Molecular design principles of ionic liquids with a sulfonyl fluoride moiety

Article abstract of DOI:10.1039/d0nj05603k

The continued success of ionic liquids in applications ranging from energy to medicine poses the challenge to rapidly find new functional ionic liquids with desirable properties while developing practical, scalable syntheses. As a SuFExable functionality, the sulfonyl fluoride has become widely adopted throughout the field of chemical biology due, in part, to its unique stability-reactivity pattern, highlighting the underappreciated potential of the S^VI-F motif in materials chemistry. For the first time, we herein report the development of a set of sulfonyl fluoride-functionalized ionic liquids with considerable structural diversityviaan efficient, modular, and orthogonal fluorosulfonylethylation procedure. The resulting SO₂F-functionalized ionic milieu has properties consistent with its classification as ionic liquids. We employed a combination of molecular design, synthesis, computational modeling, and X-ray crystallographic studies to gain in-depth understanding of their structure-property correlations. The diversification of the SO₂F-bearing salts is extended to include active pharmaceutical precursors, allowing for access to functional materials witha priorilow toxicity.

Full text of DOI:10.1039/d0nj05603k

NJC
View Article Online
View Journal | View Issue
PAPER
Molecular design principles of ionic liquids with a
sulfonyl fluoride moiety†
Cite this: New J. Chem., 2021,
45, 2443
a
a
a
a
David J. Siegel,‡ Grace I. Anderson,‡ Noah Cyr, Daniel S. Lambrecht,
b
c
a
Matthias Zeller, Patrick C. Hillesheim * and Arsalan Mirjafari
*
The continued success of ionic liquids in applications ranging from energy to medicine poses the
challenge to rapidly find new functional ionic liquids with desirable properties while developing practical,
scalable syntheses. As a SuFExable functionality, the sulfonyl fluoride has become widely adopted
throughout the field of chemical biology due, in part, to its unique stability–reactivity pattern, highlight-
VI
ing the underappreciated potential of the S –F motif in materials chemistry. For the first time, we herein
report the development of a set of sulfonyl fluoride-functionalized ionic liquids with considerable
structural diversity via an efficient, modular, and orthogonal fluorosulfonylethylation procedure. The
Received 16th November 2020,
Accepted 15th January 2021
2
resulting SO F-functionalized ionic milieu has properties consistent with its classification as ionic liquids.
We employed a combination of molecular design, synthesis, computational modeling, and X-ray
crystallographic studies to gain in-depth understanding of their structure–property correlations. The
DOI: 10.1039/d0nj05603k
2
diversification of the SO F-bearing salts is extended to include active pharmaceutical precursors, allowing
rsc.li/njc
for access to functional materials with a priori low toxicity.
the paradoxical combination of high kinetic barriers and robust
thermodynamic driving forces. Breaking the strong S –F bond
Introduction
VI
1,2
+
+
ꢀ
As a key reagent for Sulfur(VI) Fluoride Exchange (SuFEx), the requires the assistance of H , DBU, R
3
Si or bifluoride (FHF )
under rigorous spatial and kinetic restrictions. Con-
covalent probe sequently, unlike its chloride analog, the –SO F unit is pro-
2
F functionality serves as a foundly stable in acidic, basic, and physiological conditions.
2
,13–16
sulfonyl fluoride group (–SO
2
F) has recently reemerged as a salts
3–5
clickable motif in medicinal chemistry,
2
6
7–9
design and drug discovery. The SO
modular –SO – connector, joining the desired reactive moieties to Further, sulfonyl fluorides are highly resistant toward
2
VI
the scaffold of interest, allowing for facile synthesis of larger reduction due to the S –F bond cleavage being exclusively
molecular compounds. The unique reactivity of the –SO F func- heterolytic in contrast to sulfonyl chloride. The unique properties
2
tional group allows for selective and irreversible enzyme inhibition. and reactivity of the –SO
For example, the simplest structure of this family, methanesulfonyl the insightful comments of K. B. Sharpless: ‘‘[SO
2
F group is perhaps best summarized by
F] thereby attains
2
fluoride (MSF), was tested for the effective palliative treatment of a hallmark of click reaction function, remaining’’invisible’’ under most
2
Alzheimer’s disease and was found to act as a potent inhibitor of conditions and coming to life only when desired...
acetylcholinesterase AChE. Furthermore, PyFluor is an efficient
10
11
7
Originally discovered by B. R. Baker, the desirable electro-
and inexpensive deoxyfluorination reagent (Fig. 1). Moreover, the philicity of sulfonyl fluorides enables site-specific targeting
relative non-reactive nature of sulfonyl fluorides was utilized to under various chemical and biological contexts, mitigating
18
12
develop F-radiolabeled tracer compounds.
The unique reactivity of sulfonyl fluorides with nucleophiles
(e.g. amino acid residues of protein binding sites) hinges upon
a
Department of Chemistry and Physics, Florida Gulf Coast University, Fort Myers,
Florida 33965, USA. E-mail: amirjafari@fgcu.edu
b
c
Department of Chemistry, Purdue University, West Lafayette, Indiana 47907, USA
Department of Chemistry and Physics, Ave Maria University, Ave Maria,
Florida 34142, USA. E-mail: patrick.hillesheim@avemaria.edu
†
Electronic supplementary information (ESI) available. CCDC 2032771–2032774.
For ESI and crystallographic data in CIF or other electronic format see DOI:
1
0.1039/d0nj05603k
Fig. 1 Examples of sulfonyl fluorides as potent covalent enzyme inhibitors,
‡
These authors contributed equally to this work.
and as the Doyle’s deoxyfluorinating reagent.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 2443--2452 | 2443

View Article Online
Paper
NJC
off-target damage and toxicity while retaining suitable aqueous a point for structural-core diversification via Menshutkin reac-
physiological stability. The hydrogen bonding (HB)-mediated tion to access a diverse array of ILs. (ii) In each case, we employed
ꢀ
reactivity-switch-on mechanism makes the reactivity of the fluoride the non-coordinating bistriflimide anion (Tf N ) which is a
2
leaving group sensitive to the microenvironment of the binding widely used fluorous anion in IL formulations. Our decision is
sites with the proper local HBs. As indicated by Sharpless, grounded in the anion’s intrinsic thermal, electrochemical, and
‘
‘. . .protein surfaces provide molecular information that sulfonyl hydrolytic stability along with its aptness to form sub-ambient
2
fluorides are particularly adept at reading.’’ Accordingly, the melting ion pairs. In fact, its high level of charge delocalization
ability of sulfonyl fluorides to form covalent bonds with enzymes and conformational flexibility would allow us to invariably create
is highly selective compared to other electrophiles–which is the salts with melting points as low as might be achievable. (iii) The
key reason for their popularity in biological contexts.
2
SO F group acts both as a traditional functional group and a
Despite the wealth of literature on sulfur(VI) oxyfluorides, ‘‘SuFExable’’ site for further diversification (i.e. through the
their development as scaffolds for functional materials remains one-step conversion to sulphonamides, sulfonates, sulfonic
17,18
limited to few reports.
For instance, in 2020, Xue et al. acids, or sulfonyl-based dicationic ILs). This feature might be
developed the ‘‘full fluorosulfonyl’’ electrolyte for high-voltage useful to covalently immobilize biomolecules (e.g. enzymes) on
2
3
rechargeable lithium-metal batteries (LMBs) by dissolving lithium ILs for the potential development of biosensors.
18
bis(fluorosulfonyl) imide (LiFSI) in Me
2
N–SO
2
F solvent. Shortly
The ease and generality of the fluorosulfonylethylation
afterward, Shen et al. discovered that using Pefabloc improves strategy was established via the preparation of a series of
19
performance and stability of perovskite solar cells.
2
ethyl-linked SO F-functionalized ILs, illustrated in Fig. 2. We
Over 15 years ago, ‘‘task-specific’’ ionic liquids (TSILs) were first employed a variety of common imidazolium, pyridinium, 1,2,4-
introduced by J. H. Davis as a conceptual framework for the triazolium(aromatic), and pyrrolidinium, morpholinium (non-
controlled assembly of functional molecular units into an ionic aromatic) cations. Further core-diversification was accomplished
20
liquid (IL) structure, achieving characteristic chemical behaviors.
using electron-deficient (3c, 3d and 3i) and electron-rich (3g
These materials were identified as ILs where a functional group is
covalently tethered to the cation and/or anion. Such ionic com-
pounds invariably have a combination of diverse physical and
chemical properties. Their inherent tunability enables the design
of new functional materials, while retaining the core, desired
features of conventional ILs, viz., nonvolatility, high ionic and
thermal conductivity, and excellent electrochemical and thermal
21
stability. A fundamental obstacle that synthetic IL chemists face
is particularly clear: although ILs have moderate structural complex-
ity, transferring synthetic organic reactions to the ionic environment
remains a bottleneck due to the challenges associated with their
purification (i.e. distillation, crystallization and chromatographic
separation are impractical). The incorporation of new design
elements (e.g. click chemistry) has led to recent advances towards
22
novel functional IL systems. However, the paradigm in developing
TSILs remains the incorporation of new functionalities into IL
structures with great reliability as part of a linear – effective albeit
slow, approach for the discovery of new functional organic materials.
2
It is noteworthy that until now, IL–SO F systems were absent
from the TSILs catalog. Essentially, the combination of ILs’
architectural platform and relatively sluggish reactivity of SO F –
2
which we regard as an advantage rather than a liability – underpin
our enthusiasm to explore the development of SO F-bearing ILs as
2
a emergent class of practically useful and fundamentally interesting
materials. To address the task above, we herein report the first
systematic study on synthesis and empirical structure–property
relationships of this new family of functional ILs.
Results and discussion
Synthesis
Fig. 2 Two-step synthesis (top) and structures (bottom) of novel ILs 3
constructed through the fluorosulfonylethylation of diverse heterocyclic
ꢀ
headgroups 1. The counter-ion of each of these cations is Tf
2
N
that is
Three structural elements are introduced into our design:
omitted for clarity. Reactions occurred at ambient temperature except 3c,
(
i) the modular incorporation of various heterocyclic cations as 3d and 3i required reflux condition.
2444 | New J. Chem., 2021, 45, 2443--2452
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021

View Article Online
NJC
Paper
and 3h) heterocycles. Consistent with this proposal, incorpora-
tion of electron-withdrawing substitutions on the imidazolium
and pyridinium rings did not lead to attenuated reactivity.
Notably, the presence of various functional groups on the
above-mentioned aromatic rings gives the adopted synthetic
procedure an orthogonal dynamic, allowing for wide applicability
of the proposed synthetic routes.
The diversification of the SO
include active pharmaceutical precursors, allowing for access
to ILs with a priori low toxicity. In turn, we selected Scheme 1 Common (dashed box) and employed (grey box) synthetic
-nitroimidazole and 4-aminopyridine as the cations for the routes to ethyl-linked sulfonyl fluorides.
compounds 3c and 3g. 5-Nitroimidazole is the backbone of
2
F-bearing salts is extended to
5
2
4,25
several antimicrobial compounds
and 4-aminopyridine
ꢀ
1
s
addition. However, the reagent’s cost ($54.60 g for 95% purity)
and poor shelf stability render it unsuitable for our chemistry.
Strikingly, our synthetic objectives are met by using 2 (Scheme 1)
in lieu of ESF. Our decision was based upon two considerations:
forms the core for dalfampridine (Ampyra ) which is used
for the treatment of multiple sclerosis.
2
6
The IL products 3a–k were prepared via a two-step procedure
based upon the N-alkylation of nitrogen-containing hetero-
cycles 1 with 2-chloroethanesulfonyl fluoride 2, followed by
anion metathesis and combining the requisite chloride salts
(i) ESF is an extremely toxic substance (orally and intraperitoneally
to laboratory animals§) and is a severe lachrymator that must be
used under specific safety guidance; and (ii) avoid sacrificing the
extraordinary efficiency of the synthesis by adding an extra step,
i.e., MgO-mediated elimination with 73% yield. Described by
with KNTf in water (Fig. 2). The formation of the IL products
2
1
13
19
was confirmed by H, C and F NMR, MS, and by single
crystal X-ray diffraction (SC-XRD). Further elaboration on synth-
esis and characterization of the IL products and NMR spectra
can be found in the ESI.† Of particular note with respect to the
29
Sharpless et al., 2 is readily made via the on-water sulfonyl
chloride–fluoride exchange (halex process) in 100% yield, using its
chloride congener and saturated KHF solution (Scheme 1).
2
1
9
NMR data is the presence of a peak at ca. ꢀ80 ppm in the
F
spectrum of each IL, consistent with presence of an aliphatic
sulfonyl fluoride group. With the exception of compounds 3c
and 3d, which required elevated temperature conditions perhaps
due to the presence of the deactivating nitro group, reactions
were completed within few minutes at ambient temperature,
indicated by the precipitation or phase separation of the products.
For instance, the chloride counterparts of 3a, 3g, and 3j instantly
formed precipitates upon mixing of equimolar amounts of the
corresponding starting materials in ethyl acetate (see Fig. S1 in the
ESI,† to see a video about the instantaneous formation of an
onium chloride salt). Strikingly, purification was rarely required
for each step, and the resultant IL were harvested by simple
concentration and filtration or phase separation. We found that
the IL products are soluble in hot water and methanol but not in
methylene chloride and EtOAc (noted as useful purification
solvents to separate the ILs from their precursors). Yields of both
the intermediate chloride salts and the final products are quanti-
tative. As anticipated, the ILs exhibit robust thermodynamic
stability toward hydrolysis and thermolysis, making them reliable
materials of choice for a diverse set of environmental conditions.
All compounds were handled and stored on the benchtop for over
Thermophysical properties
Thermophysical analysis of these ILs was carried out using
thermogravimetric analysis (TGA) and differential scanning
calorimetry (DSC) and the data is summarized in Table 1.
Evaluating the thermal stability, the TGA data of 3a–k sub-
stantially varied between new ILs, exhibiting an initial decom-
position step between 159.2–434.6 1C where B5% mass loss is
reached (Tonset5%). Presumably, this indicates that the decom-
position proceeds initially through the cleavage of the C–SO2F
bond (defluorosulfonylation). The thermal stability of the
PyFluor-derived IL (3i) is noticeably higher than others with a
DTonset5% between structurally related compounds 3f and 3i of
284.3 1C. We attribute this dramatic increase in stability to the
interactions between to sulfonyl fluoride groups on the cation
and the anion. It should be noted that thermal stability of these
ILs can be modulated via incorporation of a protic functional
group readily leading to formation strong HB networks. Examining
the DTonset5% for the related series of compounds 3g (301.8 1C), 3f
(186.3 1C), and 3h (231.6 1C) incorporation of HBs in 3g signifi-
cantly increases the thermal stability as compared to the structures
without the amine groups. It should be noted, however, that HBs is
only one factor of many which can influence the properties of ILs.
Evaluating the experimental values of melting points (Tm)
for the ILs, all fall below the customary benchmark of 100 1C
used classify a salt as an IL. Compounds 3b, 3d, 3f, 3j, and 3k are
27
6
0 days and no elimination products (i.e. sulfenes) were detected
30
1
13
by H and C NMR spectroscopy. Due to the facile nature of the
procedure and the accessibility of starting materials, the ILs can
be synthesized in multigram scale – e.g., 3a was scaled up to ca.
20 g in 94% yield.
m
crystalline white solids with T values above room-temperature
While several synthetic routes to ethyl-linked sulfonyl fluorides
were developed, a common method used in the present work is
ꢀ
1
ꢀ1
§
The oral LD50 is approximately 50 mg kg for rats and approximately 10 mg kg
that shown in Scheme 1. Ethenesulfonyl fluoride (ESF) – ‘‘the most
for mice. The intraperitoneal LD50 is 1–5 mg kg for rats and o5 mg kgꢀ1 for
mice. The liquid was absorbed through the intact skin, and the skin absorption
ꢀ1
2
,28
perfect Michael acceptor ever found’’
SuFEx reagent to incorporate the SO
– is the most powerful
F via the conjugate (Michael) LD50 is 1–5 mL kgꢀ
.
1 28
2
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 2443--2452 | 2445

View Article Online
Paper
NJC
Table 1 The melting point (T
thermal decomposition temperature (Tonset5%) values of the IL products
m
), glass transition temperature (T
g
), and result in a slight increase in T
of 3d (DT
= 3.0 1C) in comparison
m
m
to 3b.
To further provide insight toward into structure–property
relationships of the new ILs, the dynamic and kinetic viscosity,
and density of a representative IL (3a) was measured as func-
tion of temperature at 1 bar and are shown in Fig. 4 and Table
S1 (see ESI†). As anticipated, the incorporation of sulfonyl
fluoride functionality has great influence on the viscosity of
3a relative to a corresponding IL with all-carbon and the same
IL (1C) (1C) onset5% (1C)
T
m
T
g
T
3
3
3
3
3
3
3
3
3
3
3
a
b
c
d
e
f
g
h
i
—
59.6
—
66.4
—
79.5
—
—
—
48.9
93.1
ꢀ4.7
—
237.2
184.9
200.4
159.2
211.9
186.3
301.8
231.6
434.6
229.8
266.8
ꢀ15.7
—
ꢀ30.2
—
—
ꢀ30.4
18.6
—
2
number of side-chain backbone atoms, i.e., [emim][NTf ] has a
3
2
Z value of 34.0 mPa s compared to 3a with Z value of 1081.9
mPa s s at 20 1C. It is consistent with increased inter-particle
interaction brought upon by the grafting of the powerfully
associating functional groups. For many ILs, viscosity is a
strong function of temperature near room temperature, as is
the case for representative IL, with viscosity varying over ca. 27
orders of magnitude in the temperature range studied.
j
k
—
(Fig. 3). All other compounds remain as free-flowing liquids at
room temperature. It is noted that the T peaks (onset of an
m
exothermic peak upon heating) was only observed in the first
DSC run presumably due to the high viscosity of the ILs and the
tendency to form amorphous glasses instead of ordered crystalline X-ray crystallography
ꢀ
1
solids even at slower cooling rate (2 1C min ). Generally, as is
often the case for ILs, recrystallization on a practical timescale is
not always achievable due to their dramatically increasing
viscosities and the existence of supercooled states, resulting
To better understand the structural impact incorporation of the
SO F group has on the compounds, we conducted crystallo-
graphic analysis allowing for direct insight into the spatial
relationship between the cations and the anion as well as
2
2
1
in formation of viscous glassy liquids. The remaining pro-
ducts exhibited no well-defined transitions from ordered solids
to isotropic liquids, albeit weak, low-temperature glass transi-
tions (T ; midpoint of a broad endothermic peak upon heating)
were observed and these varied slightly on their thermal history
Table 1). This behavior is likely due to, in part, the symmetry of
the cation frustrating solid-phase packing, resulting in the
formation of a matrix that ossifies into an amorphous solid
at low temperatures. Predictably, we note that the presence of
SO F has the effect of increasing the T values. For instance, 3b
has DTm of ca. 34.6 1C, relative to the common [edmim][Tf N]
edmim = 1-ethyl-2,3-dimethylimidazolium), which is likely
due to the presence of inter- and intra-ionic interactions arising
from the SO F group as observed in the solid-state structures.
Interestingly, a comparison of the T values of 3b and 3d indicates
that the paradoxical effects of the high polarity of NO and
reduced p interactions (due to its electron-withdrawing nature)
enhanced understanding of their physicochemical properties.
ꢀ
Typically, due to the conformational freedom of the Tf
2
N anion,
ILs with this anion display complex crystallization behavior and
g
33
growing diffraction quality crystals is very challenging. How-
ever, we succeeded in forming suitable crystals of 3b, 3c and 3f
for SC-XRD study and their asymmetric units are depicted in
Fig. 2 and 5. The presence of the SO F group results in inter-
2
(
molecular interactions that encourages crystallization perhaps
due to its high polarity. Despite considerable efforts, unfortu-
nately, we were unable to produce the crystals of other IL products
of suitable quality. In turn, we isolated a crystal from the cognate
2
1
2
m
2
3
1
(
ꢀ
of 3a, paired with BETI (bis[(pentafluoroethyl)sulfonyl]imide)
anion, confirming the formation of the corresponding SO F-
2
2
bearing imidazolium cation (Fig. 5). It should be noted that we
m
ꢀ
prepared the BETI analog of 3a because we had ready access to a
2
stock of LiBETI, a relatively hard-to-acquire material.
Fig. 3 DSC thermograms for compounds 3b (red), 3d (blue), 3f (green),
and 3k (purple). Measurements were taken in open pans under N gas at
2
ꢀ1
ramp rate of 5 1C min . The thermograms have been offset for clarity, but
Fig. 4 The dynamic (blue) and kinetic (teal) viscosity, and density (purple)
of IL 3a.
not rescaled.
2446 | New J. Chem., 2021, 45, 2443--2452
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021

View Article Online
NJC
Paper
Fig. 5 Top row: the asymmetric units of 3b, 3c, 3f and 3a – BETI salts. Middle row: dnorm mapped on the Hirshfeld surfaces. Bottom row: Fingerprint
plots.
Further, to seek greater insight into intermolecular interac- fingerprint plots for the crystal structures are shown in Fig. 5
tions within a crystal structures, we concluded that a solid-state and Table S2 (see ESI†) contains the relative percentages of
Hirshfeld surface analysis of these new materials would be interactions obtained from the fingerprint plots. In an effort to
useful. Several important insights emerge from the obtained evaluate the impact the SO F groups have on the solid-state
2
data as depicted in Fig. 5. Hirshfeld surface analysis is performed structure, discussion presented herein focuses predominantly on
via the CrystalExplorer. Hirshfeld surface analysis is a ‘‘whole analysis of the cationic portions of the structures only.
molecule approach’’ to examining the intermolecular interactions
Crystallographic data for IL 3b. 3b crystals crystallize in the
34,35
and packing mode within a crystal structure.
In brief, the P1% space group with a single ion pair in the asymmetric unit
Hirshfeld surface mapped with dnorm allows for facile visualization (see ESI,† Table S3). Both the cation and anion are disordered
of interactions outside (d ) and within (d ) the boundaries of the (solvable) over two positions with the larger component of the
surface. The exact definitions of d , d , and dnorm are provided disorder accounting for approximately 90% occupancy. The
within the relevant literature and are defined, briefly, herein.
e
i
i
e
34,36,37
major portion of the disorder was isolated and used for surface
The color coding on the Hirshfeld surface, that is blue to white to analysis. Examining the fingerprint plot of the cation, one can
red, indicate relative distances of interactions, with blue indicating see the expected peaks arising from Hꢁ ꢁ ꢁO|Oꢁ ꢁ ꢁH reciprocal
i e
distances longer the sum of the van der Waals radii and red being H-bonding at d E 0.9, d E 1.3, as well as at the reciprocal
3
7
less than the sum of the radii. These interactions can be decon- distances. The shortest of these interactions are between the
structed into two dimensional plots of d and d known as hydrogens on the ring, namely C4–H, and the sulfonyl oxygens
i
e
37
fingerprints. These fingerprint plots not only allow for qualitative on the anion (d = 2.3469(17) Å) followed by the Oꢁ ꢁ ꢁH–C
comparisons based on shape and structural features of the plot, interactions (d = 2.38–3.56 Å) with symmetry adjacent anions.
but can also be broken down to a ‘per atom’ basis allowing for
granular, percentage-based comparisons of interactions arising the oxygens and fluorine atoms in the SO
A number of cation–cation interactions arising from both
F group can be seen
2
from distinct sets of atoms within a structure. The fingerprint as the red spots on the Hirshfeld surface. As seen with com-
plots and Hirshfeld surfaces, along with a traditional approach to pounds 3a-BETI and 3c (vide infra) the sulfonyl oxygen atoms
crystallographic analysis, allows for a wealth of information to be are linking cations together through Hꢁ ꢁ ꢁOꢁ ꢁ ꢁH chains with
extracted from the structure of compounds. Particularly, they interactions ranging from 2.4783(12) Å to 3.1029(11) Å. How-
provide direct insight into the spatial relationships between ever, the cation-cation interactions in 3b are distinct. While
the cation and anion and structural features of the ILs, providing there are the Hꢁ ꢁ ꢁOꢁ ꢁ ꢁH chains linking the cations, the sulfonyl
a strong basis to further understand both short range (e.g. oxygens are also bridging with the p cloud of the adjacent
3
8
39,40
H-bonding) and coulombic interactions. The surfaces and cations, interacting specifically with the nitrogen in the cation.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 2443--2452 | 2447

View Article Online
Paper
NJC
This interaction is seen as a ‘pocket’ in the surface and is high- individual cations in the asymmetric unit display unique finger-
lighted in Fig. 5. prints (see ESI,† Fig. S2). The discussion herein will pertain only
The fluorine atom of the SO F also contributes to several to cation C, however it can be extended to some degree to all of
2
cation–cation interactions, the shortest of which are between the cations.
the fluorine and several hydrogens on both the ring and
The pyridinium-based IL offers a chance to examine inter-
ethylene chain. The closest interactions are with the C4–H on actions of the sulfonyl fluoride group with a non-imidazolium
ꢀ
the imidazolium ring at distances of 2.7688(10) Å, forming a based set of aromatic hydrogens. The Tf
2
N anion shows strong
sort of dimeric, paired cation motif. These interactions are interactions with the a hydrogens, that is C1–H and C5–H, of
unique from those observed in the other compounds as the the pyridinium ring. Further, the anion also has the shortest
fluorine appears to only form the dimers rather than link interactions with the methylene hydrogens on the alkyl chain at
cations together in a chain-type structure.
distances of 2.35 Å to 3.01 Å. These interactions are seen as the
Crystallographic data for IL 3c. 3c crystallizes in the P2 /c bright red spots near the relevant hydrogens on the Hirshfeld
1
space group with a single cation–anion pair in the asymmetric surface in Fig. 5. The sulfonyl fluoride oxygens on the cation
unit with no disorder observed in either moiety (see ESI,† Table show longer-distance interactions with the b hydrogens (C2–H
S4). With the inclusion of the nitro group on the imidazole & C4–H) and g hydrogens (C3–H) at distances of approximately
cation one can see two bright red spots near one of the nitrogen 3.0 Å. These interactions exist between several of the crystal-
atoms in the heterocycle indicating an interaction with the p lographically unique cations within the asymmetric unit. The
system of the ring (circled in Fig. 5). These interactions are seen shortest Oꢁ ꢁ ꢁH interaction, however, is at 2.500(2) Å between
in the fingerprint plot as the wing-type structure near d E 1.9, cation C and cation B and is seen as the H-bonding spike at di
i
4
1,42
de E 1.5.
While all the other cations do exhibit these p-type E 1.4, d E 1.1. As is observed with the imidazolium-based
e
interactions, to a varying degree, 3c is the only one to display a cations, the SO F moiety bridges aromatic hydrogens in the
2
defined feature in the plot. Further, examining Table S2 (ESI†) typical Hꢁ ꢁ ꢁOꢁ ꢁ ꢁH chain manner.
shows that 3c has the lowest percentage of Hꢁ ꢁ ꢁH interactions
2
As seen by the red spot near the fluorine atom in the SO F
of the three imidazolium compounds examined, despite having group, the fluorine atom has several close contacts with hydro-
only one less hydrogen when compared to 3a-BETI (vide infra). gens on adjacent cations. Specifically, the fluorine is interact-
The decreased percentage of Hꢁ ꢁ ꢁH interactions is easily seen ing with the g hydrogen on cation A at a distance of 2.635(2) Å.
when contrasting the fingerprint plots of 3b versus 3c (Fig. 5). These interactions are also seen by a set of Hꢁ ꢁ ꢁF|Fꢁ ꢁ ꢁH peaks
The sulfonyl fluoride interactions in IL 3c are also unique noted in the fingerprint plot.
ꢀ
when compared to the other imidazolium-based structures.
Specifically, while the fluorine atom is acting as a bridge compound (T
between cations as observed in 3a, the points of interaction crystallizes in the P2
Crystallographic data for BETI analog of 3a. The 3a-BETI
m
= 39.6 1C; T
g
= ꢀ12.8 1C; Tonset5% = 274.0 1C)
1
/c space group with a single cation–anion
ꢀ
are unique with Fꢁ ꢁ ꢁH–C6 (d = 2.589(2) Å) and Fꢁ ꢁ ꢁH–C7 (d = pair in the asymmetric unit (see ESI,† Table S6). The BETI
3.153(2) Å). Additionally, the oxygen atoms are seen bridging anion is disordered over multiple points by rotation. The
cations through interactions with the C6–H methyl group and Hirshfeld surface for the cation was calculated with all portions
the C7–H methylene groups in similar manners discussed with of the disordered anion as part of the structure, likely giving
compounds 3a and 3b. The distances of these interactions are rise to some of the disperse sets of interaction points seen
also comparable, ranging from about 2.53–2.68 Å.
along the boundaries of the fingerprint plot. The fingerprint
As expected, however, the percentage of interactions origi- plot for the cation shows a number of defining features, which
nating from oxygen is the highest of all the compounds are shared among all of the cations examined, including a
examined. This is also noticeable when looking at the differ- number of H-bonding spikes and a wing-type structure char-
3
7
ence in shape of the H-bonding spikes from the fingerprints. acteristic of interactions with the p cloud of the heretocycle.
While ILs 3b and 3c share some similarities, 3c is unique in Interactions with the methylene and methyl hydrogens on the
several ways. For example, the Hꢁ ꢁ ꢁO|Oꢁ ꢁ ꢁH peaks in 3c are ancillary chains comprise the majority of the HꢁꢁꢁO interactions
much wider than in the other structures, accounting for the seen in the spikes at d E 1.0, d E 1.2–1.3. These interactions are
i
e
increased percent contribution of oxygen to the overall interac- clearly seen as the red spots on the surfaces surrounding the
tions. Further, looking at the fingerprints as a whole, the appropriate regions on the cation. As indicated in Fig. 5, there are
interactions in 3c are more dense, especially at the longer two distinct HꢁꢁꢁO peaks, one larger diffuse peak in blue and one
i e
d /d values. Disperse interactions are theorized to indicate sharper peak in green. The larger blue peak arises from interac-
4
2
suboptimal crystal packing but are a common feature in tions between the methylene hydrogens and the disordered
4
1
fingerprints of ILs with the Tf
2
N anion.
sulfonyl oxygens of the anion. Additional interactions are from
Crystallographic data for IL 3f. 3f crystallizes in the P1% space close-contacts with a non-disordered portion of the anion. Overall,
group with a four cation–anion pairs in the asymmetric unit the HꢁꢁꢁO interactions from the methylene hydrogens to the
0
ꢀ
(
Z = 4, Z = 8) (see ESI,† Table S5). While structures of ILs have sulfonyl oxygens on the BETI anion range from approximately
0
43,44
been previously reported with similar Z values,
reported herein is the first example of a pyridinium IL with Z = 4
containing the Tf
the structure 2.36 Å to 2.70 Å, bearing in mind the disordered parts of the anion.
0
2
Focusing specifically on interactions arising from the SO F
anion. Following on this, all four of the moiety on the cation, there are Oꢁ ꢁ ꢁH and Fꢁ ꢁ ꢁH interactions
ꢀ
2
N
2448 | New J. Chem., 2021, 45, 2443--2452
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021

View Article Online
NJC
Paper
observed at d
i
E 1.4, d
e
E 1.2. As detailed in Table S2 (ESI†), polarization consistent basis set family was designed to enable
interactions arising from the sulfonyl oxygens constitute 19.7% systematic improvement of energetics for self-consistent field
47
of the total interactions while the fluorine atom contributes to approaches, similar in spirit to the correlation-consistent basis
49
9.0%. Both oxygens, that is O1 and O2, are interacting with set family for correlation energies.
distinct, symmetry adjacent cations. The closest interactions
Initial geometries for structure prediction were obtained
are between the oxygen atoms and the hydrogens on C4–H, from the crystal structures reported in this work. Our predic-
C5–H, and C6–H at distances ranging from 2.56 Å to 2.72 Å. The tions included one cation–anion pair excised from the crystal
sulfonyl oxygens act as bridges, linking multiple cationic units structure of 3b either treated in vacuum for exploratory calcula-
together through Hꢁ ꢁ ꢁOꢁ ꢁ ꢁH chains. The fluorine is also inter- tions and comparison or using the ‘‘conductor-like’’ polarizable
5
0,51
acting with the hydrogens on the methyl group (C6–H), acting continuum model (C-PCM) to incorporate solvent effects.
as a bridge between cations via the methyl group C6–H at Since this work is the first to present results on 3b, a range of
distances ranging from 2.5991(14) Å up to 3.0630(15) Å. These dielectric constants for use in the C-PCM treatment was tested
Fꢁ ꢁ ꢁH interactions are another common trait observed in the to assess the impact on predicted properties. Based on published
5
2
cations and are manifested by the blunted H-bonding spikes results for loosely related ILs, three dielectric constants were
indicated in Fig. 5.
r
included to bracket the estimated constant for the system (e = 6,
Overall, examining all of the structures presented herein, 12, 24). Starting with the initial geometries and solvent treatment
Oꢁ ꢁ ꢁH and Fꢁ ꢁ ꢁH interactions from the sulfonyl fluoride moiety as just described, molecular geometries were optimized at the
add a significant quantity of cation–cation interactions that is oB97X-D/pc-1 level of theory. Subsequent property calculations,
not present in simple alkyl-based ILs. The percentages listed in such as dipole moments, polarizabilities, and electrostatic
Table S2 (ESI†) show that inclusion of SO F moieties added potential maps were calculated at the oB97X-D/pc-2 level. All
2
approximately 20% intermolecular interactions arising from oB97X-D/pc-1 and oB97X-D/pc-2 calculations were performed
5
3
the oxygen and about 8% of interactions from the fluorine. with a development version of the Q-Chem program package.
Given that the shortest, and thus the strongest of these inter- Dielectric constants were predicted at the BP86/def2-TZVP level
5
4,55
actions are due to hydrogen atoms, the increased crystalline of theory
nature of the SO F-bearing ILs can likely be attributed to the Krossing group. These calculations were performed both
increased hydrogen interactions between cations. It should be within Q-Chem and with the Orca quantum chemistry program
following a computational approach proposed by
5
6
2
5
7,58
noted, however, that not all the interactions may be favorable as packages.
both the fluorine and oxygen atoms exhibit potentially unfavorable
The computationally predicted geometries are in good
agreement with the experimentally obtained crystal structures
45
OꢁꢁꢁO, OꢁꢁꢁF, FꢁꢁꢁF interactions in all crystals examined.
Several bridging motifs are observed in the four structures (Table 2). For example, using the shortest cation–anion distance
arising from both the oxygen and fluorine atoms in the SO as a representative geometric measure, the distance between the
moiety. Interactions were observed between aromatic and N–Me on the imidazolium ring and the nitrogen of the NTf
2
F
2
aliphatic hydrogen atoms. Further, these interactions appear anion, computational predictions produce between 3.013 and
to be competitive with the anion. Specifically, the donor atoms 3.079 Å compared to 3.345 Å in the experiment. Likewise, other
on the cation –SO F group and the anion are very similar. geometric parameters agree well with the experimental findings
2
Granted, the anionic SQO groups will have a higher negative (see ESI,† Tables S7–S12). This agreement between experiment
charge thus making them more likely HB acceptors. Further, and theory is encouraging, as it suggests sufficiently accurate
several O/Fꢁ ꢁ ꢁp interactions arising from the cations were also modelling at least of short-range interactions and confirms
seen arising from the SO
those observed arising from the SO
2
F moiety, which are again similar to experimental assignments.
ꢀ
2
groups on the Tf
2
N anions.
The predicted dipole moments range between 16.7 and 18.2
Debye (excluding the vacuum result) and depend on the C-PCM
dielectric constant used. As expected, embedding the ion pair in
Computational approach
In tandem with X-ray analysis, since several computational a more polar dielectric environment leads to an increase in
approaches were performed to study (and predict) the ILs’ predicted dipole moment due to stabilization of polarization.
properties as both function of the cation structure and the
anion identity, it seemed reasonable to do so with one of the
synthesized ILs. Accordingly, density functional theory predic-
tions were performed to confirm and interpret experimental
findings and to predict properties of 3b. To this end, the
Table 2 Computationally predicted properties for 3b depending on the
solvent dielectric constant employed in the C-PCM treatment. R denotes
the shortest cation/anion distance. q(e) denotes the charge on the cation
taken as the sum of the Mulliken charges. Deviations from the integer (+1)
oB97X-D dispersion-correct range-separated functional was
employed in combination with the polarization consistent cation–anion pair
pc) family of basis sets of double- and triple-zeta quality (pc-1
charge provide an estimate of the extent of charge transfer between the
(
3
Solvent e
r
R (Å) Total dipole (Debye) Polarizability (Bohr ) q (e)
46,47
and pc-2, respectively).
geometries across a broad range of benchmarks, including for
example non-covalent interactions which are reproduced typically
oB97X-D yields accurate energetics and
1
6
1
(vacuum) 3.013 13.7
198.9
198.3
198.3
198.6
0.9621
3.013 16.7
3.013 17.1
3.079 18.2
0.8935
0.8936
0.9042
2
ꢀ
1 48
within a root mean square deviation of 1–2 kcal mol
.
The 24
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 2443--2452 | 2449

View Article Online
Paper
NJC
We also note that due to differences in our software we modified
the approach compared to the one presented by the Krossing
5
6
group. Specifically, we used the C-PCM solvation model and
6
2
not COSMO as in the originally proposed method and used
in-house Python code to determine molecular volumes utilizing
the Mendeleev package¶ to import atomic radii. Moreover,
solvation energies were obtained for cation–anion pairs com-
bined. Due to the differences between our computational
approaches, some of the calculated parameters differed from
5
6
those of Krossing and co-workers. However, we found that
scaling of the parameters led to good agreement with the
Fig. 6 Electrostatic potential map. Electrostatic potentials were calcu- published results. Specifically, we scaled our molecular volumes
5
4,59,60
lated at the B3LYP/pc-2 level of theory
on a 300 ꢂ 240 ꢂ 200 grid
and solvation energies, respectively (see ESI,† Tables S13–S15).
with approx. 20 grid points per Angstrom. Plots were created in the VMD
61
With these adjustments, our computational approach provided
graphical interface. Shown is the ESP on an electron density isosurface
with contour value 0.1.
5
6
an excellent reproduction of the published results for a
combined training/test set of seven ILs with a root mean square
deviation of 1.3 units in the dielectric constant (Table S16, ESI†).
Using this approach, we obtained a dielectric constant of
On the other hand, the predicted polarizabilities do not show a
significant dependence on the polarity of the environment. The 26.8 for 3b, which places 3b in the upper range for dielectric
overall sizable dipole moment becomes apparent from the constants among reported values because of the presence of the
electrostatic potential (ESP) plot showing a pronounced positive sulfonyl fluoride group. We note that it is not yet clear whether
potential over the imidazolium ring of the cation and pronounced these predictions based on an empirical analysis between
negative potential regions distributed over the anion (Fig. 6). computational descriptors and experimental dielectric constants
A noteworthy exemption is that the SO F moiety on the cation are transferable to 3b structure and electrostatic properties.
2
also shows negative electrostatic potential, as expected based on However, at the very least we believe the result further empha-
the electronegativity of the atoms involved. A region of pro- sizes the unique properties of 3b.
nounced negative ESP on the cation of an IL is unusual and can
be expected to lead to short-ranged intermolecular interactions
not typically found with the cations of an IL.
Conclusions
The predicted charges for the cation range between 0.8935
Fundamental to the field of ILs is the understanding that
and 0.9042 electron charges depending on the dielectric constant
structure begets function. For the first time, we developed a
used for the C-PCM (again omitting the vacuum results as a
series of novel task-specific ILs containing a sulfonyl fluoride
special case). This result corresponds to a transfer of electron
moiety, as a thermally stable and chemical robust functionality.
density between the cation and anion ranging between 0.0958
2
Generally, the SO F functionality is much more thermodynamically
and 0.1065 electrons depending on the polarity of the solvent
environment. While Mulliken charges are basis set dependent
and should not be interpreted as absolute numbers, the magni-
tude of deviation from the integer (+1) charge suggests that there
is a considerable transfer of electron density between the cation
and anion and indicates an electronic interaction between the
cation and anion that goes beyond pure electrostatics.
stable than other sulfonyl halides toward hydrolysis, reduction,
2
and thermolysis. Along with its biocompatibility, these features
incentivize the investigation and development of sulfonyl fluorides
and allude to the reasoning behind their successful inclusion in
drug development applications. In this work, eleven structurally
diverse IL-SO F systems, including three containing biologically
2
active cores were synthesized in gram-quantity scale. Using active
pharmaceutical ingredients (APIs) with established toxicological
We have predicted dielectric constants of 3b based on the
5
6
computational approach presented by the Krossing et al.,
2
profiles and the biocompatible SO F group may serve to develop
which is based on a fit of computationally predicted descriptors
to experimental dielectric constants for a range of ILs. Specifi-
cally, the dielectric constant was calculated from the equation:
ILs with reduced toxicity. The products can be simply scaled-up
without sacrificing their yields or purity. To aid defining their
practical range of use for various applications, the thermophysical
properties of new ILs, including melting points, glass transition
temperatures, thermal stability, density, and temperature-
dependent viscosity were reported.
Aside from providing a practical route to ILs with useful
properties, the exceptionally facile synthetic procedure allowed
us to produce the ILs with great reliability through a consistent,
well-controlled pathway. Simple reaction conditions and non-
Vm
V0
DGs
G0
er ¼ a þ b
þ c
where a = ꢀ97.8, b = 51.4, c = 0.206 were determined by Krossing
5
6
and co-workers.
V
m
is the molecular volume obtained from
3
density functional theory prediction and V = 1 nm . DG is the
0
s
solvation energy again obtained from the solvent model and
ꢀ
1
G
0
= 1 kJ mol . Note that the sign of constant c is flipped
compared to the definition of Krossing et al., which was
necessary to reproduce the previously published results.
¶
L. M. Mentel, mendeleev – A Python resource for properties of chemical elements,
5
6
ions and isotopes, available at: https://github.com/lmmentel/mendeleev.
2
450 | New J. Chem., 2021, 45, 2443--2452
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021

View Article Online
NJC
Paper
chromatographic purification procedures were employed with 10 D. E. Moss, R. G. Fariello, J. Sahlmann, I. Sumaya, F. Pericle
quantitative yields throughout. Despite the common practice of
and E. Braglia, Br. J. Clin. Pharmacol., 2013, 75, 1231–1239.
using ESF as a highly reactive Michael acceptor to prepare ethyl- 11 M. K. Nielsen, C. R. Ugaz, W. Li and A. G. Doyle, J. Am. Chem.
linked sulfonyl fluorides, we chose an alternative method,
Soc., 2015, 137, 9571–9574.
allowing efficient and rapid access to diverse SO F-bearing ILs 12 J. A. H. Inkster, K. Liu, S. Ait-Mohand, P. Schaffer, B. Gu ´e rin,
2
and bypassing the higher costs and toxicity associated with ESF
T. J. Ruth and T. Storr, Chem. – Eur. J., 2012, 18, 11079–11087.
as a raw material. The synthetic technique presented herein is a 13 B. Gao, L. Zhang, Q. Zheng, F. Zhou, L. M. Klivansky, J. Lu,
facile approach for creating
functionazlied low-melting salts and is integral to the progres-
a
wide variety of SO
2
F-
Y. Liu, J. Dong, P. Wu and K. B. Sharpless, Nat. Chem., 2017,
9, 1083–1088.
sion of the rapidly expanding field. And, since we have recently 14 T. Hmissa, X. Zhang, N. R. Dhumal, G. J. McManus,
begun to elucidate the types of modifications that can be made
X. Zhou, H. B. Nulwala and A. Mirjafari, Angew. Chem., Int.
Ed., 2018, 57, 16005–16009.
to SO F–IL systems without compromising their melting point,
2
we expect in our future work to exploit this knowledge to 15 T. Guo, G. Meng, X. Zhan, Q. Yang, T. Ma, L. Xu,
prepare variants of the sulfur(VI) oxyfluoride ILs.
K. B. Sharpless and J. Dong, Angew. Chem., Int. Ed., 2018,
7, 2605–2610.
5
16 J. Dong, K. B. Sharpless, L. Kwisnek, J. S. Oakdale and
V. V. Fokin, Angew. Chem., Int. Ed., 2014, 53, 9466–9470.
17 T. Fujigaya, Y. Sibasaki, S. Ando, S. Kishimura, M. Endo,
M. Sasago and M. Ueda, Chem. Mater., 2003, 15, 1512–1517.
18 W. Xue, Z. Shi, M. Huang, S. Feng, C. Wang, F. Wang,
J. Lopez, B. Qiao, G. Xu, W. Zhang, Y. Dong, R. Gao, Y. Shao-
Horn, J. A. Johnson and J. Li, Energy Environ. Sci., 2020, 13,
Conflicts of interest
There are no conflicts to declare.
Acknowledgements
2
12–220.
This material is based upon work supported by the National
Science Foundation under Grant No. CHE-1952846 (RUI), and
CHE-1919785 and CHE-1625543 (MRI). Acknowledgment is
made to the Donors of the American Chemical Society Petroleum
Research Fund for partial support of this research. A. M., D. J. S.
and G. I. A. are grateful for the generous supports provided by
the Brodie and Blair Foundation Scholarships, administered by
the Whitaker Center. A. M. thanks Professor Jim Davis at
University of South Alabama for generously providing a stock
of LiBETI.
1
9 C. Shen, Y. Wu, S. Zhang, T. Wu, H. Tian, W.-H. Zhu and
L. Han, Solar RRL, 2020, 4, 2000069.
2
2
0 J. J. H. Davis, Chem. Lett., 2004, 33, 1072–1077.
1 D. R. MacFarlane, M. Kar and J. M. Pringle, Fundamentals of
Ionic Liquids: From Chemistry to Applications, Wiley-VCH
Verlag GmbH & Co. KgaA, 2017.
2 A. Mirjafari, Chem. Commun., 2018, 54, 2944–2961.
3 F. Ghorbanizamani and S. Timur, Anal. Chem., 2018, 90,
2
2
6
40–648.
2
2
4 D. I. Edwards, J. Antimicrob. Chemother., 1993, 31, 9–20.
5 Y. Miyamoto, J. Kalisiak, K. Korthals, T. Lauwaet, D. Y. Cheung,
R. Lozano, E. R. Cobo, P. Upcroft, J. A. Upcroft, D. E. Berg,
F. D. Gillin, V. V. Fokin, K. B. Sharpless and L. Eckmann, Proc.
Natl. Acad. Sci. U. S. A., 2013, 110, 17564–17569.
Notes and references
1
2
W. Steinkopf, J. Prakt. Chem., 1927, 117, 1–82.
J. Dong, L. Krasnova, M. G. Finn and K. B. Sharpless, Angew. 26 S. Sedehizadeh, M. Keogh and P. Maddison, Clin. Neuro-
Chem., Int. Ed., 2014, 53, 9430–9448.
pharmacol., 2012, 35, 191–200.
A. Narayanan and L. H. Jones, Chem. Sci., 2015, 6, 27 J. F. King, Acc. Chem. Res., 1975, 8, 10–17.
3
4
2
650–2659.
28 Q. Chen, P. Mayer and H. Mayr, Angew. Chem., Int. Ed., 2016,
55, 12664–12667.
O. Fadeyi, M. D. Parikh, M. Z. Chen, R. E. Kyne Jr.,
A. P. Taylor, I. O’Doherty, S. E. Kaiser, S. Barbas, S. Niessen, 29 Q. Zheng, J. Dong and K. B. Sharpless, J. Org. Chem., 2016,
M. Shi, S. L. Weinrich, J. C. Kath, L. H. Jones and
R. P. Robinson, ChemBioChem, 2016, 17, 1925–1930.
E. C. Hett, H. Xu, K. F. Geoghegan, A. Gopalsamy,
81, 11360–11362.
30 C. Roth, T. Peppel, K. Fumino, M. K o¨ ckerling and
R. Ludwig, Angew. Chem., Int. Ed., 2010, 49, 10221–10224.
5
R. E. Kyne, C. A. Menard, A. Narayanan, M. D. Parikh, 31 R. Tao, G. Tamas, L. Xue, S. L. Simon and E. L. Quitevis,
S. Liu, L. Roberts, R. P. Robinson, M. A. Tones and
L. H. Jones, ACS Chem. Biol., 2015, 10, 1094–1098.
L. H. Jones and J. W. Kelly, RSC Med. Chem., 2020, 11, 10–17.
B. R. Baker, Acc. Chem. Res., 1969, 2, 129–136.
Z. Liu, J. Li, S. Li, G. Li, K. B. Sharpless and P. Wu, J. Am.
Chem. Soc., 2018, 140, 2919–2925.
J. Chem. Eng. Data, 2014, 59, 2717–2724.
32 M.-J. Deng, P.-Y. Chen, T.-I. Leong, I. W. Sun, J.-K. Chang
and W.-T. Tsai, Electrochem. Commun., 2008, 10, 213–216.
33 R. A. O’Brien, M. S. Zayas, S. T. Nestor, J. C. Gaitor,
L. M. Paul, F. A. Edhegard, S. Minkowicz, R. E. Sykora,
Y. Sheng, S. F. Michael, S. Isern and A. Mirjafari, New
J. Chem., 2016, 40, 7795–7803.
6
7
8
9
C. Sadlowski, B. Park, C. Araujo Borges, S. Das, D. Lucas
Kerr, M. He, H. Han, L. Riley and N. Murthy, Mol. Syst. Des. 34 M. A. Spackman and D. Jayatilaka, CrystEngComm, 2009, 11,
Eng., 2018, 3, 599–603. 19–32.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 2443--2452 | 2451

View Article Online
Paper
NJC
3
3
3
3
3
4
4
4
4
4
5 J. J. McKinnon, A. S. Mitchell and M. A. Spackman, Chem. –
Eur. J., 1998, 4, 2136–2141.
A. D. Dutoi, R. G. Edgar, S. Fatehi, L. Fusti-Molnar,
A. Ghysels, A. Golubeva-Zadorozhnaya, J. Gomes,
M. W. D. Hanson-Heine, P. H. P. Harbach, A. W. Hauser,
E. G. Hohenstein, Z. C. Holden, T.-C. Jagau, H. Ji, B. Kaduk,
K. Khistyaev, J. Kim, J. Kim, R. A. King, P. Klunzinger,
D. Kosenkov, T. Kowalczyk, C. M. Krauter, K. U. Lao,
A. D. Laurent, K. V. Lawler, S. V. Levchenko, C. Y. Lin,
F. Liu, E. Livshits, R. C. Lochan, A. Luenser, P. Manohar,
S. F. Manzer, S.-P. Mao, N. Mardirossian, A. V. Marenich,
S. A. Maurer, N. J. Mayhall, E. Neuscamman, C. M. Oana,
R. Olivares-Amaya, D. P. O’Neill, J. A. Parkhill, T. M. Perrine,
R. Peverati, A. Prociuk, D. R. Rehn, E. Rosta, N. J. Russ,
S. M. Sharada, S. Sharma, D. W. Small, A. Sodt, T. Stein,
D. St u¨ ck, Y.-C. Su, A. J. W. Thom, T. Tsuchimochi,
V. Vanovschi, L. Vogt, O. Vydrov, T. Wang, M. A. Watson,
J. Wenzel, A. White, C. F. Williams, J. Yang, S. Yeganeh,
S. R. Yost, Z.-Q. You, I. Y. Zhang, X. Zhang, Y. Zhao,
B. R. Brooks, G. K. L. Chan, D. M. Chipman, C. J. Cramer,
W. A. Goddard, M. S. Gordon, W. J. Hehre, A. Klamt,
H. F. Schaefer, M. W. Schmidt, C. D. Sherrill,
D. G. Truhlar, A. Warshel, X. Xu, A. Aspuru-Guzik, R. Baer,
A. T. Bell, N. A. Besley, J.-D. Chai, A. Dreuw, B. D. Dunietz,
T. R. Furlani, S. R. Gwaltney, C.-P. Hsu, Y. Jung, J. Kong,
D. S. Lambrecht, W. Liang, C. Ochsenfeld, V. A. Rassolov,
L. V. Slipchenko, J. E. Subotnik, T. Van Voorhis,
J. M. Herbert, A. I. Krylov, P. M. W. Gill and M. Head-
Gordon, Mol. Phys., 2015, 113, 184–215.
6 J. J. McKinnon, D. Jayatilaka and M. A. Spackman, Chem.
Commun., 2007, 3814–3816.
7 J. J. McKinnon, M. A. Spackman and A. S. Mitchell, Acta
Crystallogr., Sect. B: Struct. Sci., 2004, 60, 627–668.
8 P. M. Dean, J. M. Pringle and D. R. MacFarlane, Phys. Chem.
Chem. Phys., 2010, 12, 9144–9153.
9 S. K. Singh and A. Das, Phys. Chem. Chem. Phys., 2015, 17,
9596–9612.
0 S. M. Malathy Sony and M. N. Ponnuswamy, Cryst. Growth
Des., 2006, 6, 736–742.
1 P. M. Dean, J. M. Pringle, C. M. Forsyth, J. L. Scott and
D. R. MacFarlane, New J. Chem., 2008, 32, 2121–2126.
2 M. A. Spackman and J. J. McKinnon, CrystEngComm, 2002,
4, 378–392.
3 D. Rauber, F. Heib, M. Schmitt and R. Hempelmann, J. Mol.
Liq., 2016, 216, 246–258.
4 X. Han, J. Ke, N. Suleiman, W. Levason, D. Pugh, W. Zhang,
G. Reid, P. Licence and M. W. George, Phys. Chem. Chem.
Phys., 2016, 18, 14359–14369.
4
5 K. Reichenb ¨a cher, H. I. S u¨ ss and J. Hulliger, Chem. Soc. Rev.,
2
005, 34, 22–30.
6 J.-D. Chai and M. Head-Gordon, Phys. Chem. Chem. Phys.,
008, 10, 6615–6620.
4
2
4
4
7 F. Jensen, J. Chem. Phys., 2001, 115, 9113–9125.
8 N. Mardirossian and M. Head-Gordon, Mol. Phys., 2017,
1
15, 2315–2372.
54 A. D. Becke, Phys. Rev. A: At., Mol., Opt. Phys., 1988, 38,
3098–3100.
4
5
9 T. H. Dunning Jr., J. Chem. Phys., 1989, 90, 1007–1023.
0 T. N. Truong and E. V. Stefanovich, Chem. Phys. Lett., 1995, 55 F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005,
40, 253–260. 7, 3297–3305.
1 M. Cossi, N. Rega, G. Scalmani and V. Barone, J. Comput. 56 P. Eiden, S. Bulut, T. K o¨ chner, C. Friedrich, T. Schubert and
2
5
Chem., 2003, 24, 669–681.
2 H. Weing ¨a rtner, J. Mol. Liq., 2014, 192, 185–190.
3 Y. Shao, Z. Gan, E. Epifanovsky, A. T. B. Gilbert, M. Wormit,
I. Krossing, J. Phys. Chem. B, 2011, 115, 300–309.
57 F. Neese, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2012, 2,
73–78.
5
5
J. Kussmann, A. W. Lange, A. Behn, J. Deng, X. Feng, 58 F. Neese, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2018,
D. Ghosh, M. Goldey, P. R. Horn, L. D. Jacobson, 8, e1327.
I. Kaliman, R. Z. Khaliullin, T. Ku ´s , A. Landau, J. Liu, 59 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter
E. I. Proynov, Y. M. Rhee, R. M. Richard, M. A. Rohrdanz, Mater. Phys., 1988, 37, 785–789.
R. P. Steele, E. J. Sundstrom, H. L. Woodcock, 60 S. H. Vosko, L. Wilk and M. Nusair, Can. J. Phys., 1980, 58,
P. M. Zimmerman, D. Zuev, B. Albrecht, E. Alguire, 1200–1211.
B. Austin, G. J. O. Beran, Y. A. Bernard, E. Berquist, 61 W. Humphrey, A. Dalke and K. Schulten, J. Mol. Graphics
K. Brandhorst, K. B. Bravaya, S. T. Brown, D. Casanova, Modell., 1996, 14, 33–38.
C.-M. Chang, Y. Chen, S. H. Chien, K. D. Closser, 62 A. Klamt and G. Sch u¨ u¨ rmann, J. Chem. Soc., Perkin Trans.,
D. L. Crittenden, M. Diedenhofen, R. A. DiStasio, H. Do, 1993, 799–805.
2
452 | New J. Chem., 2021, 45, 2443--2452
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021