To Anion–π or not to Anion–π: The Case of Anion-Binding to Divalent Fluorinated Pyridines in the Gas Phase

Article abstract of DOI:10.1002/chem.201800893

A series of mono- and divalent fluorinated pyridine derivatives is investigated by electrospray ionization (tandem) mass spectrometry and quantum chemical calculations with respect to their capability to bind anions in the gas phase. The pyridine derivatives differ not only in valency, but also with regard to the degree of fluorination of the pyridine rings, the positions of the fluorine atoms, the rigidity of the spacers connecting the two pyridines in the divalent compounds, and the relative configuration. While the monovalent compounds did not form anion complexes, the divalent analogues exhibit anion binding even to weakly coordinating anions such as tetrafluoroborate. Three different tandem mass spectrometric experiments were applied to rank the gas-phase binding energies: (i) collision-induced dissociation (CID) experiments in a Fourier transform ion-cyclotron-resonance (FTICR) mass spectrometer on two different, simultaneously mass-selected complexes with different receptors, (ii) determination of the collision energy required to fragment 50 % of the mass-selected complexes in an ESI-QToF mass spectrometer, and (iii) CID of heterodimers formed from two different, competing pyridine receptors and indigo carmine, a dianion with two identical binding sites. All three experiments result in consistent binding energy ranking. This ranking reveals surprising features, which are not in agreement with binding through anion–π interactions. Density functional theory (DFT) calculations comparing different potential binding modes provide evidence that the ranking can instead nicely be explained, when C?H???anion interactions with the spacers are invoked. These results are supported by gas-phase IR spectroscopy and ion mobility-mass spectrometry (IM-MS) on a selected set of chloride pyridine complexes.

Full text of DOI:10.1002/chem.201800893

A Journal of
Accepted Article
Title: To Anion-pi or not to Anion-pi: The Case of Anion-Binding to
Divalent Fluorinated Pyridines in the Gas Phase
Authors: Melanie Göth, Felix Witte, Marcel Quennet, Phillip Jungk,
Gabriel Podolan, Dieter Lentz, Waldemar Hoffmann, Kevin
Pagel, Hans-Ulrich Reissig, Beate Paulus, and Christoph A
Schalley
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201800893
Link to VoR: http://dx.doi.org/10.1002/chem.201800893
Supported by

10.1002/chem.201800893
Chemistry - A European Journal
FULL PAPER
To Anion- or not to Anion-: The Case of Anion-Binding to
Divalent Fluorinated Pyridines in the Gas Phase
Melanie Göth,[a] Felix Witte,[a] Marcel Quennet,[a] Phillip Jungk,[a] Gabriel Podolan,[a] Dieter Lentz,[b]
Waldemar Hoffmann,[a,c] Kevin Pagel,[a,c] Hans-Ulrich Reissig,*[a] Beate Paulus,*[a] and Christoph A.
Schalley*[a]
Abstract: A series of mono- and divalent fluorinated pyridine
derivatives is investigated by electrospray ionization (tandem) mass
spectrometry and quantum chemical calculations with respect to
their capability to bind anions in the gas phase. The pyridine
derivatives differ not only in valency, but also with regard to the
degree of fluorination of the pyridine rings, the positions of the
fluorine atoms, the rigidity of the spacers connecting the two
pyridines in the divalent compounds, and the relative configuration.
While the monovalent compounds did not form anion complexes, the
divalent analogues exhibit anion binding even to weakly coordinating
anions such as tetrafluoroborate. Three different tandem mass
spectrometric experiments were applied to rank the gas-phase
binding energies: (i) collision-induced dissociation (CID) experiments
binding motifs were experimentally observed: (i) the C-H···X-
hydrogen bond, (ii) the mostly covalent Meisenheimer complex,
(iii) the weakly covalent anion-donor--acceptor interaction and
5
]
(iv) the non-covalent anion- interaction.[
The anion-
interaction can be described as an attractive force between the
anion and the positive -acidic surface of an electron-deficient
aromatic ring with permanent quadrupole moment. The anion is
6
]
then located above the center of the aromatic ring.[
Hexafluorobenzene and triazine, for example, form complexes
with anions – in contrast to benzene, which has a quadrupole
moment with negative  face. Gas-phase IR spectroscopy
revealed that subtle changes in the structure of the aromatic
molecule can cause a complete shift from one binding mode to
another one.[7] While hexafluorobenzene binds a chloride ion
above the aromatic ring, pentafluorobenzene realizes a C-H···Cl-
interaction as the -acidity of the aromatic ring is reduced and
the C-H bond is strongly polarized.
Although the origins of research on anion-interactions date
back to the 1980ies,[8] theoretical studies by Deyà et al.[9] and
Mascal et al.[10] published in 2002 attracted the attention of a
broader supramolecular community, since the anion-
interaction was predicted to be significantly attractive (~20-
50 kJ/mol) in the gas phase. As anions are involved in many
chemical and biochemical processes, research in this field is
motivated by the desire for new and specific anion
receptors[11]and transport systems.[12] In the last years, a huge
number of publications appeared on a broad variety of aspects
related to anion- interactions.[4g,13] In many of these complexes,
other comparatively strong attractive forces such as hydrogen
bonding or stronger electrostatic interactions are the primary
binding force. The anion- interaction itself is relatively weak in
solution and thus usually observed as a modulation of these
major effects. It is thus difficult to differentiate, whether the
anion- interaction is attractive or repulsive in these complexes.
Consequently, experimental evidence for “pure” anion-
interactions remains hard to obtain.[14]
in
a Fourier-transform ion-cyclotron-resonance (FTICR) mass
spectrometer on two different, simultaneously mass-selected
complexes with different receptors, (ii) determination of the collision
energy required to fragment 50 % of the mass-selected complexes
in an ESI-QToF mass spectrometer, and (iii) CID of heterodimers
formed from two different, competing pyridine receptors and indigo
carmine,
a dianion with two identical binding sites. All three
experiments result in consistent binding energy ranking. This ranking
reveals surprising features, which are not in agreement with binding
through anion- interactions. Density functional theory (DFT)
calculations comparing different potential binding modes provide
evidence that the ranking can instead nicely be explained, when C-
H···anion interactions with the spacers are invoked. These results
are supported by gas-phase IR spectroscopy and ion mobility-mass
spectrometry (IM-MS) on
complexes.
a selected set of chloride pyridine
Introduction
Besides well-known and intensely studied non-covalent forces
such as hydrogen bonding,[ 1 ] - stacking,[ 2 ] and cation-
interactions,[3] the non-covalent bonds between anions and 
systems have attracted broad interest recently.[4] Four different
Tandem mass spectrometry in concert with theory can help
circumventing this difficulty, because it offers the advantage that
non-solvated anion-receptor complexes can be examined. In the
gas phase, the anion- interactions do not compete with
solvation or counterion effects. Consequently, no additional
primary interaction is required and the observation of an anion-
host complex directly implies an attractive force between both.
This advantage of the gas phase, however, comes at a price, as
the major challenge is to determine the binding mode of these
complexes. It is thus pivotal to combine the experiments with
high-level calculations.
[a]
M. Göth, F. Witte, M. Quennet, Dr. P. Jungk, Dr. G. Podolan, Prof.
Dr. H.-U. Reissig, Prof. Dr. B. Paulus, Prof. Dr. C. A. Schalley
Institut für Chemie und Biochemie, Freie Universität Berlin
Takustraße 3, 14195 Berlin, Germany
E-Mail: hreissig@chemie.fu-berlin.de; b.paulus@fu-berlin.de;
c.schalley@fu-berlin.de
[b]
[c]
Prof. Dr. D. Lentz
Institut für Chemie und Biochemie, Freie Universität Berlin
Fabeckstraße 34-36, 14195 Berlin, Germany
W. Hoffmann, Prof. Dr. K. Pagel
Here, we present a combined mass spectrometric and
quantum chemical study of non-covalent complexes of the
fluorinated mono- and divalent aminopyridines and one divalent
alkoxypyridine depicted in Figure 1 with different anions.
Fritz-Haber-Institut der Max-Planck-Gesellschaft
Faradayweg 4-6, 14195 Berlin, Germany
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800893
Chemistry - A European Journal
FULL PAPER
M8, D4 and D5) provide insight into the role of sandwich-like
complexes and the effect of the nitrogen or oxygen substituents.
(ii) Comparing divalent pyridines with flexible ethylene spacers
with those that bear a rigid cyclohexane spacer (e.g. D4 and
DC4) reveal whether conformational fixation plays a role. (iii)
Diastereomers such as DC4 and DC5 refine these arguments
that are based on the idea that preorganization of the two
pyridine rings in the divalent receptors might play an important
role. (iv) Different degrees of fluorination (e.g. D1, D3 and D4)
show how important fluorination is. (v) Isomers, which differ only
with respect to the positions of the fluorine substituents (e.g. D1
and D2), may shed light on the binding mode.
Results and Discussion
Conceptual Considerations. As briefly discussed above,
anion- interactions are usually weak in solution due to the
competing effects of counterions and solvent. A study of the
anion gas-phase binding thus has its advantages, but suffers
from the lack of direct structural information. As the
determination of exact absolute binding energies in the gas
phase is not a trivial task and requires specialized equipment,[15]
we aim at establishing a ranking of relative binding forces of
anions to the different pyridine derivatives discussed above. On
one hand, the determination of relative binding strengths for
example by experiments, in which pairwise competitions are
monitored, is particularly suited, when the binding energy
differences are small. On the other hand, a comparison between
the ranking expected for anion- interactions and that obtained
by experiment may well provide qualitative insight into the
binding mode, which can then be complemented by theory.
In order to cross-check the binding ranking obtained, three
different experiments were performed: Method A: Two
complexes with the same anion, but different receptors were
ionized, brought into the cell of
a Fourier-transform ion-
cyclotron-resonance mass spectrometer (FTICR), where both
complexes are simultaneously mass-selected and fragmented at
increasing collision energies. The more weakly bound complex
will start to fragment at lower collision energies and the intensity
ratio of both signals shift towards the more strongly bound
complex. As the mass differences between the two complexes
are small, we do not expect mass discrimination to play a
significant role. Method B: Each complex was furthermore
studied by collision-induced dissociation and the collision energy,
at which 50 % of the parent ion had dissociated, was determined
and compared for the different complexes (survival yield
method).[16] Method C: Finally, a direct competition was achieved
by using indigo carmine (3,3’-dioxo-2,2’-bisindolyden-5,5’-
disulfonic acid disodium salt), a dianion with a rigid spacer
separating the two sulfonate groups. Mixing two different
receptors with this dianion generated the homo- and
heterodimers of two receptor molecules bound to the dianion
simultaneously. Mass-selection of the heterodimer followed by
CID also provides insight, which of the receptors binds more
strongly. This method is a qualitative variant of Cooks’ kinetic
method.[17] Note that the sodium salt is used here instead of the
Figure 1. Fluorinated mono- and divalent pyridines with different numbers and
positions of fluoro substituents, flexible ethylene (D1-D5) or rigid cyclohexane
spacers (DC1-DC6) and different relative configuration (DC4 vs. DC5).
Pyridines were used as they are already -electron-deficient
aromatic compounds, which are further made -acidic by the
fluorine substituents. Theoretical studies have shown that
pentafluoropyridine binds anions even stronger than
hexafluorobenzene.[6] The pyridines used here differ with respect
to several features so that comparison of different series will
provide more detailed insight into binding trends: (i) Mono- and
divalent compounds with the same degree of fluorination (e.g.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800893
Chemistry - A European Journal
FULL PAPER
prepared from pentafluoropyridine
and ethylene glycol under basic
conditions (equation 4).
Crystallization of the divalent
4-aminopyridine derivatives D4
and DC4 with a flexible and a
rigid spacer, respectively, from
benzene/hexane (1:1) (D4) and
directly after column chromate-
graphy (DC4) gave crystals
suitable for single crystal X-ray
structure determinations. Both
compounds crystallize in the
monoclinic space group C2/c,
with half a molecule forming the
asymmetric unit and both mole-
cules possess crystallographic C2
symmetry. The molecular struc-
ture and the atom-numbering
scheme are depicted in Figure 3.
Due to a less rigid structure,
the fluorinated pyridine moieties
of D4 are in closer contact with
the shortest distance of 3.13 Å
found between C4 and its
symmetry related atom C4i. In
Figure 2. Syntheses of 4-aminopyridines M8 and M4 as well as divalent compounds DC4, DC1, DC2, and D5.
contrast, the shortest carbon-
tetrabutylammonium salt employed in the other experiments. As
the counterion is absent in any of the gas-phase complexes that
are mass-selected and studied, this should not have any effect
on the gas-phase stability ranking.
It is much more conclusive to compare the experimental
binding sequences with trends calculated for the different
binding modes with high-level quantum theoretical methods than
to try to theoretically match individual experimental binding
energies. Also, this argument supports that the determination of
the ranking is a reasonable approach.
carbon distance (C4···C4i) between the aromatic ring systems of
DC4 amounts to 3.94 Å. However, a short fluorine contact exists
between F3 and C4i (3.26 Å). In D4 and DC4, the amino
nitrogen atom adopts a gauche conformation with dihedral
angles of 69.0° and -55.2°, respectively.
Synthesis of fluorinated 4-aminopyridine derivatives and
crystal structures of D4 and DC4. The experimentally studied
4-aminopyridine receptors were prepared by nucleophilic
substitution at the 4-position of pentafluoropyridine or
2,4,6-trifluoropyridine followed by regioselective defluorination or
N-methylation.[18] This approach is exemplified by the reactions
leading to compounds M4, M8, DC1, DC2 and D4 (Figure 2).
Pentafluoropyridine and dimethylamine react to provide
exclusively the 4-amino-substituted pyridine derivative M8 that
was
exposed
to
a
novel
titanocenium-catalyzed
hydrodefluorination reaction.[18a] Target compound M4 was
regioselectively obtained in overall good yield (equation 1).
The divalent compounds DC4 and DC1 were prepared with
reasonable efficiency following the same route (equation 2). For
divalent 4-aminopyridine derivative DC2 it was advantageous to
introduce the N-methyl groups after the nucleophilic substitution
in order to achieve at least moderate regioselectivity in the
nucleophilic substitution reaction (equation 3).[18] The divalent 4-
alkoxy-substituted pyridine derivative D5 was analogously
Figure 3. Molecular structures (ORTEP[19]) of a) D4 and b) DC4. Ellipsoids are
drawn at a 50 % probability level.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800893
Chemistry - A European Journal
FULL PAPER
which two receptor/chloride complexes are compared pairwise in
a CID experiment (method A). Gradually increasing the collision
energy leads to the preferential dissociation of the more weakly
bound complex. For pairs of receptors that exhibit large
differences in binding energies, this experiment is difficult to do
or even impossible, because their chloride complexes appear in
the mass spectrum with substantially different intensities. In
extreme cases, one is even absent due to the strong competition.
Nevertheless, pairs of receptors with similar binding energies
can be compared easily and a binding energy ranking is
obtained by proceeding step-wise from one pair to the next. It
should also be mentioned that the DC4/DC5 pair could not
directly be compared because both receptors differ only in their
relative configuration; both have the same elemental
composition and thus form complexes with the same m/z value.
Here, the comparison of both receptors with others provides
insights into their relative binding strengths. Using as many pairs
as possible demonstrates the binding strength ranking to be
consistent (Figure S1, Table S2). From these measurements,
we obtain the following ranking:
Table 1. Detected complexes of anions with divalent pyridines.
-
-
-
-
Cl-
Br-
I-
NO3
BF4
PF6
H3CSO3
D1
–
+
+
+
+
–
+
+
+
+
–
–
+
+
+
+
–
+
+
+
+
–
–
+
–
+
+
–
+
+
+
–
–
–
+
–
+
+
–
+
+
+
+
–
–
+
–
+
+
–
+
–
+
–
–
–
+
–
–
–
–
–
–
–
–
–
–
+
–
+
–
–
+
+
+
+
D2
D3
D4
D5
DC1
DC2
DC3
DC4
DC5
DC6
DC2 > DC4 > D2 > D4 ≈ DC3 > D5 > DC5 > D3
In both compounds, the amino nitrogen atoms exhibit an
almost trigonal planar substitution by the adjacent carbon atoms
with a sum of the bond angles of 358.0° (D4) and 359.4° (DC4).
As expected, the C-N distances to the aromatic ring are about
0.1 Å shorter than those to the sp3 carbon atoms of the
substituents.
Generation of gaseous anion-receptor complexes by ESI
mass spectrometry. To generate anion complexes, all pyridine
derivatives were incubated with the corresponding tetrabutyl-
ammonium salt of the anion and electrosprayed from a dichloro-
methane : acetonitrile (4:1) solution. The three monovalent 4-
aminopyridine derivatives M4, M5 and M8 (Figure 1) had been
synthesized to be examined as first test systems. They feature
different degrees of fluorination as well as different positions of
fluorine substituents and were regarded therefore as meaningful
examples. As all attempts failed to ionize intact complexes of the
halides Cl-, Br-, I- and of NO3- or of weakly coordinating anions
-
-
BF4 and PF6 with these aminopyridines, the remaining
monovalent fluorinated aminopyridines where thus not
synthesized, but studied only theoretically. Among the divalent
receptors D1, DC1 and DC6 did not yield any detectable
complexes with the anions tested (Table 1). In marked contrast,
electrospray ionization of the divalent pyridine analogues D2-D5
and DC2-DC5 provided access to quite abundant complex ions.
-
Some of the receptors even bind to the weakly coordinating BF4
-
anion. Even PF6 is bound by one of the divalent receptors, i.e.
Figure 4. Method A. CID mass spectra for two simultaneously isolated
chloride complexes. a) DC2 vs. DC4: Different degrees of fluorine substitution.
b) D4 vs. DC4: Flexible and rigid spacers. Insets: Plot of the intensity depletion
of the weaker chloride complex relative to the stronger complex over
increasing collision energies (CE).
D2. This result is particularly intriguing as D2 bears only doubly
fluorinated pyridine rings. It is certainly counterintuitive for anion-
-interactions that this pyridine derivative forms complexes,
while the structurally equivalent, but more highly fluorinated
receptors D3 and D4 do not.
Gas-phase ranking of anion binding to divalent fluorinated
Clearly, the pairs DC2/DC4 (Figure 4a) and D2/D4 (Figure
pyridines. Figure 4 shows two representative experiments, in
S1a) reveal the derivatives doubly fluorinated in the 2,6-
This article is protected by copyright. All rights reserved.

10.1002/chem.201800893
Chemistry - A European Journal
FULL PAPER
positions of the pyridine rings to bind more strongly to chloride
than their perfluoro analogues. Also, the more rigid spacers in
DC2 and DC4 promote the binding strength, while the flexible
ethylene spacer in D2 and D4 reduces it (Figure 4b and Figure
S1b). The rather weak binding of DC3, which has the lowest
conformational flexibility due to the methylene bridge, evidences
free chloride is determined for each step in order to identify at
which collision energy 50 % of the ions have fragmented (Figure
5b). This experiment is somewhat limited for very weakly bound
complexes, as the collision energy, at which 50 % of the
complexes have fragmented is then below the instrument’s
minimum (start) value of 2 V. However, the degree of
fragmentation at this collision energy can be compared even for
those complexes and thus, a ranking can be obtained. Table 2
reports the 50 % survival yield collision energies for all
complexes that bind the chloride sufficiently strong. The ranking
of binding strengths obtained here is consistent with that
obtained with method A:
a
clear dependence on structural details. In addition, a
pronounced stereochemical difference is observed, which puts
the trans-configured isomer DC4 among the strongest binders,
while the corresponding cis-isomer DC5 is among the weakest
ones. The series D2 > D4 > D3 also shows that the binding
strength does not only depend on the degree of fluorine
substitution. As D1 does not form any anion complexes and D3
is weakly interacting with chloride, fluorine substituents in the
3,5-positions of the pyridine ring appear not to be advantageous.
The sequence of D4 > D5 shows that substitution of the N-
methylamino units by oxygen atoms does not lead to a stronger
interaction with chloride.
DC2 > DC4 > D2 > DC3 > D4 (> D5 ~ DC5 ~ D3)
Table 2. Collision energy at 50 % complex dissociation for the chloride
complexes of different divalent pyridine derivatives. Values correspond to
the average survival yield of measurements (in triplicate). Data is well
reproducible (standard deviation < 1 %).
DC2
DC4
D2
5.36 V
3.19 V
3.08 V
2.21 V
< 2 V
DC3
D4
Finally, method C has been applied, in which two of the
divalent pyridines are attached simultaneously to indigo carmine
(IC), a dianion with two sulfonated groups (Figure 6a). The
dianion enables a direct comparison of relative binding strengths
of the pyridine receptors. The heterodimeric complex is mass-
selected and subsequently fragmented. The first pyridine to
leave the complex is the one which has the weaker affinity.
Figure 6 shows exemplarily two experiments with the pairs
D2/D4 and D4/DC4. At low collision voltages (5 V), the indigo
carmine complex with D2 and D4 barely dissociates (Figure 6b),
top). With increasing collision voltages, two main fragments
evolve, which correspond to the doubly charged indigo carmine
(m/z 210) and the doubly charged complex of D2 with indigo
carmine (m/z 367). As the D4 indigo carmine complex is not
detected (m/z 403), this implies that D4 dissociates first and thus
is clearly the weaker binding pyridine. A slightly different case is
the D4/DC4 pair (Figure 6c): Already at low collision voltages,
fragments form with high intensity. In addition to the doubly
charged bare indigo carmine ion (m/z 210), significant signals for
the D4 indigo complex (m/z 403) and the DC4 indigo carmine
complex (m/z 430) are detected. DC4 binds slightly stronger
than D4 indicated by the somewhat higher intensity of the
DC4/IC intermediate. This indicates that the more rigid receptor
DC4 has a higher binding affinity to indigo carmine than the
flexible D4. Performing this experiment with several pairs of
divalent pyridines (Figure S3, Table S3), the following ranking
evolves:
Figure 5. Method B. a) Mass spectrum of the m/z-selected complex [DC4·Cl]-
at 2 V collision energy (CE). b) Plot of the survival yield against the CE and fit
with a sigmoidal function. At a CE of 3.19 V, 50 % of the complex has
dissociated.
The ranking of chloride complex binding strengths obtained
by method A is confirmed by the application of the survival yield
approach (method B). Figure 5 displays exemplarily for the
complex DC4·Cl- a typical experiment, during which the collision
energy is increased in a stepwise manner by increasing the
transfer collision energy of the ions. The ratio of complex and
D2 ≥ DC2 > DC4 > D4 > D5
This article is protected by copyright. All rights reserved.

10.1002/chem.201800893
Chemistry - A European Journal
FULL PAPER
This experiment was not possible with all of the pyridines.
For instance, DC3 does not bind to the dianion and the pairs
D3/DC5 and D3/D5 form only complexes where indigo carmine
is bound to two receptors of the same type (homodimers). The
pair D5/DC5 forms a heterodimer, however, the interaction is too
weak so that it does not survive mass-selection in the
quadrupole (Table S3). For the complexes, which could be
ranked, the sequence is the same as those obtained with
methods A and B – with one notable exception: D2 exhibits a
similar binding strength as DC2, although its chloride complexes
are weaker than those of DC2. We attribute this to the fact that
indigocarmine differs in structure significantly from the small,
spherical chloride anion.
Density functional calculations of anion binding with
monovalent pyridines. Quantum chemical calculations were
carried out to provide insight into the binding energies and
binding modes of the pyridine-anion complexes. Therefore,
chloride complexes with the monovalent 4-aminopyridine
derivatives M1-M8 (Figure 1) were investigated at the TPSS-
D3(BJ)/def2-TZVP level to establish the theoretical methods with
a series of model compounds. Two starting points were selected,
where chloride is coordinated either in close proximity to the
hydrogen atoms of the N-methyl groups (Figure 7a), or above
the aromatic ring as required for an anion- interaction (Figure
7b). The calculations reveal that an anion- interaction does not
yield a stable minimum structure, whereas the favored binding
site is in every case the coordination through C-H···anion
hydrogen bonds as depicted in Figure 7a. The distances
between the chloride anion and the hydrogen atoms of the N-
methyl groups are in the range of 2.36-2.41 Å (see Supporting
Information, Table S5), which is smaller than the sum of the two
van der Waals radii of hydrogen (1.10 Å) and chloride (1.74
Å).[20]
Figure 7. Starting geometries of the monovalent 4-aminopyridine M8 with a
chloride anion attached a) through C-H···anion hydrogen bonds at the N-
methyl groups and b) through anion-interaction. The color code is as follows:
carbon, grey; hydrogen, white; nitrogen, blue; fluorine, orange; chlorine, green.
Figure 6. Method C. a) Direct comparison of the relative binding strengths with
indigo carmine (IC). b) The heterodimeric complex [D2·D4·IC]2- was subjected
to a CID experiment. The exclusive formation of [D2·IC]2- as primary fragment
clearly shows D2 to bind stronger than D4. c) The preferential formation of
[DC4·IC]2- over [D4·IC]2- from mass-selected [D4·DC4·IC]2- indicates a slightly
higher affinity of the more rigid receptor DC4.
The resulting free association enthalpies (ΔGa) are
summarized in Table 3. Comparing the values of the complexes
of M4 with M5 (- 53.4 vs. -70.1 kJ/mol) and other isomer pairs,
This article is protected by copyright. All rights reserved.

10.1002/chem.201800893
Chemistry - A European Journal
FULL PAPER
which differ only in the fluorination pattern, it becomes apparent
that fluorine substituents in position 2 and 6 of the pyridine ring
lead to stronger binding, whereas they weaken the interaction
when located in positions 3 and 5. This can be explained by
taking a closer look at the electron densities and partial charges
of the aminopyridines: 2,6-fluorine substituents lead to more
polarized C-H bonds of the two coordinating N-methyl groups
and to a lower electron density at the corresponding hydrogen
atoms, which strengthens the hydrogen-chloride interaction. In
contrast, fluorine atoms in 3,5-position weaken the interaction.
The electron densities intuitively correspond to the +M and -I
effects of the fluorine substituents. The mono-fluorinated
pyridine derivatives have similar partial charges, with only
slightly deviating association enthalpies (~5 kJ/mol). Comparing
the values of the non-fluorinated pyridine derivative M1
(-56.1 kJ/mol) with the perfluorinated M8 (- 67.2 kJ/mol), M8
leads to more positively polarized hydrogen atoms and therefore
to a stronger interaction with chloride.
close proximity to hydrogen atoms and a stable minimum for a
pure anion- interaction has never been obtained.
Table 3 lists the resulting free association enthalpies (ΔGa)
of chloride complexes with the divalent pyridine derivatives.
Stable minima were obtained either when chloride is located on
top of the spacer (a) or between the pyridine rings in close
proximity to the spacer (b). ΔGa values range from -68.1 to
-132.6 kJ/mol, depending on a variety of parameters that will be
examined in detail below. The corrected free enthalpy
contribution in the form of ΔGRRHO is nearly the same for all
complexes (ranging from ca. 30 to 40 kJ/mol). Therefore, it does
not matter, whether the electronic energy or ΔGa is used for
discussion. No minima were obtained, where chloride was
exclusively in proximity to one or two pyridine rings. In all cases,
at least one C-H bond was nearby. Divalent complexes with
chloride being on top of the linker always yielded stronger
associations than their counterparts, where the anion is initially
placed between the pyridine rings. The energetic differences
between these two binding motifs range vastly from 10 to
30 kJ/mol and may be attributed to the lower number of possible
hydrogen bonds between the rings. Generally speaking,
numerous and strong C-H···anion hydrogen bonds rather than
anion- interactions are responsible for the binding of anions to
divalent pyridines. From Table 3, the following ranking evolves,
with strongest to weakest association from left to right:
Table 3. Free association enthalpies (ΔGa) of the chloride complexes with
monovalent and divalent fluorinated 4-aminopyridine derivatives computed
at the TPSS-D3(BF)/def2-TZVP level at T = 298.15 K.
ΔGa [kJ/mol]
-56.1
ΔGa [kJ/mol][a]
-81.5
ΔGa [kJ/mol][b]
M1
M2
M3
M4
M5
M6
M7
M8
D1
-
-58.5
D2
-118.2[c]
-94.0
-
DC2 > D2 > DC4 > DC3 ~ D4 > D5 ~ DC5 ~ D3 > DC1 > D1
-63.3
D3
-68.1
-80.5
-86.0
-
Similar arguments as for the monovalent pyridines apply for
the divalent receptors. A comparison of D1 vs D2, or DC1 vs
DC2, further underlines the assumption that 2,6-fluorine
substitution increases binding energies to chloride, whereas
fluorine atoms in 3,5-position weaken these non-covalent
interactions. Moreover, perfluoro-substituted compounds do not
show the highest binding energies, which fully supports our
experimental data. When taking a closer look at the overall trend
within the row D1 to D4, electrostatic repulsion and +M effect of
the fluorine atoms are in competition with their –I effect. Thus, it
can be concluded that the fluorination pattern is a more
important factor than the degree of fluorine substitution, as was
already the case for the monovalent complexes.
-53.4
D4
-107.9
-96.2[c]
-90.8
-70.1
D5
-60.5
DC1
DC2
DC3
DC4
DC5
-72.3
-132.6
-108.0[c]
-112.7
-95.8[c]
[b]
-
-67.2
-
-67.7
-69.0
[a]
Chloride located on top of the spacer.
Chloride located between
The calculations also offer an explanation, why DC2 and D2
[c]
pyridines in close proximity to the spacer. The receptor structure in the
complex closely resembles the most stable structure of the free divalent
pyridine.
bind chloride particularly strong
– as observed in the
experiments as well. As the 3- and 5-positions of the pyridine
are not fluorinated, the C-H bonds at these positions participate
in chloride binding with an additional C-H···Cl- interaction, which
cannot occur, when these positions are fluorinated. This C-
Density functional calculations of anion binding with
divalent pyridines. TPSS-D3(BJ)/def2-TZVP density functional
calculations were then performed for chloride complexes with
divalent pyridine derivatives. For these receptors, several
binding sites are possible and therefore different starting
structures were selected. This is exemplarily shown in Figure 8
for compound DC2, in which the two pyridine units are
connected by a rigid cyclohexane spacer. The chloride anion
can interact with the spacer either through the hydrogen atoms
of the N-methyl groups (Figure 8a) and/or through those of the
cyclohexane moiety (Figure 8b). As in the monovalent cases,
structure optimizations show that chloride is always located in
H···chloride interaction exhibits
a
H···Cl- distance that is
somewhat longer than the C-H···Cl- contacts to aliphatic
moieties (2.7 Å vs. 2.5 Å). Nevertheless, this additional contact
contributes an additional attractive force to the anion. As these
theoretical findings are in excellent agreement with the
experimental trends, we conclude that indeed the C-H···anion
contacts rather than anion- interactions are the pivotal non-
covalent forces that mediate anion binding to the divalent
pyridines.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800893
Chemistry - A European Journal
FULL PAPER
Figure 8. Optimized geometries of divalent aminopyridine derivative DC2 with a chloride anion attached. Three binding sites could be identified: a) Chloride
located between the N-methyl groups, b) chloride located between the two pyridine rings, and c) chloride located above the cyclohexyl spacer.
Comparing D1 vs. DC1, or D2 vs. DC2, indicates that a
flexible spacer between the two pyridine rings leads to weaker
interactions than a rigid spacer. This again strongly supports the
results from mass spectrometry. However, the hydrogen
bonding patterns with flexible or rigid spacer do not significantly
differ from one another. In both cases, there are the same
amount of linear and bent hydrogen bridges. Moreover, the N-
methyl groups in the rigid spacer case are slightly more
polarized due to the higher charge stabilizing effect of the longer
alkyl chain. This can be deduced from the respective atomic
populations (see supporting information). It could be argued that
the loss of binding energy for DC3 compared to DC2
(~ 25 kJ/mol) originates from its gain in rigidity. However,
simultaneously we see that DC3 provides one less hydrogen
bonding site than DC2, which is probably the stronger argument
and better in line with the other results. The molecular structure
of DC5 makes it geometrically impossible to display the same
amount of C-H···Cl- interactions as DC4. Hence, we see a major
difference in binding energy (~ 17 kJ/mol) between the two, even
though they are just stereoisomers.
unambiguously identified as the best binding 4-aminopyridine
derivative and the compounds, for that no binding could not be
detected in the experiment, i.e. DC1 and D1, are the same that
occur at the end of the experimental ranking with binding
energies only slightly higher than those of the monovalent
compounds. The preferred coordination site of the chloride anion
is never directly above the pyridine ring, but always near to the
N-methyl groups. In D5 – a compound without N-methyl groups
– the chloride anion only binds near the ethylene spacer.
Furthermore, electrostatic potential surfaces (ESP, Figure S6)
indicate that despite being highly fluorinated there is still a
significant amount of electron density in the area of the pyridine
moieties that is very likely due to the strong electron-donating
effect of the 4-amino substituents. While the potentials become
significantly positive with values up to 200 kJ/mol (isovalue:
0.01) at the methyl groups, the aryl units on average display
potentials far below 100 kJ/mol and can even become negative
in some monovalent cases. All of these results exclude anion-
interactions and strongly imply C-H···anion hydrogen bonding as
the main binding motif.
The most stable structures of free divalent pyridines usually
do not represent the conformations that show the strongest
associations to chloride (Figure 8c). This has a crucial effect on
the ranking, especially in the case of DC4. Arranging the
molecule in such a way that the two N-methyl groups point into
the same direction causes the binding energy to gain around
15 kJ/mol as compared to the assumed most stable
conformation shown by the crystal structure of the free molecule
(Figure 3 and Figure S5). For DC1 and DC2, this also has an
effect of around 9 kJ/mol, but does not change their position in
the ranking. For DC5, rearranging the N-methyl groups does not
lead to a stronger binding as it does not significantly change the
nature nor the number of hydrogen bonds and gives rise to an
unfavorable interaction of the cyclohexane ring with the
perfluorinated pyridine moiety.
Iodide complexes with receptors D3 and DC5 could not be
detected in the mass spectrometric experiments. With D5 a
complex could be detected, but it was identified as one of the
weakest binders. It was therefore worthwhile to compute these
structures as well. A comparison of the resulting ∆Ga values of
D3 (-66.5 kJ/mol), DC5 (-70.0 kJ/mol), and D5 (-71 kJ/mol)
shows a considerable weakening of association by more than
25 kJ/mol compared to the corresponding chloride complexes.
The iodide complex with receptor D5 could be detected in the
experiment, but was identified as one of the weakest binders.
Interestingly, the enthalpies of chloride complexes with the
monovalent receptors are also around -70 kJ/mol.
Gas-phase spectroscopy. To support the results from theory
and mass spectrometry, we further recorded size- and mass-
selected gas-phase infrared (IR) spectra for a selected set of
chloride pyridine complexes (DC2, DC4, D2, and D4) using a
home-built ion mobility-mass spectrometry (IM-MS) instrument,
Comparing these observations to the ranking of the mass
spectrometric results, we see an overall good agreement of
theory and experiment. There are no major differences, but only
small deviations. One such deviation lies in the comparison of
D2 and DC4. However, their calculated difference of association
enthalpies is below 6 kJ/mol, which is not significant. DC2 is
21
]
which is coupled to
a
free electron laser (FEL).[
IR
spectroscopy is sensitive towards inter- and intramolecular
interactions[22] and therefore can further support our previous
This article is protected by copyright. All rights reserved.

10.1002/chem.201800893
Chemistry - A European Journal
FULL PAPER
Figure 9. Experimental (black trace) and theoretical (red trace) gas-phase IR spectra for chloride complexes a) with the pyridine receptors DC2 (top) and DC4
(bottom) and b) with the pyridine receptors D2 (top) and D4 (bottom). A comparison between experimentally determined collision cross sections in Å2 (black)
and theoretically determined values (red) shows an excellent agreement for all complexes. Theoretical IR spectra were obtained at the TPSS-D3(BJ)/def2-TZVP
level. A Gaussian distribution with a width of 0.5% with respect to the IR wavenumber was used to convolute the theoretical IR spectra.
assignment based on MS. The recorded IR spectra in the range
of 1000 to 1800 cm-1 show multiple bands (Figure 9). Each pair
(DC2/D2 and DC4/D4) shows similar vibrational shifts and
similar relative intensities. For the DC2/D2 pair, the most intense
vibration is observed at ~1630 cm-1, corresponding to a C=C
vibration, while for the chloride complexes with DC4 and D4 CH2
and CH3 vibrations gain in relative intensity (1400 – 1530 cm-1).
In general, the comparison between experimental (black) and
theoretical (red) IR spectra is in good agreement. Furthermore,
we measured the experimental collision cross section (CCS) of
these complexes.[23] The CCS is a molecular property, which
provides information of the ion’s overall shape and it can also be
calculated theoretically. The CCSs of the investigated
In this study, a series of fluoro-substituted mono- and
divalent 4-aminopyridines and one divalent 4-alkoxypyridine
were investigated with respect to their capability to bind anions
using tandem mass spectrometry and density functional theory
at the TPSS-D3(BJ)/def2-TZVP level. As the aromatic systems
of these receptors are highly electron deficient, the main
question was whether they interact through anion- interactions
(-hole interactions) or through an alternative binding mode. The
4-aminopyridines differ in degree and pattern of fluorine
substitution, configuration, and nature of the spacer, which
connects the two pyridine substituents in the divalent
compounds. The experiments show that monovalent pyridines
do not form detectable complexes with anions, whereas most of
complexes range from 110 to 122 Å2 (DC2: 122 Å2, DC4: 117 Å2, the divalent derivatives interact even with weakly coordinating
D2: 110 Å2, D4: 112 Å2) and they are in excellent agreement
with theoretically predicted values (DC2: 122 Å2, DC4: 117 Å2,
D2: 109 Å2, D4: 112 Å2) using the projection approximation (PA)
within the software sigma.[ 24 ] The good agreement of the
structures and CCS values from IR spectroscopy and IM-MS
strongly support the evolved rankings from mass spectrometry
and theory.
anions. The ranking of the qualitative binding affinity from
experimental data agrees nicely with the trends found in density
functional calculations. The molecules with rigid spacers showed
a higher or at least equal affinity to chloride than their flexible
analogues. The optimized structures of the monovalent and
divalent chloride complexes clearly show that the anion binds to
several C-H moieties on the spacers in the most stable
structures, while anion- interactions do not play a substantial
role. Surprisingly, the best binding receptors are not the
perfluoro-substituted aminopyridines, but the 2,6-fluorinated
derivatives. This is easily explained by an additional C-H···anion
contact to one of the pyridine C-H bonds. In conclusion, our
Conclusions
This article is protected by copyright. All rights reserved.

10.1002/chem.201800893
Chemistry - A European Journal
FULL PAPER
Ionspec QFT-7 FT-ICR mass spectrometer (Agilent/Ionspec, Lake Forest,
results strongly indicate that the investigated fluoro-substituted
4-aminopyridine derivatives do not exhibit significant anion--
interactions with chloride, but rather form strong hydrogen bonds
with polarized C-H-bonds.
CA, USA), equipped with
a 7 T superconducting magnet and a
Micromass Z-spray electrospray ionization source (Waters Co., Saint-
Quentin, France). The sample solutions were introduced into the ion
source at flow rates of 4 μL/min. A constant spray and highest intensities
were achieved with a capillary voltage of 3.8 – 4.2 V and a source
temperature of 40 °C. The parameters for sample cone and extractor
cone voltages as well as the ion optics were optimized for maximum
abundances of the desired complex ions. Multiple scans (up to 50) were
averaged for each spectrum in order to improve the signal-to-noise ratio.
Collision-induced dissociation (CID) was achieved by employing an
isolation function for the cyclotron frequencies of the two corresponding
pyridine anion complexes and integrating a 5 ms gas pulse in the setup.
Experimental Section
For general information see Supporting Information.
The syntheses of compounds M5[25 ] and M8[26] was reported in the
literature. The preparation of D1, D2, D3, D4, DC1, DC2, DC3, DC4 and
DC5 has previously been reported by our group.[18]
Method B: Electrospray ionization quadrupole-time-of-flight high-
resolution mass spectrometric (ESI-Q-TOF-HRMS) experiments were
performed with a Synapt G2-S HDMS (Waters Co., Milford, MA, USA) in
negative ion mode. Typical instrument parameters were as follows: flow
rate: 20 µL min-1; capillary voltage: 1.6 kV; sample cone voltage: 10 V;
source offset: 80 V; nebulizer gas: 6 bar; desolvation gas flow: 500 L/h.
Near these settings, these parameters were optimized for maximum
abundance of the desired intact [M·anion]- complex and minimum
abundance of the corresponding gas-phase fragmentation products. CID
experiments were employed for MS/MS experiments with Ar as the
fragmentation gas. Complex ions of interest were m/z-selected in the
quadrupole region and fragmented in the transfer cell with increasing
collision energies from 0 to 12 V.
3,5-Difluoro-(N,N-dimethylamino)pyridine (M4): 2,3,5,6-Tetrafluoro-
(N,N-dimethylamino)pyridine M8 (0.300 g, 1.55 mmol), titanocene
difluoride (0.050 g, 0.23 mmol, 0.15 equiv.), diphenylsilane (1.71 mL,
9.27 mmol, 6 equiv.) and dry 1,4-dioxane (4 mL) were placed into a
Schlenk flask and the resulting mixture was subsequently degassed by
several freeze-pump-thaw cycles. Subsequently, the mixture was heated
for a few seconds with a heat gun until the yellow color changed into
dark-brown and the solution was then stirred at 110 oC for 16 h. The
reaction was quenched with methanol and water was added. Extraction
with Et2O, drying with Na2SO4, concentration and purification by FLC
(hexanes/EtOAc 5:1) furnished the desired product M4 in 71 % yield
(0.176 g) as a slightly yellow oil which solidified in the fridge. 1H NMR
(CDCl3, 500 MHz): δ = 3.01 (mc, 6 H, CH3), 8.02 (mc, 2 H, Py) ppm. 13C
NMR (CDCl3, 125 MHz): δ = 42.7 (q, CH3), 134.8-135.0 (m, C-2, C-4, C-
6), 152.0 (dd, JCF = 253.9, 4.4 Hz, C-3, C-5) ppm. 19F NMR (CDCl3, 471
MHz): δ = -136.9 (s, 2 F, 3-F, 5-F) ppm. HRMS (ESI-TOF): calcd. for
C7H8F2N2: 159.0734 [M+H]+; found: 159.0731.
Quantum chemical calculations. The fluorinated aminopyridine
derivatives and their anion complexes were investigated by density
functional theory (DFT) at the TPSS-D3(BJ)/def2-TZVP[27] level. D3(BJ)
refers to Grimme's D3 dispersion correction[ 28 ] with Becke-Johnson
damping.[29] We opted for this meta-GGA functional due to its outstanding
performance in terms of computational efficiency, especially in
combination with the RI-DFT[30] method. Preliminary work showed that
this level of theory competes well with computationally much more
demanding M06-2X/aug-cc-pVTZ [27] calculations[ 31 ] (Table S4). All
structures were fully optimized, whereby local minima were verified via
normal mode analyses showing no imaginary frequencies. With these
optimized structures, free enthalpies of association (ΔGa) were computed
by a method proposed by Grimme[32] that incorporates a so-called rigid-
rotor harmonic-oscillator approximation (RRHO) as a correction to low-
1,2-Bis[(2,3,5,6-tetrafluoropyridine-4-yl)oxy]ethane (D5): Ethylene
glycol (0.148 g, 2.38 mmol), pentafluoropyridine (0.81 g, 4.79 mmol, 2
equiv.) and K2CO3 (0.794 g, 5.75 mmol) were dissolved in DMSO (5 mL)
and stirred for 16 h at rt. After addition of water extraction with Et2O and
purification by FLC (hexanes/EtOAc 6:1) furnished product D5 in 49 %
yield (0.425 g) as a colorless oil, which solidified in the fridge. 1H NMR
(CDCl3, 500 MHz): δ = 4.87 (mc, 4 H, CH2) ppm. 13C NMR (CDCl3, 125
MHz): δ = 72.7 (t, CH2), 135.1, 144.4, 146.8 (3 mc, C-2, C-3, C-4, C-5, C-
6) ppm. 19F NMR (CDCl3, 471 MHz): δ = -158.6 (mc, 4 F, 3-F, 5-F), -89.9
(mc, 4 F, 2-F, 6-F) ppm. HRMS (ESI-TOF): calcd. for C12H4F8N2O2:
361.0223 [M+H]+; found: 361.0228.
33
]
lying frequencies, which
often
show anharmonicites:[
ΔGa = ΔECP + ΔGRRHO. ΔECP is the counterpoise-corrected (CP) binding
energy as described by Boys and Bernardi[34] to account for basis set
superposition error (BSSE) and ΔGRRHO is the corrected free enthalpic
contribution. Further explanations can be found in the designated
literature. All calculations were performed with the program package
TURBOMOLE[35] (version 7.0.1). For every pyridine derivative, several
binding motifs were computed, whereby always the lowest binding
energies are presented. Using the pdb structures for the lowest binding
energies, theoretical CCS values can be calculated with the projection
approximation implemented within the software sigma.[24]
Mass spectrometric experiments. Tetra-n-butylammonium salts were
purchased from Sigma Aldrich (Taufkirchen, Germany). Samples for
electrospray-ionization mass spectrometric experiments were prepared
by adding the corresponding pyridine and the corresponding tetra-n-
butylammonium salt in equimolar amounts. Therefore, a 0.5 mM stock
solution of the pyridines in dichloromethane was mixed with a 2 mM
tetra-n-butylammonium anion salt acetonitrile solution in a 4 to 1 ratio to
obtain a concentration of 400 µM each (method B). For experiments
following method A equimolar amounts of two pyridines were mixed
analogous with the corresponding anion solution. Mesylate complexes
did not form at a 1:1 pyridine:anion ratio, so they were mixed in a 1:5
Gas-phase infrared spectroscopy and ion mobility-mass
spectrometry. Gas-phase IR spectra of the size- and mass-selected
chloride/pyridine receptor complexes were measured using an in-house
constructed drift-tube ion mobility-mass spectrometer, similar to one
described previously.[22a,36] Briefly, ions are drift-time pre-selected before
irradiation with intense IR light. The complexes were sprayed from in-
house prepared Pd/Pt-coated borosilicate capillaries and ionized via
pyridine:anion excess. For experiments of method
C a 1 mM
indigocarmine solution (H2O: MeOH 1:1) was mixed with the two desired
pyridines and diluted to a final concentration of 50 µM pyridine and 25
µM indigocarmine.
nano-electrospray ionization using voltages of 0.6
ionization, the generated ions are transferred and stored in an entrance
funnel. Subsequently, the ions are released into a drift tube, which is
– 1.0 kV. After
Method
A and C: Electrospray-ionization (ESI) mass spectra of
anion/pyridine complexes were recorded in negative ion mode on an
This article is protected by copyright. All rights reserved.

10.1002/chem.201800893
Chemistry - A European Journal
FULL PAPER
filled with helium buffer gas. Under the influence of a weak electric field
(10 – 20 V/cm) the ions traverse the tube and the drift time of the ions is
recorded. The drift velocity of a particular ion depends on its mobility,
which in turn is based on its overall size, shape and charge. After exiting
the ion mobility cell, the ions are mass-selected using a quadrupole mass
filter and their arrival time distributions (ATDs) can be recorded by
measuring the time dependent ion current of the m/z selected species
after release of the ion trap. Collision cross section (CCS) values can be
calculated from these ATDs with the Mason Schamp equation.[23]
[3]
[4]
D. A. Dougherty, Science 1996, 271, 163-168.
a) P. Gamez, T. J. Mooibroek, S. J. Teat, J. Reedijk, Acc. Chem. Res.
2007, 40, 435-444; b) B. L. Schottel, H. T. Chifotides, K. R. Dunbar,
Chem. Soc. Rev. 2008, 37, 68-83; c) T. J. Mooibroek, C. A. Black, P.
Gamez, J. Reedijk, Cryst. Growth Des. 2008, 8, 1082-1093; d) O. B.
Berryman, D. W. Johnson, Chem. Commun. 2009, 3143-3153; e) A.
Frontera, P. Gamez, M. Mascal, T. J. Mooibroek, J. Reedijk, Angew.
Chem. Int. Ed. 2011, 50, 9564-9583; f) M. Giese, M. Albrecht, K.
Rissanen, Chem. Rev. 2015, 115, 8867-8895; g) M. Giese, M. Albrecht,
K. Rissanen, Chem. Commun. 2016, 52, 1778-1795; h) Y. Liu, A.
Sengupta, K. Raghavachari, A. H. Flood, Chem 2017, 3, 411-427.
As the ion mobility-mass spectrometer is connected to a free electron
laser, conformer- and mass-selected IR spectra can be recorded.
Therefore, a narrow drift-time window is selected (100 µs width) using
electrostatic deflection prior to mass selection. This m/z- and ion mobility-
selected ion cloud was further irradiated by an intense 10 µs pulse
(54 - 72 mJ/pulse) of IR photons. The photofragmentation is detected by
a time-of-flight (ToF) mass analyzer and IR spectra are obtained by
[5]
[6]
B. P. Hay, V. S. Bryantsev, Chem. Commun. 2008, 0, 2417-2428.
I. Alkorta, I. Rozas, J. Elguero, J. Am. Chem. Soc. 2002, 124, 8593-
8598.
[7]
[8]
H. Schneider, K. M. Vogelhuber, F. Schinle, J. M. Weber, J. Am. Chem.
Soc. 2007, 129, 13022-13026.
plotting the fragmentation yield as
a function of the tunable IR
a) K. Hiraoka, S. Mizuse, S. Yamabe, The Journal of Physical
Chemistry 1987, 91, 5294-5297; b) H.-J. Schneider, F. Werner, T.
Blatter, J. Phys. Org. Chem. 1993, 6, 590-594; c) H.-J. Schneider,
Angew. Chem. Int. Ed. 2009, 48, 3924-3977.
wavenumber. The final IR spectrum represents an average of at least
two individual scans, where each scan was obtained by scanning in
3 cm-1 wavenumber steps and averaging at least 40 spectra for each
step.
The tunable mid-IR light is provided by the Fritz Haber Institute Free
Electron Laser and is transported to the instrument via an evacuated
beam line, where the last two meters of the beam line are flushed with
dry nitrogen to avoid water absorption.
[9]
D. Quiñonero, C. Garau, C. Rotger, A. Frontera, P. Ballester, A. Costa,
P. M. Deyà, Angew. Chem. Int. Ed. 2002, 41, 3389-3392.
[10] M. Mascal, A. Armstrong, M. D. Bartberger, J. Am. Chem. Soc. 2002,
124, 6274-6276.
[11] a) M. Wenzel, J. R. Hiscock, P. A. Gale, Chem. Soc. Rev. 2012, 41,
480-520; b) S. Chakravarty, Z.-Z. Sheng, B. Iverson, B. Moore, FEBS
Lett. 2012, 586, 4180-4185.
Crystal structure determination. X-ray single crystal diffraction of
compounds D4 and DC4 was performed on a Bruker AXS Smart CCD.
The structures were solved by direct methods, using SHELXS-97.[37]
Refinement was done with the least squares method (SHELXL
Version2017/1)[37]. Molecular Graphics: ORTEP-3 for Windows.[19] The
CCDC numbers 1578872 (D4) and 1578873 (DC4) contain the
supplementary crystallographic data for this paper. These data can be
obtained free of charge from the Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
[12] a) V. Gorteau, G. Bollot, J. Mareda, A. Perez-Velasco, S. Matile, J. Am.
Chem. Soc. 2006, 128, 14788-14789; b) J. Mareda, S. Matile, Chem.
Eur. J. 2009, 15, 28-37; c) R. E. Dawson, A. Hennig, D. P. Weimann, D.
Emery, V. Ravikumar, J. Montenegro, T. Takeuchi, S. Gabutti, M.
Mayor, J. Mareda, C. A. Schalley, S. Matile, Nat. Chem. 2010, 2, 533-
538; d) A. Vargas Jentzsch, D. Emery, J. Mareda, P. Metrangolo, G.
Resnati, S. Matile, Angew. Chem. Int. Ed. 2011, 50, 11675-11678.
[13] a) Y. S. Rosokha, S. V. Lindeman, S. V. Rosokha, J. K. Kochi, Angew.
Chem. Int. Ed. 2004, 43, 4650-4652; b) B. L. Schottel, H. T. Chifotides,
M. Shatruk, A. Chouai, L. M. Pérez, J. Bacsa, K. R. Dunbar, J. Am.
Chem. Soc. 2006, 128, 5895-5912; c) G. Gil-Ramírez, E. C. Escudero-
Adán, J. Benet-Buchholz, P. Ballester, Angew. Chem. Int. Ed. 2008, 47,
4114-4118; d) H. T. Chifotides, B. L. Schottel, K. R. Dunbar, Angew.
Chem. Int. Ed. 2010, 49, 7202-7207; e) M. Giese, M. Albrecht, C.
Bannwarth, G. Raabe, A. Valkonen, K. Rissanen, Chem. Commun.
2011, 47, 8542-8544; f) M. Giese, M. Albrecht, C. Plum, D. Hintzen, A.
Valkonen, K. Rissanen, Supramol. Chem. 2012, 24, 755-761; g) D.-X.
Wang, M.-X. Wang, J. Am. Chem. Soc. 2012, 135, 892-897; h) H. T.
Chifotides, K. R. Dunbar, Acc. Chem. Res. 2013, 46, 894-906; i) M.
Albrecht, H. Yi, O. Köksal, G. Raabe, F. Pan, A. Valkonen, K. Rissanen,
Chem. Eur. J. 2016, 22, 6956-6963; j) A. Kim, R. Ali, S. H. Park, Y.-H.
Kim, J. S. Park, Chem. Commun. 2016, 52, 11139-11142; k) L. M. Eytel,
A. K. Gilbert, P. Görner, L. N. Zakharov, D. W. Johnson, M. M. Haley,
Chem. Eur. J. 2017, 23, 4051-4054; l) Y.-Z. Liu, K. Yuan, L. Liu, Z.
Yuan, Y.-C. Zhu, J. Phys. Chem. A 2017, 121, 892-900.
Acknowledgements
We thank Dr. Andreas Springer for introducing M.G. to the mass
spectrometers and for inspiring discussions. We further
acknowledge the Fritz-Haber-Institute free electron laser facility,
in particular S. Gewinner, W. Schöllkopf and G. von Helden for
providing us with many IR photons. This work was made
possible using the High-Performance Computing resources of
the Zentraleinrichtung für Datenverarbeitung (ZEDAT) of Freie
Universität
Berlin.
Funding
from
the
Deutsche
Forschungsgmeinschaft (CRC 765, GK 1582 and project
LE423/17-1) is gratefully acknowledged.
[14] B. P. Hay, R. Custelcean, Cryst. Growth Des. 2009, 9, 2539-2545.
Keywords: fluorinated pyridines • anion- interactions • C-H-
anion interactions • tandem mass spectrometry • density
functional theory
[15] a) W. D. Price, P. D. Schnier, E. R. Williams, J. Phys. Chem. B 1997,
101, 664-673; b) E. C. Kampen, J. S. Brodbelt, Anal. Chem. 2000, 72,
5411-5416; c) P. B. Armentrout, Top. Curr. Chem. 2003, 225, 233-262.
[16] A. emboeuf, A. asioudis, S. ndelicato, . ollreis , A. k. uki, S. n.
ki, . . van den Brink, . r. key, L. s. rahos, Anal. Chem. 2010,
82, 2294-2302.
References
[1]
a) C. L. Perrin, J. B. Nielson, Annu. Rev. Phys. Chem. 1997, 48, 511-
544; b) I. Alkorta, J. Elguero, Chem. Soc. Rev. 1998, 27, 163-170.
[17] R. Graham Cooks, J. S. Patrick, T. Kotiaho, S. A. McLuckey, Mass
Spectrom. Rev. 1994, 13, 287-339.
[2]
S. Grimme, Angew. Chem. Int. Ed. 2008, 47, 3430-3434.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800893
Chemistry - A European Journal
FULL PAPER
[18] a) G. Podolan, D. Lentz, H.-U. Reissig, Angew. Chem. Int. Ed. 2013, 52,
9491-9494; b) G. Podolan, P. Jungk, D. Lentz, R. Zimmer, H.-U.
Reissig, Adv. Synth. Catal. 2015, 357, 3215-3228.
[27] a) J. Tao, J. P. Perdew, V. N. Staroverov, G. E. Scuseria, Phys. Rev.
Lett. 2003, 91, 146401; b) F. Weigend, R. Ahlrichs, PCCP 2005, 7,
3297-3305.
[19] L. J. Farrugia, J. Appl. Crystallogr. 1997, 30, 565.
[28] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132,
154104.
[20] A. Bondi, J. Phys. Chem. 1964, 68, 441-451.
[29] S. Grimme, S. Ehrlich, L. Goerigk, J. Comput. Chem. 2011, 32, 1456-
[21] S. Warnke, G. von Helden, K. Pagel, Proteomics 2015, 16, 2804-2812.
1465.
[22] a) S. Warnke, J. Seo, J. Boschmans, F. Sobott, J. H. Scrivens, C.
Bleiholder, M. T. Bowers, S. Gewinner, W. Schöllkopf, K. Pagel, G. von
Helden, J. Am. Chem. Soc. 2015, 137, 4236-4242. b) W. Hoffmann, M.
Marianski, S. Warnke, J. Seo, C. Baldauf, G. von Helden, K. Pagel,
Phys. Chem. Chem. Phys. 2016, 18, 19950-19954.
[30] M. Von Arnim, R. Ahlrichs, J. Comput. Chem. 1998, 19, 1746-1757.
[31] T. H. D. Jr., J. Chem. Phys. 1989, 90, 1007-1023.
[32] S. Grimme, Chem. Eur. J. 2012, 18, 9955-9964.
[33] P. Y. Ayala, H. B. Schlegel, J. Chem. Phys. 1998, 108, 2314-2325.
[34] S. B. Boys, F. Bernardi, Mol. Phys. 1970, 19, 553-566.
[23] E. A. Mason, H. W. Schamp, Ann. Phys. 1958, 4, 233-270.
[24] G. von Helden, M. T. Hsu, N. Gotts, M. T. Bowers, J.Phys.Chem. 1993,
97, 8182-8192.
[35] R. Ahlrichs, M. Bär, M. Häser, H. Horn, C. Kölmel, Chem. Phys. Lett.
1989, 162, 165-169.
[25] M. Schlosser, C. Bobbio, T. Rausis, J. Org. Chem. 2005, 70, 2494-
2502.
[36] P. R. Kemper, N. F. Dupuis, M. T. Bowers, Int. J. Mass Spectrom.
2009, 287, 46-57.
[26] R. Ranjbar-Karimi, M. Mashak-Shoshtari, A. Darehkordi, Ultrason.
Sonochem. 2011, 18, 258-263.
[37] G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr. 2008, 64,
112-122.
This article is protected by copyright. All rights reserved.

10.1002/chem.201800893
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
Anion- Interactions 
Do electron-deficient fluorinated
pyridines bind anions through anion-
interactions in the gas phase? Tandem
mass spectrometric experiments with
anion complexes of divalent fluorinated
pyridines reveal trends of binding
strengths that clearly rule out anion-
interactions to be important for complex
formation. Density functional
calculations instead suggest binding
through C-H···anion contacts with
polarized C-H bonds of the spacers
between the two pyridines.
M. Göth, F. Witte, M. Quennet, P.
Jungk, G. Podolan, D. Lentz, W.
Hoffmann, K. Pagel, H.-U. Reissig,* B.
Paulus,* Christoph A. Schalley*
……...… Page – Page
To Anion- or not to Anion-: The
Case of Anion-Binding to Divalent
Fluorinated Pyridines in the Gas
Phase
This article is protected by copyright. All rights reserved.