Olejniczak, A. Katrusiak, A. Vij, CrystEngComm 11 (2009) 1073.
Due to the strongly oxidizing nature of compound 1, an equivalent of cyclohexene had to be added to the
reaction in order to quench the elemental bromine that was formed by oxidation of bromide. We have
seen no similar reactivity for compounds 4a and 4b.
[18]
There is only one previous report in which a polyfluorinated halogen-bond donor has been postulated to
be active in organic synthesis / organocatalysis, see: A. Bruckmann, C. Bolm, Synlett (2008) 900.
[19]
[20]
Y. Zhao, D. G. Truhlar, Theor. Chem. Acc. 120 (2008) 215.
(a) F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 7 (2005) 3297;
(b) K.A. Peterson, D. Figgen, E. Goll, H. Stoll, M. Dolg, J. Chem. Phys. 119 (2003) 11113.
[21]
Gaussian 09, Revision B.01, M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R.
Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li,
H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R.
Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J.A.
Montgomery, Jr., J.E. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N.
Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J.
Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J.
Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski,
R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich,
A.D. Daniels, Ö. Farkas, J.B. Foresman, J.V. Ortiz, J. Cioslowski, and D. J. Fox, Gaussian, Inc.,
Wallingford CT, (2010).
[22]
[23]
J. Tomasi, B. Mennucci, R. Cammi, Chem. Rev. 105 (2005) 2999.
Bidentate coordination to bromide was chosen in both cases for the sake of comparison. In solution,
monodentate binding of bromide to the halogen-bond donors might also occur, but the trends found for
the bidentate binding should also apply to the monodentate complexes.
[24]
[25]
CYLview, 1.0b; C.Y. Legault, Université de Sherbrooke (2009) (http://www.cylview.org).
A common visualization of the electronic origin of halogen bonding is the region of positive electrostatic
potential on the surface of the respective halogen substituent, the so-called “-hole” (see T. Clark, M.
Hennemann, J.S. Murray, P. Politzer, J. Mol. Mod. 13 (2007) 291). We have computed the maximum
11
value of positive electrostatic potential VS,max for both halogen-bond donors 1 and 4a on an isodensity
Page 12 of 13

surface of 0.001 electron/Bohr-3. The corresponding values (VS,max(1) = 139 kcal/mol, VS,max(4a) = 32
kcal/mol) further support the finding that compound 1 is a markedly more potent halogen bond donor.
We note, however, that caution should be exercised in the interpretation of VS,max values (compare
reference 2i).
[26]
We note that the complexation Gibbs free energies are estimated to be +4 kcal/mol for the adduct with
4a and -1 kcal/mol for the complex with 1, and thus feature the same trend. However, the calculation of
exact Gibbs free energies is likely to be even more challenging than for the energies itself.
12
Page 13 of 13