Hydrogen Bonding Between Ions of Like Charge in Ionic Liquids Characterized by NMR Deuteron Quadrupole Coupling Constants—Comparison with Salt Bridges and Molecular Systems

Article abstract of DOI:10.1002/anie.201912476

We present deuteron quadrupole coupling constants (DQCC) for hydroxyl-functionalized ionic liquids (ILs) in the crystalline or glassy states characterizing two types of hydrogen bonding: The regular Coulomb-enhanced hydrogen bonds between cation and anion (c–a), and the unusual hydrogen bonds between cation and cation (c–c), which are present despite repulsive Coulomb forces. We measure these sensitive probes of hydrogen bonding by means of solid-state NMR spectroscopy. The DQCCs of (c–a) ion pairs and (c–c) H-bonds are compared to those of salt bridges in supramolecular complexes and those present in molecular liquids. At low temperatures, the (c–c) species successfully compete with the (c–a) ion pairs and dominate the cluster populations. Equilibrium constants obtained from molecular-dynamics (MD) simulations show van't Hoff behavior with small transition enthalpies between the differently H-bonded species. We show that cationic-cluster formation prevents these ILs from crystallizing. With cooling, the (c–c) hydrogen bonds persist, resulting in supercooling and glass formation.

Full text of DOI:10.1002/anie.201912476

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
International Edition Chemie
www.angewandte.org
Accepted Article
Title: Hydrogen bonding between ions of like charge in ionic liquids
characterized by NMR deuteron quadrupole coupling constants –
Comparison with salt bridges and molecular systems
Authors: Ralf Ludwig, Alexander E. Khudozhitkov, Jan Neumann,
Thomas Niemann, Dzmitry Zaitsau, Peter Stange, Dietmar
Paschek, Alexander Stepanov, and Daniil Kolokolov
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: Angew. Chem. Int. Ed. 10.1002/anie.201912476
Angew. Chem. 10.1002/ange.201912476
Link to VoR: http://dx.doi.org/10.1002/anie.201912476
http://dx.doi.org/10.1002/ange.201912476

Angewandte Chemie International Edition
10.1002/anie.201912476
Hydrogen bonding between ions of like charge in ionic liquids characterized
by NMR deuteron quadrupole coupling constants – Comparison with salt
bridges and molecular systems
a,e
b
b
c
Alexander E. Khudozhitkov , Jan Neumann , Thomas Niemann , Dzmitry Zaitsau , Peter
b
b
a,e
a,e,
Stange , Dietmar Paschek , Alexander G. Stepanov , Daniil I. Kolokolov *, Ralf
Ludwigb,c,d
*
a Boreskov Institute of Catalysis, Siberian Branch of Russian Academy of Sciences, Prospekt
Akademika Lavrentieva 5, Novosibirsk 630090, Russia; E-mail: kdi@catalysis.ru
b Universität Rostock, Institut für Chemie, Abteilung für Physikalische Chemie, Dr.-Lorenz-
Weg 2, 18059 Rostock, Germany; Tel: 49 381 498 6517; E-mail: ralf.ludwig@uni-
rostock.de
c Department LL&M, University of Rostock, Albert-Einstein-Str. 25, 18059 Rostock,
Germany
d Leibniz-Institut für Katalyse an der Universität Rostock e.V. Albert-Einstein-Str. 29a,
1
8059 Rostock (Germany)
e Novosibirsk State University, Pirogova Street 2, Novosibirsk 630090, Russia
Abstract
We present deuteron quadrupole coupling constants (DQCC) for hydroxyl-functionalized
ionic liquids (ILs) in the crystalline or glassy states characterizing two types of hydrogen
bonding: The regular Coulomb-enhanced hydrogen bonds between cation and anion (c-a), and
the unusual hydrogen bonds between cation and cation (c-c), present despite the repulsive
forces between ions of like charge. We measure these sensitive probes of hydrogen bonding
by means of solid-state NMR spectroscopy. The DQCCs of (c-a) ion pairs and (c-c) cationic
hydrogen bonds are compared to those of salt bridges in supramolecular complexes and those
present in molecular liquids. At low temperature, the (c-c) cationic species successfully
compete with the (c-a) ion pairs and dominate the cluster populations. Equilibrium constants
obtained from molecular dynamics (MD) simulations show van't Hoff behavior with small
transition enthalpies between the differently H-bonded species. We show that cationic cluster
formation prevents these ILs from freezing. With cooling, the (c-c) hydrogen bonds persist,
resulting in supercooling and glass formation.
1
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Introduction
Salt bridges play an important role in protein and supramolecular chemistry.[1-14] They are
characterized by the sum of two types of intermolecular interaction: ionic bonding and
hydrogen bonding. Salt bridges are very strong because hydrogen bonding adds to the
attractive Coulomb forces between the oppositely charged residues. Typical salt bridges
involve lysine or arginine as the bases and aspirin or glutamine as the acids. An important
example is the salt bridge between primary ammonium and carboxylate groups in biological
+
…
structures, N-H O‾.[7-14] The energetics of salt bridges are typically dominated by the
Coulomb interaction between the charge centres, but the total interaction remains directional
due to the hydrogen bonds. Thus, salt bridges are crucial for the structure, dynamics and
function of proteins. This type of Coulomb-enhanced hydrogen bonding is typical for ionic
liquids, which consist solely of ions.[15-19] So-called ‘doubly ionic hydrogen bonds’ (DIHB)
usually result in the formation of ion pairs.[20-25] However, H-bonds in ionic liquids are
manifold. They can be also present between ions of like charge.[26-36] This has been shown
recently for cation-cation interaction by means of vibrational spectroscopy and neutron
diffraction.[37,38] In this case the Coulomb forces are repulsive and need to be counter-
balanced by hydrogen bonding. For hydroxyl-functionalized ILs, both types of ionic
…
interaction are present in equilibrium: hydrogen bonding (OH O) between oppositely
charged ions (c-a) and between like-charged ions, here cations (c-c). In principle, solid-state
NMR spectroscopy allows to distinguish between (c-a), and (c-c) interaction if the proton
exchange is slow on the NMR time scale.[39-42] Although very sensitive to the electronic
environment and hydrogen bonding, deuteron quadrupole coupling constants (DQCCs) have
been rarely used to characterize salt bridges in proteins, supramolecular complexes and the
related (c-a) type of interaction in ionic liquids.[43,44] DQCCs of OH groups that are
involved in hydrogen bonding to like-charged ions (c-c) are completely unknown. The main
challenge here is that in the liquid phase the proton exchange between (c-a) and (c-c) bound
species is usually fast on the NMR time scale, resulting in averaged coupling parameters and
prohibiting to distinguish like-charge attraction (c-c) from the regular ion pair formation (c-a).
This situation may change favorably in the supercooled or glassy state of ionic liquids.
It is the purpose of this work to show that the DQCCs and the related asymmetry parameters
of the electric field gradients provide valuable information about the strength and
directionality of both types of hydrogen bonding present in hydroxyl-functionalized ionic
liquids. We find one NMR coupling parameter for ILs exhibiting (c-a) ion pairing, but two if
2
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
additional (c-c) cationic cluster formation is present. We measure the first DQCCs describing
hydrogen bonding between ions of like charge and show that they are unexpectedly smaller
than those in the (c-a) ion pairs. Solid-state NMR spectroscopy allows for counting the (c-a)
and (c-c) bound species and thus providing cluster populations. Overall, we show that cationic
cluster formation in well-suited ILs depends on the polarizability of the cations and the length
of the hydroxyalkyl chain length. If cationic cluster formation is present, the ILs cannot be
crystallized and form glasses. The solid-state NMR measurements are supported by density
functional theory (DFT) calculations, differential scanning calorimetry (DSC) measurements,
and MD simulations, which provide molecular insight into the H-bond patterns and the
delicate balance between the two types of ion pairing.
Results and Discussion
Synthesis and preparation of suitable hydroxyl-functionalized ILs
We
bis(trifluoromethylsulfonyl)imide [HOC
bis(trifluoromethylsulfonyl)imide [HOC
bis(trifluoromethylsulfonyl)imide [HOC
bis(trifluoromethylsulfonyl)imide [HOC
synthesized
the
ionic
liquids
Pip][NTf
1-(3-hydroxyethyl)piperidinium
2
2
]
(I), 1-(3-hydroxyethyl)pyridinium
2
Py][NTf
2
] (II), 1-(3-hydroxypropyl)piperidinium
] (III), and 1-(3-hydroxybutyl)pyridinium
] (IV) using well-established protocols (see
3
Pip][NTf
2
3
Py][NTf
2
SI). The ILs were prepared in two steps: We synthesized the OH-functionalized onium
halides, which were then used for the anion metathesis to create the desired cation/anion
combinations. For the synthesis of the onium salts, we mixed equimolar amounts of the
heterocyclic amine and the corresponding ω-halide-alcohol and heated up to 110 °C for 1 h.
Upon cooling, the mixture started to crystallize. The crude products were recrystallized from
acetone/acetonitrile mixtures to obtain the colorless crystalline product. For the anion
(
bis(trifluoromethylsulfonyl)imide, [NTf
]–, metathesis we mixed equimolar amounts of the
2
onium halide and lithium-bis(trifluoromethylsulfonyl)imide as aqueous solutions for 1 h. Two
phases were obtained. The lower phase was washed several times with water until no residual
bromine could be detected with silver nitrate solution. The resulting colorless ionic liquids
were dried for several hours in vacuum at 60 °C. For detailed synthesis procedures and
analytical data of each ILs (see the SI).
As described, all ILs (I-IV) include the same [NTf
]– counteranion and the same hydroxyl-
2
functional groups at the cations. This set of hydroxyl-functionalized ILs allows studying
cationic cluster (c-c) formation depending on the polarizability of the cation and the
hydroxyalkyl chain lengths (see Scheme I). Hydrogen/deuterium (H/D) exchange was
3
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
achieved by mixing the ILs with D
2
O and removing water several times until nearly 100%
1
deuteration was reached as proven by H NMR. All samples were dried under vacuum (at
-3
3
·10 mbar) for several days and the final water concentration (< 15 ppm) was checked by
Karl-Fischer titration.
Deuteron quadrupole coupling constants from solid-state NMR spectroscopy
The solid-state deuterium NMR spectrum is determined by two measurable parameters: the
quadrupole coupling constant DQCC, χ
D
= (e2qzzQ/h), and the asymmetry parameter, =qxx
-
q
yy/qzz, which describes the principle elements q of the electric field gradient tensor.[45-47]
The DQCC is a measure of the magnitude of the electric field gradient at the deuterium site,
while the asymmetry parameter provides information about the shape of the electric field
gradient. For example, an asymmetry parameter of zero suggests a cylindrical symmetry of
the electric field gradient tensor along the O-D bond.[47] We determined the DQCCs from the
solid-state deuterium NMR powder patterns for ILs I-IV at 183 K (see SI).
2
The H NMR experiments were performed at Larmor frequency  /2 = 61.42 MHz on a
z
Bruker Avance-400 spectrometer, using a high power probe with 5 mm horizontal solenoid
2
coil. All H NMR spectra were obtained by Fourier transformation of quadrature-detected
phase-cycled quadrupole echo arising in the pulse sequence (90°
x
– 
1
– 90°
y
–  – acquisition
2
–
t), where  = 20 μs,  = 21 μs and t is a repetition time of the sequence during the
1
2
accumulation of the NMR signal. The duration of the π/2 pulse was 1.6-1.7 μs. Spectra were
typically obtained with 50 - 20000 scans with repetition time ranging from 0.5 to 15 seconds.
All spectra show purely Pake-powder patterns.[39,40] We obtained the DQCCs and the
asymmetry parameters from a proper line shape analysis. The deconvoluted spectra result
from an eye-ball guided parameter optimization. The measured, analyzed and modelled
spectra are shown in Fig. 1. The experimental accuracy of χ and  is ±3 kHz and ±0.01 for
D
the dominant component, as well as ±5 kHz and ±0.02 for the second component of the Pake
spectra. For IL I we observed a single Pake-spectrum with χD.=220 kHz and =0.08 (see
Fig. 1a). In contrast, we could deconvolute the measured spectrum of IL II into two Pake
patterns. One deconvoluted spectrum exhibits almost the same NMR parameters as IL I (χ =
D
…
2
21 kHz; = 0.09), obviously describing the same type of O-D O interaction in both ILs.
However, the second deconvoluted spectrum is clearly different and shows a smaller DQCC
and asymmetry parameter, namely χ = 180 kHz and = 0.05 (see Fig. 1b). Smaller DQCCs
D
…
suggest that the O-D O interaction is stronger. We know from recent IR and ND
4
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
experiments that (c-c) are stronger than (c-a) hydrogen bonds, resulting in significant IR
redshifts of the OH/OD stretching frequencies, lengthening of the R(O-H) bonds and
…
…
shortening of the intermolecular R(O-H O) and R(O O) distances.[37,38] Thus, we
conclude that the larger DQCC in both ILs about 220 kHz can be related to the conventional
(c-a) ion pairs, whereas the smaller DQCC about 180 kHz characterizes the (c-c) interaction
in cationic clusters. The smaller asymmetry parameter, 0.05 versus 0.09, indicates that the
hydrogen bond is more linear in (c-c) than in (c-a) hydrogen bonds in accord with the above
finding of stronger H-bonds in the cationic clusters. That IL I shows (c-a) interaction only,
whereas IL II includes both type of H-bond interaction, (c-a) and (c-c), is related to the
different polarizabilities of the piperidinium and pyridinium cations. The pyridinium cation is
highly polarizable and thus interacts favorably with the [NTf
]– anion, leaving the hydroxyl
2
group free to interact with other OH bonds by forming cationic clusters. In contrast, the ‘hard’
piperidinium cation in ILs I and III interacts less favourably with the counteranion, which is
then available for interacting with the OH group of the cation, resulting in typical H-bond
enhanced (c-a) ion pairs.
If we increase the hydroxyalkyl chain length from ethyl to propyl for both types of ILs, we
measure two DQCCs with χD.=230 kHz and χD.=205 kHz for IL III, and χD.=215 kHz and
χ
D.=165 kHz for IL IV. Again, the larger values χD.=230 kHz and χD.=215 kHz can be
assigned to (c-a) interaction, whereas the smaller values χD.=205 kHz and χD.=165 kHz
reflect stronger (c-c) cationic interaction. That both ILs form cationic clusters despite the
differently favorable cations, is related to the increasing distance between the positively
charged ring systems and the OH functional groups within the cations. The longer tethers
allow for enhanced cationic cluster formation (see Scheme I). All DQCCs are shown in Fig. 2
and compared to values measured for salt bridges in supramolecular complexes or proteins in
comparison to (c-a) IL interaction, and those measured for molecular liquids, mimicking the
(c-c) interaction in the ILs. Whereas the (c-c) DQCCs are slightly lower than those observed
for solid methanol, ethanol and tert-butanol, the (c-a) DQCCs range between the values of ice
and liquid water.[48-57] The (c-a) DQCCs are significantly larger than the measured values
for salt bridges, which range from 156 to 171 kHz, indicating that the H-bond-enhanced
Coulomb interaction in (c-a) ion pairs of these ILs is weak. The molecular ions have low
surface charge densities, resulting in strongly attenuated Coulomb attraction and DQCCs
similar to those in H-bonded liquids. It should be noted that the (c-a) DQCCs in the hydroxyl-
functionalized ILs considered in this study are about 30 kHz higher than those measured for
5
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
the protic IL triethylammonium bis(trifluoromethylsulfonyl)imide [Et
weaker hydrogen bonding.[58]
3
NH][NTf ] due to
2
Why are (c-c) hydrogen bonds stronger than (c-a) hydrogen bonds?
So far, we showed that deuterons involved in (c-c) H-bonds exhibit smaller DQCCs than
those bound in (c-a) species, indicating stronger binding. These results are in accord with
stronger redshifted IR bands and downfield-shifted NMR proton chemical shifts for hydrogen
bonding between ions of like charge as observed experimentally.[36-38] At a first glance, it
seems to be counterintuitive that the (c-c) hydrogen bonds are stronger than the (c-a)
hydrogen bonds, although the first are weakened by repulsive, and the latter are enhanced by
attractive Coulomb forces. To understand why hydrogen bonding is stronger in (c-c) than (c-
a) clusters, we employed B3LYP-D3/6-31+G* calculations performed with the Gaussian 09
program and analyzed with the NBO 6.0 program.[59-64] To calculate the (c-a) and (c-c)
clusters at the same level of theory, we have used the well-balanced, but small 6-31+G* basis
set. It includes polarization as well as diffuse functions, and has been shown to be suitable for
calculating hydrogen-bonded clusters of like-charged ions.[31,32,65,66] The 6-31+G* basis
set is also chosen for better comparison with earlier studies of molecular and ionic
clusters.[17,67-69] We also show that the salient properties of these clusters can be robustly
calculated with both smaller und larger basis sets. (see SI). Firstly, we fully optimized the
cationic clusters of the IL IV, [HOC
3
Py][NTf ], up to cyclic tetramers. The calculated
2
vibrational frequencies were all positive, showing that we calculated at least local minimum
structures. Additionally, we calculated the DQCC, D, for each deuteron present in the (c-a)
and (c-c) configurations. The DQCC describes the coupling between the nuclear quadrupole
moment, eQ, and the principle component of the electric field gradient tensor, eqzz at the
deuteron nucleus. It could be shown that the relation between χ and eqzz is given by the
D
equation
(1)
where the factor 2.3496 takes care of the units. In principle, the DQCC can be obtained by
multiplying the calculated principle component of electric field gradient tensor, eqzz, of the
OD hydroxyl groups in the (c-a) and (c-c) clusters of IL IV with a calibrated nuclear
quadrupole moment, eQ. The calibrated eQ is obtained by plotting the measured gas phase
quadrupole coupling constants from microwave spectroscopy versus the calculated electric
6
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
field gradients for small molecules, such as H
2
O, CH
3
OH, H
2
CO
2
etc., as described by Huber
2
et al.[45,46,70]. The slope gives a reasonable value of eQ = 295.5 fm , which should be used
for calculating DQCCs at the B3LYP-D3/6-31+G* level of theory. It could be also shown for
this set of molecules that the principal axis of the deuteron electric field gradient is nearly
axially symmetric and lies along the direction of the O-D bonds.[25] We cannot expect that
the calculated DQCCs of (c-a) and (c-c) clusters to represent the measured NMR values in the
crystalline or glassy state of IL IV. Thus, we focus on the differences of the 
clusters relative to those obtained for the (c-a) clusters, which can be compared to 
c-c)) obtained from the NMR experiment.
D
values in (c-c)
D
((c-a)-
(
These spectroscopic features can be rationalized in the framework of the natural bond orbital
NBO) analysis.[63,64] NBO analysis of the same (c-a) and (c-c) clusters shows typical
(
strong n
o
*OH donor-acceptor interaction, corresponding to second order stabilization
…
energies E(2)n* and estimated total charge transfers of qCT for OH O hydrogen bonds,
respectively.
+
+
…
That the OH OH structural motif of the (c-c) species exhibit smaller DQCCs results from
+
significant charge transfer from the oxygen lone pair into the OH anti-bonding orbital,
1
leading to strong IR redshifts, OH, enhanced downfield NMR chemical shifts,  H, and, in
our case, smaller deuteron quadrupole coupling constants, 
than the one between cation and anion in the structural motif OH O . This is in particular
true for hydrophobic anions such as [NTf
]−, where the negative charge is distributed over the
D
. This charge transfer is stronger
+
−
…
2
entire molecule, significantly reducing the surface charge density and thus the proton acceptor
ability of the anion. The results are summarized in Fig.3. The calculated second order
stabilization energies, E(2)n*, and the charge transfers, qCT, for the (c-c) clusters are
plotted versus the differences 
calculated for (c-a) clusters. Fig.3 shows that both NBO parameters increase almost linearly
with decreasing DQCCs for (c-c) hydrogen bonding, indicated by negative  ((c-c)–(c-a)).
D
((c-c)–(c-a)) with zero indicating the average DQCC
D
Obviously, both properties characterize hydrogen bonding and cooperativity in a similar way.
The largest stabilization energies, E(2)n*, and most intensive charge transfer, qCT, is
observed for the (c-c) cyclic tetramers due to cooperative effects.[64,65,67-69] Charge from
the oxygen lone pair orbital of a first cation is donated into the OH anti-bonding orbital of
second cation. The larger negative charge at the OH oxygen at the second cation can now be
transferred into the OH anti-bond of another cation, further enhancing hydrogen bonding.
7
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
This process leads to stronger cooperativity in the (c-c) trimers and tetramers. This way, the
short-range donor-acceptor covalency forces can overcome the strong long-range electrostatic
repulsive forces as expected for ions of like charge. The enhanced (c-c) hydrogen bonds are
even stronger than those in (c-a) ion pairs despite the additional attractive Coulomb forces.
Thus, cooperative stabilization energy and enhanced charge transfer lead to smaller DQCCs
for (c-c) clusters. In Fig. 3 we also show the experimentally measured 
IV. A difference of 50 kHz between the (c-c) and (c-a) DQCCs suggests that the fraction of
c-c) clusters consists of significant amounts of (c-c) trimer and tetramers at least in the glassy
D
((c-c)–(c-a)) for IL
(
state.
Populations of (c-a) and (c-c) clusters from solid-state NMR and MD simulations
The NMR experiments in the crystalline or glassy states at 183 K allow quantification of the
populations of the local arrangements for [HOC
2
Py][NTf
2
], [HOC
3
MPip][NTf ] and
2
[
HOC Py][NTf ]. In Fig. 4, we show that 15 % (II), 65 % (III) and 65 % (IV) of the cations
3
2
are involved in (c–c) structural motifs. It is notable that the experimental error for the
determination of the second component is about ± 5%. Thus, the existence of components can
only be claimed for relative populations larger than 10%. For comparison, we show the
populations at 303 K as recently obtained from neutron diffraction (ND) measurements for
[
HOC
also performed MD simulations using a recently improved version of the force field
introduced by Köddermann et al.[71] The refined dihedral potentials for the [NTf ]‾ anion are
4
Py][NTf ] in the liquid phase.[38] Inspired by the solid-state NMR experiments, we
2
2
based on extensive ab initio calculations and are leading to a better representation of the
conformational space of the anion.[72] In detail, we performed NpT molecular dynamics
simulation using Gromacs 5.0.6 [73-77] at temperatures of 300 K, 320 K, 340 K, 360 K, 380
K, and 400 K and a pressure of p = 1 bar, respectively. The ILs were represented by a cubic
simulation box containing 512 ion pairs. The box was first equilibrated for 2 ns at T = 500 K
employing the Berendsen thermostat as well as the Berendsen barostat [78] with coupling
times of 
temperature followed. Production runs of 100 ns length utilizing the Nosé-Hoover thermostat
79,80] with  = 1 ps and Rahman-Parrinello barostat [81,82] with  = 2 ps were performed
T
= 
p
= 0.5 ps. After that, another equlibration run for 2 ns at the desired
[
T
p
for each temperature. All simulations were done with a 2.0 fs time step employing periodic
boundary conditions and the LINCS algorithm [83] for fixed bond lengths. The smooth
particle mesh Ewald summation [84] was applied with a mesh spacing of 0.12 nm, a real
8
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
space cutoff 0.9 nm and 4th order interpolation. The relative accuracy of the Ewald sum was
−
5
−1
set to 10 corresponding to a convergence factor  = 3.38 nm .
The force-field of the [NTf
]− anion has been published in reference [71,85]. The pyridinium
2
force fields were derived from the OPLS force field for pyridine from Jorgensen et al.[86, 87]
The dihedral potentials of the hydroxyalkyl chain were fitted on ab initio calculations
employing second order Møller-Plesset perturbation theory using the cc-pvtz basis set. The
point charges were derived from the electrostatic potential according to the CHelpG
scheme.[88] The Lennard-Jones parameters for the cations can be found in Table 1, the point
charges are given in Scheme II. Further details on the simulations are given in the SI.
We performed MD simulations for [HOC
2
Py][NTf
2
] and [HOC
3
Py][NTf ] between 300 and
2
4
00 K to show the temperature-dependence of the cluster populations. Although we obtained
the cluster populations from NMR (183 K) and ND (303 K) measurements only at single
temperatures, and those from MD simulations only above room temperature, we can clearly
conclude: a) longer hydroxyalkyl chain lengths significantly enhance the formation of (c-c)
cationic clusters; b) for longer alkyl chain lengths the polarizability of the cation is less
important; c) the temperature dependence of the (c-c) cluster formation in the liquid phase
between 300 and 400 K can be described by van’t Hoff plots. The ratios for the (c–c) and the
(c–a) hydrogen-bonded species versus the inverse temperature obtained from MD simulations
-1
-
data result in transition enthalpies from (c–c) to (c–a) of about 31.24 kJmol (II), 9.42 kJmol
1
-1
(
IV), and 3.75 kJmol (for [HOC
4
Py][NTf
2
]), respectively. The smaller transition enthalpy
suggests that cationic clusters already exist at room temperature.
Crystallization or supercooling – (c-c) cluster formation prevents crystallization
The differential scanning calorimetry (DSC) traces of ILs I-IV as shown in Fig. 5
strongly support the interpretation of the NMR spectra at low temperatures (see also SI). The
DSC measurements were carried out by using a Mettler Toledo DSC 822e with a Huber
TC100MT cooler under N atmosphere. The samples for DSC measurements were tightly
2
sealed in Al pans of 40 l volume. All handling operations with sample were carried out in a
glove-box under a nitrogen atmosphere (residual concentrations of oxygen 1.0 ppm and water
0
.3 ppm). Pans and samples were weighted with Sartorius MSE3.6P-000-DM microbalances
-6
with the standard uncertainty of 5·10 g. The calibration of Mettler Toledo DSC 822e was
checked with melting behavior of the reference samples of indium, n-octane, and double-
9
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
distilled water. The temperature of fusion agreed with recommended value better than 0.3 K
-1
and the fusion enthalpy within 0.2 kJ·mol .
Thermograms were recorded during cooling (373-193 K) and heating (193-373 K) at
-1
cooling and heating rates of 1, 5, and 10 K min . The glass transition temperature (T
point of the heat capacity change), crystallization temperature (T ), and melting temperature
fus) were determined from DSC thermograms during the heating scans. The summary of
phase transitions is given in Table 2.
During cooling from 373 to 193ꢀK at 5 and 10ꢀK·min cooling rates, only a heat capacity
change corresponding to glass transitions (T ) could be observed in the DSC profiles of ILs II
200 K), III (206 K) and IV (200 K). The supercooled state of the (c-c) cluster-forming ILs is
obviously fairly stable. In contrast, the phase transition behavior is complex for IL I,
including melting (Tfus=276.2 K) and solid/solid phase transitions (Tss= 266.2 K and Tss
51.6 K). For the crystalline state of IL I at 183 K, we observed only one Pake pattern,
g
, middle
c
(T
−
1
g
(
=
2
indicating pure (c-a) hydrogen bonding. The strong formation of cationic clusters in ILs II,
III and IV results in supercooling and glass transition. From the combined NMR and DSC
experiments, we have clear evidence that the formation of cationic clusters prevents the ILs
from crystallization and liquid/solid phase transition. The resulting material is a glass.[89]
Our findings suggest that the phase behavior of this type of ILs can be controlled by cationic
cluster formation.
Conclusion
We measured DQCCs of hydroxyl-functionalized ionic liquids in the crystalline and glassy
states. We observed two Pake patterns for deuterons involved in normal (c-a) Coulomb-
enhanced hydrogen bonds, and in unusual (c-c) Coulomb-weakened hydrogen bonds between
cations. The DQCCs in the (c-c) cationic clusters are smaller than in (c-a) ion-pairs, indicating
stronger hydrogen bonding in accord with observed redshifts in IR spectra. These DQCCs are
close to values for molecular liquids but larger than those known in salt bridges of
supramolecular complexes. The (c-a) DQCCs are surprisingly large despite the additional
attractive Coulomb forces. Depending on the polarizability of the cation and the alkyl chain
length, the (c-c) clusters can be stronger populated than the (c-a) ion pairs at low
temperatures. The DSC traces clearly show that the ILs forming substantial amounts of (c-c)
clusters do not tend to crystallize and rather form glasses.
1
0
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Acknowledgement
This work has been supported by the Deutsche Forschungsgemeinschaft (DFG) within the
projects LU 506/14-1 and LU 506/14-2. R. L. also thanks the University of Rostock for a
Baltic Mare Fellowship provided for D.I.K. D.I.K., A.G.S. and A.E.K. acknowledge financial
support by the Ministry of Science and Higher Education of the Russian Federation (budget
Project no. AAAA-A17-117041710084-2 for Boreskov Institute of Catalysis).
1
1
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Literature
[
1]
Ionic Interactions in Natural and Synthetic Macromolecules (A. Ciferri and A. Perico,
Eds), 2012 John Wiley & Sons, Inc., p. 35 ff.
[
[
[
2]
3]
4]
F. Biedermann, H. J. Schneider, Chem. Rev. 2016, 116, 5216-300.
H. J. Schneider, Angew. Chem. Int. Ed. 2009, 48, 3924-77.
M. Ikeda, Y. Tanaka, T. Hasegawa, Y. Furusho, E. Yashima, J. Am.
Chem. Soc. 2006, 128, 6806–6807.
[
[
5]
6]
B. Kuberski, A. Szumna, Chem. Commun. 2009, 15, 1959–1961.
L. E. R. O’Leary, J. A. Fallas, E. L. Bakota, M. K. Kang, J. D. Hartgerink,
Nat. Chem. 2011, 3, 821–828.
[
[
[
7]
8]
9]
S. Kumar, R. Nussinov, J. Mol. Biol. 1999, 293, 1241-1255.
S. Kumar, R. Nussinov, ChemBioChem. 2002, 3, 604-17.
L. Cruz, B. Urbanc, J. M. Borreguero, N. D. Lazo, D. B. Teplow, and H. E.
Stanley, Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 18258-18263.
[
[
10] A. Moroni and G. Thiel, Nat. Chem. Biol. 2006, 2, 572-573.
11] J. M. Christie, A. S. Arvai, K. J. Baxter, M. Heilmann, A. J. Pratt, A. O’Hara,
S. M. Kelly, M. Hothorn, B. O. Smith, K. Hitomi, G. I. Jenkins, and E. D.
Getzoff, Science 2012, 335, 1492-1496.
[
[
[
12] P. M. J. Szell, S. Zablotny, D. L. Bryce, Nat. Comm. 2019, 10, 916.
13] X. Shi, C. M. Rienstra, J. Am. Chem. Soc. 2016, 138, 4105-4119.
14] R. Manríquez, F. A. López-Dellamary, J. Frydel, T. Emmler, H. Breitzke, G.
Buntkowsky, H.-H. Limbach, I. G. Shenderovich, J. Phys. Chem. 2009, 113, 934-940.
15] H. Weingärtner, Angew. Chem. Int. Ed. 2008, 47, 654-670.
[
[
[
16] N. V. Plechkova, K. R. Seddon, Chem. Soc. Rev. 2008, 37, 123-150.
17] K. Fumino, S. Reimann, R. Ludwig, Phys. Chem. Chem. Phys., 2014, 40, 21903-
2
1929.
[
[
[
18] P. A. Hunt, C. R. Ashworth, R. P. Matthews, Chem. Soc. Rev. 2015, 44, 1257-88.
19] R. Hayes, G. Warr, R. Atkin, Chem. Rev. 2015, 115, 6357-6426.
20] C. R. Ashworth, R. P. Matthews, T. Welton, P. A. Hunt, Phys. Chem. Chem. Phys.
2
016, 18, 18145-60.
[
[
21] S. Grimme, W. Hujo, B. Kirchner, Phys. Chem. Chem. Phys. 2012, 14 , 4875-4883.
22] E. I. Izgorodina, Z. L. Seeger, D. L. A. Scarborough, S. Y. S. Tan, Chem. Rev. 2017,
1
17, 6696-6754.
1
2
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
[
23] K. Fumino, V. Fossog, P. Stange, D. Paschek, R. Hempelmann, R. Ludwig, Angew.
Chem. Int. Ed. 2015, 54, 2792-2795; Angew. Chem. 2015, 127, 2834-2837.
24] R. Ludwig, Phys. Chem. Chem. Phys. 2015, 17, 13790-13793.
[
[
25] D. H. Zaitsau, V. N. Emel'yanenko, P. Stange, C. Schick, S. P. Verevkin, R. Ludwig,
Angew. Chem. Int. Ed. 2016, 55, 11682-11686; Angew. Chem. 2016, 128, 11856-
1
1860.
[
[
26] S. A. Katsyuba, M. V. Vener, E. E. Zvereva, Z. F. Fei, R. Scopelliti, G. Laurenczy, N.
Yan, E. Paunescu, P. J. Dyson, J. Phys. Chem. B 2013, 117, 9094-9105.
27] M. Fakhraee, B. Zandkarimi, H. Salari, M. R. Gholami, J. Phys. Chem. B 2014, 118,
1
4410-14428.
[
[
[
28] S. K. Panja, B. Haddad, J. Kiefer, ChemPhysChem 2018, 19, 3061-3068.
29] A. Knorr, R. Ludwig, Sci. Rep. 2015, 5, 17505.
30] A. Strate, T. Niemann, D. Michalik, R. Ludwig, Angew. Chem., Int. Ed. 2017, 56,
4
96-500; Angew. Chem. 2017, 129, 510 –514.
[
[
[
31] A. Strate, T. Niemann, R. Ludwig, Phys. Chem. Chem. Phys. 2017, 19, 18854-18862
32] T. Niemann, D. Zaitsau, A. Strate, A. Villinger, R. Ludwig, Sci. Rep. 2018, 8, 14753.
33] F. S. Menges, H. J. Zeng, P. J. Kelleher, O. Gorlova, M. A. Johnson, T. Niemann, A.
Strate, R. Ludwig, J. Phys. Chem. Lett. 2018, 9, 2979-2984.
[
[
[
[
[
34] T. Niemann, A. Strate, R. Ludwig, H. J. Zeng, F. S. Menges, M. A. Johnson, Angew.
Chem., Int. Ed. 2018, 57, 15364-15368; Angew. Chem. 2018, 130, 15590-15594.
35] A. Strate, V. Overbeck, V. Lehde, J. Neumann, A. M. Bonsa, T. Niemann, D. Paschek,
D. Michalik, R. Ludwig, Phys. Chem. Chem. Phys. 2018, 20, 5617-5625.
36] T. Niemann, A. Strate, R. Ludwig, H. J. Zeng, F. Menges, M. A. Johnson, Phys.
Chem. Chem. Phys. 2019, 21, 18092-18098.
37] A. Knorr, P. Stange, K. Fumino, F. Weinhold, R. Ludwig, ChemPhysChem 2016, 17,
4
58-462.
38] T. Niemann, J. Neumann, P. Stange, S. Gärtner, T. G. A. Youngs, D. Paschek, R.
Atkin, R. Ludwig, Angew. Chem. Int. Ed. 2019, 58, 12887-12892; Angew. Chem.
2
019, 131, 13019-13024.
[
[
[
[
39] J. G. Powles, J. H. Strange, Proc. Phys. Soc., London 1963, 82, 6-15.
40] G. E. Pake, J. Chem. Phys. 1948, 16, 327-336.
41] A. Abragam, 1961, The Principles of Nuclear Magnetism (Oxford Clarendon Press)
42] T.C. Farrar, E. D. Becker, Pulse and Fourier Transform NMR. Introduction to Theory
and Methods (Academic Press: New York and London, 1971.
1
3
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
[
[
43] J. W. Clymer, J. L. Ragle, J. Chem. Phys. 1982, 77, 4366-4373.
44] J. M. Jackman, J. C. Trewella, R. C. Haddon, J. Am. Chem. Soc. 1980, 102, 2519-
525.
2
[
[
45] H. Huber, J. Chem. Phys. 1985, 83, 4591-4598.
46] R. Eggenberger, S. Gerber, H. Huber, D. Searles, M. Welker, J. Chem. Phys. 1992, 97,
5
898-5904.
47] P. L. Cummins, G. B. Bacskay, N. S. Hush, B. Halle, S. Engström, J. Chem. Phys.
985, 82, 2002-2013.
[
1
[
[
48] D. T. Emonds, A. L. Mackay, J. Magn. Res. 1975, 20, 515-519.
49] D. Lankhorst, J. Schriever, J. C. Leyte, Ber. Bunsenges. Phys. Chem. 1982, 86, 215-
2
21.
[
[
[
[
[
[
[
[
50] R. Ludwig, F. Weinhold, T. C. Farrar, J. Chem. Phys. 1995, 103, 6941-6950.
51] R. Ludwig, Chem. Phys. 1995, 195, 329-337.
52] R. Ludwig, D. S. Gill, M. D. Zeidler, Z. Naturforsch. 1991, 46a, 89-94.
53] R. Ludwig and M.D. Zeidler, Mol. Phys. 1994, 82, 313-323.
54] R. Ludwig, M. D. Zeidler, T. C. Farrar, Z. Phys. Chem. 1995, 189, 19-27.
55] M. A. Wendt, T. C. Farrar, Mol. Phys. 1998, 95, 1077-1081.
56] T. D. Ferris, T. C. Farrar, Mol. Phys. 2002, 100, 303-309.
57] A. M. Nishchenko, D. I. Kolokolov, A. G. Stepanov, J. Phys. Chem. 2011, 115, 7428-
7
438.
[
[
58] A. E. Khudozhikov, P. Stange, B. Golub, D. Paschek, A. G. Stepanov, D. I. Kolokov,
R. Ludwig, Angew. Chem. Int. Ed. 2017, 56, 14310-14314: Angew. Chem. 2017, 129,
1
4500-14505.
59] Gaussian 09, Revision A.02, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, G. A. Petersson, H.
Nakatsuji, X. Li, M. Caricato, A. Marenich, J. Bloino, B. G. Janesko, R. Gomperts, B.
Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov, J. L. Sonnenberg, D.
Williams-Young, F. Ding, F. Lipparini, F. Egidi, J. Goings, B. Peng, A. Petrone, T.
Henderson, D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, W. Liang,
M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y.
Honda, O. Kitao, H. Nakai, T. Vreven, K. Throssell, J. A. Montgomery, Jr., J. E.
Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N.
Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C.
Burant, S. S. Iyengar, J. Tomasi, M. Cossi, J. M. Millam, M. Klene, C. Adamo, R.
1
4
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Cammi, J. W. Ochterski, R. L. Martin, K. Morokuma, O. Farkas, J. B. Foresman, and
D. J. Fox, Gaussian, Inc., Wallingford CT, 2016.
[
[
60] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132, 154104.
61] S. Ehrlich, J. Moellmann, W. Reckien, T. Bredow, S. Grimme, ChemPhysChem. 2011,
1
2, 3414-3420.
62] S. Grimme, A. Jansen, J. G. Brandenburg, C. Bannwarth, Chem. Rev. 2016, 116,
105-5154.
[
[
5
63] NBO 6.0. E. D. Glendening, J, K. Badenhoop, A. E. Reed, J. E. Carpenter, J. A.
Bohmann, C. M. Morales, C. R. Landis, and F. Weinhold, Theoretical Chemistry
Institute, University of Wisconsin, Madison (2013).
[
[
[
64] F.Weinhold, C. R. Landis, Valency and Bonding A Natural Bond Orbital Donor-
Acceptor Perspective, Cambridge, University Press, Cambridge, 2005.
65] A. Knorr, K. Fumino, A.-M. Bonsa, R. Ludwig, Phys. Chem. Chem. Phys. 2015, 17,
3
0978-30982.
66] T. Niemann, P. Stange, A. Strate, R. Ludwig, Phys. Chem. Chem. Phys. 2019, 21,
215-8220.
8
[
[
67] R. Ludwig, Phys. Chem. Chem. Phys. 2002, 4, 5481-5487.
68] K. M. Murdoch, T. D. Ferris, J. C. Wright, T. C. Farrar, J. Chem. Phys. 2002, 116,
5
717.
[
[
69] R. Ludwig, ChemPhysChem 2005, 6, 1369-1375
70] D. J. Searles, H. Huber, “Accurate Determination of Nuclear Quadrupole Coupling
Constants and other NMR Parameters in Liquids from the Combination of Molecular
Dynamics Simulations and ab initio Calculations”, in Encyclopedia of Nuclear
Magnetic Resonance, Volume 9, D. M. Grant and R. K. Harris (Eds.), Wiley, New
York, 2002.
[
[
[
[
71] T. Köddermann, D. Paschek, R. Ludwig, ChemPhysChem 2007, 8, 2464-2470.
72] D. Paschek, B. Golub, R. Ludwig, Phys. Chem. Chem. Phys. 2015, 17, 8431-8440.
73] E. Lindahl, B. Hess, D. van der Spoel, J. Mol. Mod. 2001, 7, 306-317.
74] H. J. C. Berendsen, D. van der Spoel, R. van Drunen, Comp. Phys. Comm. 1995, 91,
4
3-46.
75] B. Hess, C. Kutzer, D. van der Spoel, E. Lindahl, J. Chem. Theory Comput. 2008, 4,
35-447.
[
[
4
76] S. Pronk, S. Pall, R. Schulz, P. Larsson, P. Bjelkmar, R. Apostolov, M. R.
1
5
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Shirts, J. C. Smith, P. M. Kasson, D. van der Spoel, B. Hess, E. Lindahl,
Bioinformatics 2013, 29, 845-54.
[
77] M. Abraham, B. Hess, D. van der Spoel, E. Lindahl, E. Apol, R. Apostolov, H. J. C.
Berendsen, A. van Buuren, P. Bjelkmar, R. van Drunen, A. Feenstr
S. Fritsch, G. Groenhof, C. Jungshans, J. Hub, P. Kasson, C. Kutzner, B.
Lambeth, P. Larsson, J. A. Lemkul, E. Marklund, P. Meulenho_, T. Murtola,
S. Pall, S. Pronk, R. Schulz, M. Shirts, A. Sijbers, P. Tieleman, C. Wennberg,
M. Wolf, Gromacs 5.0.6, 2015.
[
78] H. J. C. Berendsen, J. P. M. Postma, W. F. van Gunsteren, A. DiNola, J.
R. Haak, J. Chem.Phys. 1984, 81, 3684-3690.
[
[
[
[
[
79] S. Nosé, Mol. Phys. 1984, 52, 255-268.
80] W. G. Hoover, Phys. Rev. A 1985, 31, 1695-1697.
81] M. Parrinello, A. Rahman, J. Appl. Phys. 1981, 52, 7182-7180.
82] S. Nosé, M. L. Klein, Mol. Phys. 1983, 50, 1055-1076.
83] B. Hess, H. Bekker, H. J. C. Berendsen, J. G. E. M. Fraaije, J. Comp. Chem. 1997, 18,
1
463-1472.
[
[
[
84] U. Essmann, L. Perera, M. L. Berkowitz, T. A. Darden, H. Lee, L. G. Pedersen, J.
Chem. Phys. 1995, 103, 8577-8593.
85] J. Neumann, B. Golub, L.-M. Odebrecht, R. Ludwig, D. Paschek, J. Chem. Phys.
2
018, 148, 193828.
86] W. L. Jorgensen, D. S. Maxwell, J. Tirado-Rives, J. Am. Chem. Soc. 1996, 118,
1225-11236.
1
[
[
[
87] W. L. Jorgensen, N. A. McDonald, J. Mol. Struc. THEOCHEM 1998, 424, 145-155.
88] C. M. Breneman, K. B. Wiberg, J. Comput. Chem. 1990, 11, 361-373.
89] Debennedetti, P. G., Stillinger, F. H., Supercooled liquids and the glass transition.
Nature 410, 259-267 (2001).
1
6
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Table 1. Lennard-Jones parameters  and  for all interaction sites of the [HOC
Py]+,
2
+
+
[
HOC
3
Py] and [HOC Py] cations. The assignment of the atoms is shown in Scheme
4
II.
-
1
site
N
 / Å
3.25
3.55
2.42
3.50
2.50
2.50
3.12
0.00
·k
B
/ K
85.55
35.23
15.10
33.20
15.10
15.10
85.60
0.00
Ca
Ha
Cc
Hc
Hm
O
Ho
Table 2. The thermodynamic parameters of observed phase transitions for ILs I-IV.
o
-1
Ionic liquid
Phase transition
] crIII-crII
T
trs, K
 H , kJ·mol
trs
m
[
(
HOC
2
MPip][NTf
2
251.9 ± 1.3
266.7 ± 1.5
276.6 ± 0.9
200.4 ± 0.1
2.7 ± 0.3
4.2 ± 0.2
4.4 ± 0.3
-
I)
crII-crI
crI – liquid
glass - liquid
[
(
[
(
HEPy][NTf ]
2
II)
-
1
HOC
3
MPip][NTf
2
] glass-liquid
204.4 ± 0.3 (1 K·min )
-
-
-1
III)
205.5 ± 0.1 (5 K·min )
-
1
2
06.8 ± 0.1 (10 K·min )
-1
[
(
HOC3Py][NTf2]
glass-liquid
197.6 ± 0.1 (1 K·min )
-1
IV)
199.8 ± 0.1 (5 K·min )
-1
2
00.6 ± 0.1 (10 K·min )
1
7
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Scheme I: a) 1-(3-hydroxyalkyl)piperidinium and b) 1-(3-hydroxyalkyl)pyridinium cations as
present in the ILs [HOC Pip][NTf ] with n=2,3 (I, III) and [HOC Py][NTf ] with n=2,3 (II,
n
2
n
2
IV). The two cations differ in polarizability. Moreover, we study (c-c) cluster formation by
variation of the hydroxyalkyl chain length.
1
8
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
+
Scheme II: Structure of the 1-(3-hydroxyethyl)pyridinium [HOC
and corresponding point charges q/e (red).
2
Py] cation with atom types
1
9
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Figure 1a.
2H NMR spectrum at 183 K and line shape analysis for ILs I
HOC MPip][NTf ]: (a) measured spectra, (b) simulated spectra. Q denotes the
[
2
2
DQCCs,  the asymmetry parameters, and p the relative populations (fractions)
of the (c-a) and (c-c) bound species. Index number 1 refers to (c-c), index
number 2 to (c-a) bound species.
2
0
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Figure 1b.
2H NMR spectrum at 183 K and line shape analysis for ILs II
[
HOC
2
Py][NTf
2
]: (a) measured spectra, (b) simulated spectra, (c) deconvoluted
(c-a) spectra and (d) deconvoluted (c-c) spectra. The designations are given in
the caption of Figure 1a.
2
1
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Figure 1c.
2H NMR spectrum at 183 K and line shape analysis for ILs III
HOC MPip][NTf ]: (a) measured spectra, (b) simulated spectra, (c)
[
3
2
deconvoluted (c-a) spectra and (d) deconvoluted (c-c) spectra. The designations
are given in the caption of Figure 1a.
2
2
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Figure 1d. 2H NMR spectrum at 183 K and line shape analysis for ILs IV
[
HOC
3
Py][NTf
2
]: (a) measured spectra, (b) simulated spectra, (c) deconvoluted
(c-a) spectra and (d) deconvoluted (c-c) spectra. The designations are given in
the caption of Figure 1a.
2
3
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Figure 2. a) Measured and deconvoluted DQCCs for ILs I-IV. The DQCCs for the (c-a)
hydrogen bonds above 210 kHz are given as open symbols, and the (c-c) hydrogen
bonds below 210 kHz as filled symbols (with n=2,3). b) We compare the DQCCs
for the (c-c) hydrogen bonds with those for water, methanol, ethanol and tert-
butanol in the liquid and solid phases as well as in phenols.[49-57] The DQCCs for
the (c-a) hydrogen bonds are related to those of salt bridges in proteins.[13, 43-44]
2
4
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
(2)
Figure 3.
NBO second order stabilization energies E n*, (red circles, right y-axes)
and estimated total charge transfers qCT (blue squares, left y-axes) for (c-c)
clusters n=2-4 plotted versus calculated differences  ((c-c)–(c-a)). The
D
almost linear dependency indicates the strong relation between NBO
stabilization energies and charge transfers with spectroscopic descriptor
DQCC. The NBO parameters in particular describe the different H-bond
strength in both species. The measured 
D
((c-c)–(c-a)) between (c-a) and (c-
c) suggests that the glassy (c-c) cluster populations consist of trimers/tetramers.
2
5
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Figure 4. Analysis of (c-a) and (c-c) cluster populations in [HOC
n
Py][NTf ] with n=2-4
2
from MD simulations for the liquid phase between 300 and 400 K. The filled
symbols show the cluster distribution obtained from the crystalline and glassy state
NMR at 183 K (this study) and neutron diffraction (ND) experiments at 303
K.[38]
2
6
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
Figure 5. The DSC traces for the ILs I-IV: For IL I we observe liquid/solid and several
solid/solid phase transitions.[33] For the crystalline state at 183 K we found a
single Pake pattern indicating (c-a) hydrogen bonded ion pairs only (see Figs. 1).
For ILs II-IV we observe supercooling and finally glass transition below 200 K.
These ILs exhibit substantial cationic cluster formation characterized by two
distinguished Pake patterns. Obviously, cationic cluster formation prevents the ILs
from crystallizing.
2
7
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201912476
TOC
Sensitive probes of like-charge attraction: Solid-state NMR spectroscopy discriminates
hydrogen bonding in conventional ion pairs from that between ions of like charge in ionic
liquids. The measured deuteron quadrupole coupling constants for both types of ‘doubly ionic
hydrogen bonds’ differ significantly. The observed cationic clusters prevent ionic liquids from
crystallizing and result in supercooling and glass formation.
2
8
This article is protected by copyright. All rights reserved.