Polarity study of ionic liquids with the solvatochromic dye Nile Red: A QSPR approach using in silico VolSurf+ descriptors

Article abstract of DOI:10.1016/j.tet.2016.04.056

The in silico VolSurf+ descriptors, accounting for both cationic and anionic structural features of ionic liquids (ILs) were used to develop a Partial Least Squares (PLS) model able to establish a Quantitative Structure Property Relationship (QSPR) correlation with their solvatochromic dye Nile Red polarity. The PLS model allowed prediction of E_NR values for 116 ILs providing an in silico ILs polarity database.

Full text of DOI:10.1016/j.tet.2016.04.056

Accepted Manuscript
Polarity study of ionic liquids with the solvatochromic dye Nile Red: a QSPR approach
using in silico VolSurf+ descriptors
Alessio Paterno', Francesca D'Anna, Cosimo G. Fortuna, Giuseppe Musumarra
PII:
S0040-4020(16)30333-7
10.1016/j.tet.2016.04.056
TET 27699
DOI:
Reference:
To appear in: Tetrahedron
Received Date: 12 February 2016
Revised Date: 7 April 2016
Accepted Date: 21 April 2016
Please cite this article as: Paterno' A, D'Anna F, Fortuna CG, Musumarra G, Polarity study of ionic
liquids with the solvatochromic dye Nile Red: a QSPR approach using in silico VolSurf+ descriptors,
Tetrahedron (2016), doi: 10.1016/j.tet.2016.04.056.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical Abstract
0

ACCEPTED MANUSCRIPT
Polarity study of Ionic Liquids with the solvatochromic dye Nile Red: a QSPR approach using
in silico VolSurf+ descriptors
Alessio Paterno’ , Francesca D’Anna , Cosimo G. Fortuna , and Giuseppe Musumarra *
a
b
a
a
†
Electronic supplementary information (ESI) available. See DOI:
a
Dipartimento di Scienze Chimiche, Università di Catania, Viale A. Doria 6, I-95125 Catania, Italy.
Dipartimento STEBICEF, Università di Palermo, Viale delle Scienze - Parco d'Orleans II, Ed.17,
b
90128 Palermo, Italy.
Abstract
The in silico VolSurf+ descriptors, accounting for both cationic and anionic structural features of
ionic liquids (ILs). were used to develop a Partial Least Squares (PLS) model able to establish a
Quantitative Structure Property Relationship (QSPR) correlation with their solvatochromic dye Nile
Red polarity. The PLS model allowed prediction of ENR values for 116 ILs providing an in silico
ILs polarity database.
Keywords
Ionic liquids; Polarity; Nile Red; QSPR
*
gmusumarra@unict.it
1

ACCEPTED MANUSCRIPT
1.
Introduction
Ionic liquids (ILs), organic salts formed of an organic cation and an inorganic or organic anion, are
nowadays a very popular class of compounds due to their versatility in applications which, from
1
their original use as reaction media, soon expanded to many others. Their low melting point made
them widely adopted as reaction media originally considered as “green” solvents due to their low
volatility, as well as to their high chemical and thermal stability.
As reaction media, they can exert very different effects. Indeed, ILs can affect reaction mechanisms
2
,3
inducing different pathways with respect to conventional organic solvents or, alternatively, they
can exert a kinetic effect giving products with higher yields in significantly lower reaction times
4
and/or temperatures.
Given the huge number of cation/anion combinations, a large variety of solvent systems having
different properties can be obtained. This may simultaneously represent an advantage and a
detriment. Although a wide range of solvents would be in principle available for the same process,
the lack of knowledge of solvent properties and the effort required to perform experimental
investigations makes the selection of a suitable solvent for a given process or reaction a hard task.
Hence the need to predict solvent properties by Quantitative Structure Property Relationships
(QSPR) models based on in silico structural descriptors. In this context a new opportunity is
5
,6
provided by the VolSurf+ approach, which uses information coded into 3D GRID Molecular
7
-10
Interaction Fields (MIFs)
to derive physicochemical molecular descriptors such as molecular
size and shape, hydrophilic and hydrophobic properties, hydrogen bonding, amphiphilic moments,
critical packing parameters and pharmacokinetic descriptors related to solubility, metabolic stability
and cell permeability.
The VolSurf+ procedure has been successfully applied in QSAR studies to predict a wide variety of
biological properties such as structure–permeation relationships,5 antitumour activities
,
11-13
1
4,15
16
phospholipidosis induction
and more recently the aquatic toxicity of ionic liquids.
In the ionic liquids field VolSurf+ descriptors take into account several cationic ILs structural
features such as heterocyclic aromatic and non-aromatic cores, alkyl chain length, presence of
oxygen atoms in the substituents as well as physicochemical properties of various inorganic and
organic anions.
In this study we wish to verify if VolSurf+ descriptors, designed for in silico modelling of
biological activities, could be suitable for the prediction of an important ILs property such as
solvent polarity. A recent review, examining several approaches to computational chemistry used
1
7
with ILs and pointing out the utility of combining computational and experimental approaches,
2

ACCEPTED MANUSCRIPT
confirms that the proposed methodology, widely adopted in medicinal chemistry and industrial
experimental design, has not been previously applied for modelling ILs physico-chemical
properties.
Among solvent properties, polarity is one of the most widely used and different empirical scales
1
8
known. In general, both neutral and charged probes have been used, for example the Kosower and
1
9,20
Reichardt
scales adopt pyridinium betaine dyes as polarity probes, while a neutral “sensor” dye
Nile Red (NR) exhibits bathocromic shifts above 100 nm.
As far as ILs are concerned, interesting experimental studies on the determination of ILs dielectric
2
1-24
constants have been carried out by Weingaertner and coworkers.
The use of empirical scales to evaluate their polarity has been the object of an intense debate.
Indeed, the response of the probe to the solvent is determined by all possible solvent-solute
interactions, and in the case of ionic solvents media, some can play a more significant role than
others. Polarity of ILs has been determined using single-molecules as spectroscopic probes, taking
advantage of the effects they are able to exert on electronic absorption, fluorescence and vibrational
2
5
spectra, as well as on paramagnetic resonance signals.
In the case of UV-Vis active probes, both charged and neutral “sensors” have been used.
Consequently, the polarity of ILs has been evaluated using pyridinium betaine dyes, transition-
+
metal complexes such as [Cu(acac)(tmen)] , merocyanine probes and Nile Red (NR). In particular,
NR has been used not only to evaluate the polarity of ILs by means of the ENR parameter, but also
to have information about their 3-dimensional organization as neat solvents or in combination with
2
6,27
other ILs or conventional organic solvents to give binary mixtures.
More than a decade ago,
2
8
Seddon et al. reported a pioneering study on NR polarity of 1-alkyl-3-methylimidazolium ILs and
showed that their polarity is comparable with that of the lower alcohols. This paper points out “a
paucity of good reproducible data” and the need “to create a reliable database”.
2
9
Studies relating different ILs structures and structural variations in both cationic and anionic
3
0
moieties to ILs polarity using the empirical Kamlet–Taft solvent parameters have been reported.
In this context we here report the polarity of 18 ILs with different aromatic and non-aromatic
cationic scaffolds, different alkyl chain length and different anions and develop a Quantitative
Structure Property Relationships (QSPR) model based on VolSurf+ descriptors with the aim to
predict ENR values for a large number of commonly used ILs.
3

ACCEPTED MANUSCRIPT
Scheme 1. Structure of ILs and of the solvatochromic probe used.
2.
Results and Discussion
3
1-33
In Table 1 we report both literature
and experimental ENR values for 18 ILs (see Scheme 1) with
different cationic and anionic structural features: mainly imidazolium cationic scaffolds (plus one
piperidinium and one pyrrolidinium), different length alkyl side chains (up to 8 carbon atoms), and
8
anions ranging from tetrafluoroborate to oxygenated anions.
The PLS analysis was carried out using a 18x176 descriptor matrix including 18 ILs (Table 1) and
1
6
1
76 variables (128 cation and 48 anion VolSurf+ descriptors) and the E polarity as the response.
NR
2
The analysis provided a 3 PLS components model (see Table 2) explaining 93.2% of y variance (Q
st
nd
2
=
0.770) where the 1 and the 2 PLS components explain already 88.1% of y variance (Q =
.728). The plot of predicted vs. experimental data (Fig. 1) shows a good agreement between
experimental and predicted ENR values.
0
4

ACCEPTED MANUSCRIPT
Table 1. Experimental ENR values for ILs used as learning and test set and PLS predictions.
a
ENR
ENR
Kcal/mol)
PLS
pred.
b
Ionic Liquids
(Kcal/
mol)
DModX Name
(
c
1-Benzyl-3-methylimidazolium 1,1,1-trifluoro-N-
(trifluoromethyl)sulfonyl]methanesulfonamide
1-Ethyl-3-methylimidazolium 1,1,1-trifluoro-N-
(trifluoromethyl)sulfonyl]methanesulfonamide
1
2
IM1-(1Ph) (CF3SO2)2N
IM12 (CF3SO2)2N
216.2
216.3
217.8
1.54
1.07
[
d
217.9
[
e
3
4
5
IM14 (1O)2PO2
IM14 N(CN)2
IM14 PF6
218.0
217.8
215.6
0.63
0.74
0.79
1-Butyl-3-methylimidazolium dimethyl phosphate
1-Butyl-3-methylimidazolium N-cyanocyanamide
f
215.7
f
217.7
218.5 217.5
218.0 217.8
1-Butyl-3-methylimidazolium hexafluorophosphate
1-Butyl-3-methylimidazolium 1,1,1-trifluoro-N-
f
6
IM14 (CF3SO2)2N
218.2
0.85
[
(trifluoromethyl)sulfonyl]methanesulfonamide
d
7
8
9
IM14 BF4
217.3
217.2 217.2
216.9
0.86
1.01
1.57
1-Butyl-3-methylimidazolium tetrafluoroborate
1-Butyl-3-Methylimidazolium trifluoroacetate
e
IM14 CF3COO
IM14 1COO
216.6
e
217.2
217.3
1-Butyl-3-methylimidazolium acetate
d
1-Butyl-2,3-dimethylimidazolium 1,1,1-trifluoro-N-
1
1
1
1
1
1
1
0 IM14-2Me (CF3SO2)2N
1 IM16 N(CN)2
217.7
218.0
215.1
217.3
216.8
217.7
215.5
216.4
0.92
0.70
0.84
1.03
0.61
1.07
1.41
[
(trifluoromethyl)sulfonyl]methanesulfonamide
f
215.3
1-Hexyl-3-methylimidazolium dicyanamide
d
1-Hexyl-3-methylimidazolium 1,1,1-trifluoro-N-
2 IM16 (CF3SO2)2N
3 IM16 SbF6
217.9
[
(trifluoromethyl)sulfonyl]methanesulfonamide
d
216.1
1-Hexyl-3-methylimidazolium hexafluoridoantimonate
d
1-Methyl-3-octylimidazolium 1,1,1-trifluoro-N-
4 IM18 (CF3SO2)2N
5 IM18 N(CN)2
217.9
[
(trifluoromethyl)sulfonyl]methanesulfonamide
f
215.7
1-Methyl-3-octylimidazolium dicyanamide
1-Benzyl-3-butylimidazolium 1,1,1-trifluoro-N-
c
6 IM4-(1Ph) (CF3SO2)2N
216.1
[
(trifluoromethyl)sulfonyl]methanesulfonamide
1-Butyl-1-methylpiperidinium 1,1,1-trifluoro-N-
(trifluoromethyl)sulfonyl]methanesulfonamide
1-Butyl-1-methylpyrrolidinium 1,1,1-trifluoro-N-
(trifluoromethyl)sulfonyl]methanesulfonamide
d
1
1
7 Pip14 (CF3SO2)2N
8 Pyr14 (CF3SO2)2N
219.0
219.5
219.6
0.87
0.77
[
d
219.9
[
1
2
2
2
2
9 IM1-10 BF4
0 IM16 PF6
1 IM16 BF4
2 IM18 PF6
3 IM18 BF4
219.2 216.9
216.8 217.0
216.8 216.7
217.6 217.4
217.7 217.1
1.40
0.87
0.96
0.86
1.01
1-Decyl-3-methylimidazolium tetrafluoroborate
1-Hexyl-3-methylimidazolium hexafluorophosphate
1-Hexyl-3-methylimidazolium tetrafluoroborate
1-Methyl-3-octylimidazolium hexafluorophosphate
1-Methyl-3-octylimidazolium tetrafluoroborate
N1 IM14 NO3
N2 IM16 NO3
N3 IM18 NO3
215.3 212.3 35.32 1-Butyl-3-methylimidazolium nitrate
216.3 211.8 35.28 1-Hexyl-3-methylimidazolium nitrate
217.4 212.2 35.29 1-Methyl-3-octylimidazolium nitrate
a
b
34
Ref. 28; DModX values indicate the distance of the observations to the X-model hyperplane. Predictions for ILs
c
d
e
f
with DModX > 1.32 should be considered with caution; Ref. 31; Ref. 32; This work; Ref. 33.
Table 2. PLS model parameters for the 3 extracted principal components (PCs).
2
2
2
2
2
Q
2
PC
R X
R X(cum) R Y R Y(cum)
Q (cum)
1
2
3
0.295
0.234
0.128
0.295
0.529
0.657
0.756
0.126
0.051
0.756
0.881
0.932
0.603
0.315
0.153
0.603
0.728
0.770
5

ACCEPTED MANUSCRIPT
3
4
SIMCA (Soft Independent Modelling of Class Analogy) models are soft models applicable within
the experimental space covered by the observations (in the present case ILs) considered in the
analyzed data matrix. Therefore the prediction affordability for each single IL depends on its
structural similarity with the structures of the ILs considered in the model derivation. The PLS
model predictions are reported in Table 1 together with DModX values, the model residuals in the
X matrix, providing an estimate of the similarity of a new observation (IL) to the training set
observations. Predictions for ILs exhibiting DModX values higher than 1.32 should be considered
3
4
with some caution, the higher the DModX value, the lower the prediction reliability.
Figure 1. PLS correlation plot reporting the experimental and the corresponding predicted ENR
values for ILs 1-18.
In order to check the experimental reliability of the model predictions, the ENR values reported by
2
8
35
Seddon are also reported as a test set in Table 1. It has been noted that although the Kamlet–Taft
parameters are well established for traditional solvents, for ILs they are very sensitive to
36
impurities. However, in the present case, the E values used as a learning set were determined in
NR
the presence of 1,4-dioxane. As previously reported, the comparison with literature values outline
3
2
that the co-solvent presence has a negligible effect. This statement is further supported by the
2
8
agreement between the ENR values measured by us with literature ones for three ILs (Table 1)
which indicates a remarkable inter-laboratory reproducibility.
The PLS model predictions for the test set are in agreement with the experimental data in Table 1.
Predictions for objects with high DModX values, such as ILs with nitrate anions absent in the
learning set, are definitely unreliable. A possible interpretation of this result might rely on both
anion geometry and size. Anions used in this study exhibit different geometries, as well as different
-
-
ionic radii, comparable for [BF ] and [PF ] anions (232 and 240 pm, respectively) but significantly
4
6
-
lower for [NO ] (179 pm) which determines a different charge distribution. Anyway, a specific
3
6

ACCEPTED MANUSCRIPT
effect which renders nitrate modelling more difficult might be responsible for their peculiar
behaviour.
As far as 1-decyl-3-methylimidazolium tetrafluoroborate is concerned, the poor agreement between
experimental and predicted values could be ascribed to segregation effects, which heavily depend
3
7,38
on the alkyl chain length.
On the grounds of the above effects, the fluorescent probe could feel
different polar and apolar microenvironments, more evident for decylimidazolium, than for
octylimidazolium tetrafluoroborate.
The PLS model predictions were extended to ENR values for a higher number of ILs (up to 116)
exhibiting DModX values lower than 2.0 in Table S2.
In addition to its prediction potentiality, the PLS model allows us to evaluate quantitatively the
importance of the VolSurf+ descriptors (Table S1 in ESI) in determining the ENR values by means
of their VIP (Variables Importance for the Projection) values.
Table 3 reports the VIP values for the top 75 descriptors, together with the coefficients referring to
st
the 1 PLS component (the most relevant one) providing information on the sign of the descriptor
contribution.
st
Table 3. VIP values for the top 75 descriptors together with the coefficients of the 1 PLS
component. VIP values describe the importance of each X-variable both in modeling the objects
and in relation to the Y response. The coefficients values express the sign and the magnitude of the
contribution between the Xs and the Y.
VIP
values
Coeff.
Values
Var ID
1
2
3
4
5
6
7
8
9
D3_Cat
CP_Cat
1.80
1.80
1.78
1.78
1.78
1.77
1.77
1.77
1.74
1.73
1.71
1.70
1.69
1.66
1.64
1.62
1.62
1.61
1.60
-0.022
-0.022
-0.022
-0.022
-0.022
-0.022
-0.022
-0.022
-0.021
0.021
D5_Cat
VD_Cat
D4_Cat
D6_Cat
CD1_Cat
CD2_Cat
CD3_Cat
CACO2_Cat
CD4_Cat
D7_Cat
1
0
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
-0.021
-0.021
-0.021
-0.021
-0.021
-0.020
-0.020
-0.020
-0.019
D2_Cat
CD7_Cat
CD8_Cat
CW2_Cat
CD6_Cat
CD5_Cat
PSA_Cat
7

ACCEPTED MANUSCRIPT
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
D8_Cat
D1_Cat
1.60
1.58
1.58
1.58
1.55
1.55
1.54
1.53
1.52
1.51
1.48
1.47
1.45
1.39
1.35
1.35
1.35
1.34
1.33
1.32
1.32
1.32
1.32
1.30
1.29
1.26
1.25
1.24
1.24
1.21
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.19
1.18
1.16
1.15
1.11
-0.020
-0.019
-0.019
-0.019
-0.018
-0.018
-0.019
-0.018
-0.018
-0.018
-0.018
-0.017
-0.017
-0.018
-0.016
-0.016
0.016
0.016
0.015
-0.016
-0.015
-0.015
-0.015
-0.015
-0.014
-0.015
0.014
-0.012
-0.013
0.014
0.013
0.013
0.013
0.013
0.013
0.013
0.013
0.013
0.013
0.013
0.013
0.013
0.012
-0.013
-0.013
-0.011
-0.013
D8_An
CD8_An
CD6_An
CD7_An
W2_Cat
D6_An
D7_An
CD5_An
CW1_Cat
CD4_An
D5_An
W3_Cat
D4_An
CW3_An
S_An
V_An
G_An
CW3_Cat
W1_Cat
CW2_An
IW4_An
CW5_An
ID4_Cat
W5_An
R_An
DD1_Cat
FLEX_Cat
W1_An
SOLY_Cat
LgS3_Cat
LgS4_Cat
LgS5_Cat
LgS6_Cat
LgS7_Cat
LgS7,5_Cat
LgS8_Cat
LgS9_Cat
LgS10_Cat
LgS11_Cat
L0LgS_Cat
LgBB_Cat
G_Cat
DRDRDR_Cat
DD2_Cat
FLEX_RB_Cat
8

ACCEPTED MANUSCRIPT
6
6
6
7
7
7
7
7
7
7
8
9
0
1
2
3
4
5
CW4_An
DD3_Cat
A_An
1.11
1.10
1.10
1.10
1.09
1.09
1.06
1.05
1.01
-0.013
-0.009
0.013
S_Cat
-0.011
-0.011
-0.012
-0.012
-0.010
0.009
HSA_Cat
MW_Cat
CW1_An
V_Cat
DIFF_Cat
VIP values give an indication (in absolute values) of what variables in the X block (VolSurf+
descriptors) are relevant to determine the dependent variable (polarity of ILs). Table 3 shows that
the major influence on the response is provided by cation descriptors, especially by those related to
their hydrophobic character (hydrophobic volumes D and hydrophobic capacity factors CD at
n
n
different energy levels). A lower contribution is given by cationic descriptors referring to a general
hydrophilic character (hydrophilic volumes indicating polarizability W1-W4 and hydrophilic
capacity factors CW1-CW3) and water solubility (intrinsic solubility SOLY and solubility at
various pH LgSs). The latter descriptors show a positive contribution to the ENR values (as
expressed by the relative coefficients), whereas the former are inversely correlated to the responses,
indicating that molecular polarizability (expressed by W1-W3) in ILs contribute to decrease the
response, whereas ILs able to be dissolved in water have higher ENR values. Similar considerations
can be done for the corresponding anionic hydrophobic character (D4-D8 and CD4-CD8) and
hydrophilic (W1,W5 and CW1-CW5) character descriptors. In particular, a positive contribution is
found for anionic size descriptors (molecular globularity G, molecular volume V, molecular surface
S and volume/surface ratio R), in agreement with the considerations made to rationalize the
-
peculiarity of [NO ]-based ionic liquids.
3
In Fig. 2 we report the PC1 vs. PC2 scores plot displaying in the same chemical space the 18 ILs
and ENR (the y response). In the upper right quadrant ILs with a non-aromatic heterocyclic scaffold
are located very close to ENR showing that they positively influence the response (i.e. exhibit higher
ENR values). Moreover the shorter the side chain (i.e. more water soluble ILs located in the lower
right part of the plot) the higher the ENR values. When considering different ILs with the same
cation (IM14) the anions size plays a relevant role evidenced by the second component, in
agreement with the positive contribution of the anionic size descriptors: the higher the PC2 score,
the higher the ENR value.
9

ACCEPTED MANUSCRIPT
Figure 2. PC1 vs. PC2 scores plot reporting the ILs 1-18 together with y-response: this plot can
show similarities and dissimilarities among the objects (ILs) and allows to interpret how they are
st
nd
related to the response (ENR variable) along the 1 and the 2 PLS components.
The above considerations show that the present QSPR approach, considering simultaneous
variations in both the cation (heterocyclic core and side chain length) and the anion structural
features by means of 176 ILs descriptors is able to provide an estimate of their importance and a
tentative interpretation of their physico-chemical significance in relation to the polarity of ILs. It is
perhaps worth mentioning here that attempted correlations with Kosower and Reichardt empirical
solvent parameters were not satisfactory probably due to the ionic structure of the solvatochromic
dyes which in solution with ILs implies the presence of two cations and two anions, a condition
quite difficult to be modelled. A neutral “sensor” such as Nile Red simplifies the above picture
rendering in silico modelling possible.
1
0

ACCEPTED MANUSCRIPT
3.
Conclusions
VolSurf+ descriptors, accounting for both cationic and anionic structural features of ILs were used
to develop a PLS model able to establish a QSPR correlation with their solvatochromic dye Nile
Red polarity. The PLS model allowing prediction of ENR values provided an in silico ILs polarity
database for over 100 ILs. The above database including structurally different ILs with similar
polarity provides the possibility to select an optimal IL according not only to its polarity but also to
other properties such as synthetic affordability, environmental sustainability in its life cycle and
economic considerations. The same approach, using readily available descriptors and an accessible
statistical procedure such as PLS, could be extended to develop QSPR models for other relevant ILs
physico-chemical properties.
4
4
. Experimental
.1 Synthesis and characterization of ILs
[
[
IM14][(1O)2PO2], was purchased and used without any other purification. [IM14][CF3COO] and
IM14][1COO] were obtained by anion exchange on resin starting from corresponding halide as
following reported. All ionic liquids were dried on a vacuum line at 60 °C for at least 2 h before
use, then stored in a dryer under argon and over calcium chloride. In all cases, the silver nitrate test
performed to verify the presence of residual halide anion, gave a negative result.
4
.1.1 General procedure for anion exchange on resin. The Amberlite resin IRA-400 (16 g) was
converted from chloride form to hydroxide form by eluting an aqueous solution of NaOH (10 mL,
0% w/v) within a column. Subsequently the resin was washed with water until the eluate was
1
neutral. The halide salt (10 mmol), dissolved in methanol/water (70/30, v/v), was eluted, and the
eluate collected into a flask containing a solution of the corresponding acid in stoichiometric
amount, until neutralization. The resulting solution was concentrated in vacuo, then treated with
activated charcoal in ethanol. Filtration and removal of the solvent afforded the ionic liquids as
viscous oils in high to quantitative yields.
1-Butyl-3-methylimidazolium trifluoroacetate ([IM14][CF3COO]).
Pale yellow oil. Yield: 94%.
-1
νMAX (Nujol) 3150, 3095, 2970, 2880, 1700 cm
1
H NMR (300 MHz, d -DMSO): δ=9.34 (s, 1 H), 7.91 (d, J=24 Hz, 2 H), 4.33 (t, J=8 Hz, 2 H),
6
4.01 (s, 3H), 1.93 (quint, J=7.6 Hz, 2 H), 1.40 (quint, J=7.6 Hz, 2 H), 1.07 (t, J=7.6 Hz, 3 H) ppm.
1
3
C NMR (300 MHz, d -DMSO): δ= 158.0, 139.4, 125.3, 122.1, 50.4, 37.1, 30.9, 30.2, 22.4, 17.6
6
ppm.
1
1

ACCEPTED MANUSCRIPT
1-Butyl-3-methylimidazolium acetate. ([IM14][1COO]).
Pale yellow oil. Yield: 99%.
-1
νMAX (Nujol) 3150, 3095, 2970, 2880, 1700 cm
1
H NMR (300 MHz, d -DMSO): δ=9.73 (s, 1 H), 7.88 (d, J=21 Hz, 2 H), 4.24 (t, J= 7.6 Hz, 2 H),
6
3.92 (s, 3 H), 1.82 (quint, J=7.5 Hz, 2 H), 1.64 (s, 3 H), 1.29 (quint, J=7 Hz, 2 H), 0.95 (t, J=7 Hz, 3
1
3
H) ppm. C NMR (300 MHz, d -DMSO) δ= 172.9, 137.4, 123.7, 122.4, 59.2, 48.5, 35.7, 25.9,
6
18.9, 13.4 ppm.
4.2
ENR measurements
The determination of polarity parameter ENR was carried out by mixing into a quartz cuvette
(optical path 0.2 cm), 500 µL of IL and 75 µL of a concentrated solution of spectroscopic probe in
-4
1
,4-dioxane. The concentration of the probe was equal to 2.0 · 10 M. The obtained solution was
thermostated at 298 K.
4.3
Computational Methods
7
In VolSurf+, the GRID force field was chosen to characterize potential polar and hydrophobic
interaction sites around target molecules by the water (OH2), the hydrophobic (DRY), the carbonyl
oxygen (O) and amide nitrogen (N1) probes. The information contained in the resulting MIFs is
transformed into a quantitative scale by calculating the volume or the surface of the interaction
contours. In particular, the VolSurf+ procedure consists of the generation of the 3D molecular fields
from the interactions of the OH2, DRY, O and N1 probes around a target molecule, next the
calculation of descriptors from the 3D maps obtained in the first step. The definition of all 128
5
,14,15,39-45
46-
VolSurf+ descriptors is given in case studies
and with the previous versions of VolSurf.
4
9
16
By adopting the procedure already reported in a recent QSAR work 176 in silico molecular
descriptors for cations and anions (128 and 48 respectively) were calculated and here used as the
X-matrix in a correlation model in which the ENR values are used as the Y-responses by means of
partial least squares model.
5
0
The Partial Least Squares Projections to Latent Structures” (PLS) chemometric tool, available in
3
4
the SIMCA Software package, was adopted.
Acknowledgements
We thank the University of Catania for partial financial support (FIR project ECDF5E) and for a
PhD grant (to AP).
References
1
2

ACCEPTED MANUSCRIPT
1
2
3
. Hayes, R.; Warr, G. G.; Atkin R. Chem. Rev. 2015, 115, 6357−6426.
. Mečiarová, M.; Toma, Š. Chem. Eur. J. 2007, 13, 1268-1272.
. Millan, D.; Rojas, M.; Pavez, P.; Isaacs, M.; Diaz, C.; Santos, J. G. New J. Chem. 2013, 37,
3281-3288
4
5
6
7
8
9
. Singh, A.; Kumar, A. RSC Adv. 2015, 5, 2994-3004.
. Cruciani, G.; Crivori, P.; Carrupt, P. A.; Testa B. J. Mol. Struct. Theochem. 2000, 503, 17–30.
. VolSurf+ manual and software available at http://www.moldiscovery.com/docs/vsplus/.
. Carosati, E.; Sciabola, S.; Cruciani, G. J. Med. Chem., 2004, 47, 5114-5125
. Goodford, P. J. J. Med. Chem., 1985, 28, 849-857.
. Boobbyer, D. N. A.; Goodford, P. J.; Mcwhinnie, P. M.; Wade, R. C. J. Med. Chem., 1989, 32,
1083-1094.
1
0. Wade, R.; Clerk, K. J.; Goodford, P. J. J. Med. Chem., 1993, 36, 140-147.
1. Barresi, V.; Bonaccorso, C.; Consiglio, G.; Goracci, L.; Musso, N.; Musumarra, G.; Satriano,
C.; Fortuna, C. G. Mol. Biosys. 2013, 9, 2426-2429.
1
12. Fortuna, C. G.; Barresi, V.; Bonaccorso, C.; Consiglio, G.; Failla, S.; Musso, N.; Musumarra,
G. Eur. J. Med. Chem. 2012, 47, 221-227.
13. Fortuna, C. G.; Barresi, V.; Musumarra, G. Bioorg. Med. Chem. 2010, 18, 4516-4523.
14. Goracci, L.; Ceccarelli, M.; Bonelli, D.; Cruciani G. J. Chem. Inf. Model. 2013, 53, 1436-1446.
15. Goracci, L.; Buratta, S.; Urbanelli, L.; Ferrara, G.; Di Guida, R.; Emiliani, C.; Cross S. Eur. J.
Med. Chem. 2015, 92, 49-63.
16. Paternò, A.; Bocci, G.; Goracci, L.; Musumarra, G.; Scirè, S. SAR QSAR Environ. Res. 2016,
27, 1-15.
1
1
1
7. Keaveney, S. T.; Harper, J. B.; Croft A. K. RSC Adv. 2015, 5, 35709-35729.
8. Kosower, E. M. J. Amer. Chem. Soc. 1958, 80, 3253-3260.
9. Reichardt C. In Solvents and Solvent Effects in Organic Chemistry-2, rev. and enl. ed.; VCH,
Ed.; D-6940 Weinheim, 1988.
20. Reichardt C. Chem. Rev. 1994, 94, 2319-2358.
21. Wakai, C.; Oleinikova, A.; Ott, M.; Weingärtner, H. J. Phys. Chem. B 2005, 109, 17028-
17030.
2
2
2
2. Weingärtner, H.; Sasisanker, P.; Daguenet, C.; Dyson, P. J.; Krossing, I.; Slattery, J. M.;
Schubert, T. J. Phys. Chem. B 2007, 111, 4775-4780.
3. Huang, M. M.; Jiang, Y.; Sasisanker, P.; Driver, G. W.; Weingärtner, H. J. Chem. Eng.Data
2011, 56, 1494-1499.
4. Weingärtner, H. J. Mol. Liq. 2014, 192, 185-190.
1
3

ACCEPTED MANUSCRIPT
2
2
2
2
2
3
3
3
3
3
3
5. Hallett, J. P.; Welton, T. Chem. Rev 2011, 111, 3508-3576.
6. D'Anna, F.; Marullo, S.; Vitale, P.; Noto, R. ChemPhysChem 2012, 13, 1877-1884.
7. Vanyúr, R.; Biczók, L.; Miskolczy, Z. Colloids Surf. A 2007, 299, 256-261.
8. Carmichael, A. J.; Seddon K. R. J. Phys. Org. Chem. 2000, 13, 591-595.
9. Spange, S.; Lungwitz, R.; Schade A. J. Mol. Liq. 2014, 192,137–143.
0. Chiappe, C.; Pomelli, C. S.; Rajamani, S. J. Phys. Chem. B 2011, 115, 9653–9661.
1. D'Anna, F.; Marullo, S.; Vitale, P.; Noto, R. Eur. J. Org. Chem. 2011, 2011, 5681-5689.
2. D'Anna, F.; La Marca, S.; Lo Meo, P.; Noto, R. Chem. Eur. J. 2009, 15, 7896-7902
3. D’Anna, F.; Millan, D.; Noto, R. Tetrahedron 2015, 71, 7361-7366.
4. SIMCA v.13.0.3 MKS Umetrics AB, Malmo, Sweden, 2013.
5. Cláudio, A. F. M.; Swift, L.; Hallett, J. P.; Welton, T.; Coutinho, J. A. P.; Freire M. G.
Phys. Chem. Chem. Phys. 2014, 16, 6593-6601.
36. Seddon, K. R.; Stark, A.; Torres M. J. Pure Appl. Chem. 2000, 72, 2275–2287.
37. Holbrey, J. D.; Seddon, K. R. J. Chem. Soc., Dalton Trans. 1999, 2133-2140.
38. Triolo, A.; Russina, O.; Bleif, H.-J.; Di Cola, E. J. Phys. Chem. B 2007, 111, 4641-4644.
39. Fortuna, C. G.; Barresi, V.; Berellini, G.; Musumarra G. Bioorg. Med. Chem. 2008, 16, 4150-
4159.
4
0. Forte, G.; Fortuna, C. G.; Salerno, L.; Modica, M. N.; Siracusa, M. A.; Cardile, V.; Romeo, G.;
Bulbarelli, A.; Lonati, E.; Pittala, V. Bioorg. Med. Chem. 2013, 21, 5233-5245.
1. Barone, M.; Santagati, A.; Graziano, A. C. E.; Fortuna, C. G.; Ronsisvalle, G.; Cardile, V. Med.
Chem. 2014, 7, 700-710.
4
4
4
4
2. Riduan, S. N.; Zhang, Y. Chem. Soc. Rev. 2013, 42, 9055-9070.
3. Crivori, P.; Cruciani, G.; Carrupt, P. A.; Testa, B. J. Med. Chem. 2000, 43, 2204-2216.
4. Cruciani, G.; Meniconi, M.; Carosati, E.; Zamora, I.; Mannhold, R. In Drug Bioavailability:
Estimation of Solubility, Permeability, Absorption and Bioavailability, H. van de Waterbeemd,
H. Lennernäs, P. Artursson, Eds., Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2003,
pp 406-419.
4
5. Berellini, G.; Cruciani, G.; Mannhold, R. J. Med. Chem. 2005, 48, 4389-4399.
6. Mannhold, C. R.; Cruciani, G.; Weber, H.; Lemoine, H.; Derix, A.; Weichel, C.; Clementi, M.
J. Med. Chem. 1999, 42, 981-991.
4
4
7. Cruciani, C. G.; Pastor M.; Guba, W. Eur. J. Pharm. Sci. 2000, 11, S29-S39.
8. Cruciani, G.; Milletti, F.; Storchi, L.; Sforna, G.; Goracci, L. Chem. Biodivers. 2009, 6, 1812-
4
1821.
1
4

ACCEPTED MANUSCRIPT
4
5
9. Milletti, F.; Storchi, L.; Goracci, L.; Bendels, S.; Wagner, B.; Kansy, M.; Cruciani, G. Eur. J.
Med. Chem. 2010, 45, 4270-4279
0. Wold, S.; Sjöström, M.; Eriksson, L. In The Encyclopedia of Computational Chemistry,
Schleyer P.v.R., Ed., John Wiley & Sons, Chichester, 1998, pp 2006-2020.
1
5