A family of chiral ionic liquids from the natural pool: Relationships between structure and functional properties and electrochemical enantiodiscrimination tests

Article abstract of DOI:10.1016/j.electacta.2018.12.060

In spite of the increasing fundamental and practical interest of electrochemistry in ionic liquids (ILs), exploration of chiral ionic liquids (CILs) in view of enantioselective electrochemistry and electroanalysis is surprisingly overdue. In this study a

Full text of DOI:10.1016/j.electacta.2018.12.060

Accepted Manuscript
A family of chiral ionic liquids from the natural pool: Relationships between structure
and functional properties and electrochemical enantiodiscrimination tests
Mariangela Longhi, Serena Arnaboldi, Elena Husanu, Sara Grecchi, Ivo Franco Buzzi,
Roberto Cirilli, Simona Rizzo, Cinzia Chiappe, Patrizia Romana Mussini, Lorenzo
Guazzelli
PII:
S0013-4686(18)32758-0
https://doi.org/10.1016/j.electacta.2018.12.060
EA 33275
DOI:
Reference:
To appear in: Electrochimica Acta
Received Date: 10 September 2018
Revised Date: 7 December 2018
Accepted Date: 10 December 2018
Please cite this article as: M. Longhi, S. Arnaboldi, E. Husanu, S. Grecchi, I.F. Buzzi, R. Cirilli, S.
Rizzo, C. Chiappe, P.R. Mussini, L. Guazzelli, A family of chiral ionic liquids from the natural pool:
Relationships between structure and functional properties and electrochemical enantiodiscrimination
tests, Electrochimica Acta (2019), doi: https://doi.org/10.1016/j.electacta.2018.12.060.
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
BioBased
chiral ionic liquid
myrthenol
Me
∆E
OH
(R)-Fc
(S)-Fc
Me2N
Fe
myrthus
myrthus communis: adapted from CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=1381763

ACCEPTED MANUSCRIPT
A family of chiral ionic liquids from the natural pool:
relationships between structure and functional properties
and electrochemical enantiodiscrimination tests
Mariangela Longhia, Serena Arnaboldia,
Elena Husanub, Sara Grecchia, Ivo Franco Buzzia, Roberto Cirillic
Simona Rizzod, Cinzia Chiappeb, Patrizia Romana Mussinia,*, Lorenzo Guazzellib,*
aUniversità degli Studi di Milano, Dipartimento di Chimica, Via Golgi 19, 20133 Milano (Italy)
bUniversità degli Studi di Pisa, Dipartimento di Farmacia, Via Bonanno 33, 56126 Pisa (Italy)
cIstituto Superiore di Sanità, Centro Nazionale per il Controllo e la Valutazione dei Farmaci,
Viale Regina Elena 299, 00161 Rome (Italy)
cCNR ISTM, Milano, Via Golgi 19, 20133 Milano (Italy)
* Corresponding Authors. Patrizia Romana Mussini, e-mail: patrizia.mussini@unimi.it;
Lorenzo Guazzelli, e-mail: lorenzo.guazzelli@unipi.it
1

ACCEPTED MANUSCRIPT
Abstract
In spite of the increasing fundamental and practical interest of electrochemistry in ionic liquids
(ILs), exploration of chiral ionic liquids (CILs) in view of enantioselective electrochemistry and
electroanalysis is surprisingly overdue. In this study a family of chiral ionic liquids (CILs) based on
natural chiral building blocks, of easy synthesis, is detailedly characterized in terms of thermal and
electrochemical properties, achieving valuable information about structure−property relationships
on account of the systematicity of available family terms. Moreover, they are submitted to a series
of chiral electroanalysis tests. Cyclic voltammetry in bulk CILs or with CIL as additives in a bulk
achiral ionic liquid IL, shows small but statistically significant potential differences for the
enantiomers of two quite different chiral probes, an interesting result since enantiodiscrimination in
terms of potential differences in chiral voltammetry (more desirable respect to current differences)
has been only seldom obtained so far. The present first example of enantioselective voltammetry in
CIL media with stereocenters as localized chirality sources also offers an important and so far
missing confirmation of the intrinsically superior level of the inherent chirality strategy, recently
resulting in larger potential differences with the same protocols implemented in "inherently chiral"
ionic liquid media. Furthermore, an alternative transduction mode, based on electrochemical
impedance spectroscopy EIS, is proposed to effectively highlight the enantiodiscrimination ability
of CIL media in analytical experiments.
2

ACCEPTED MANUSCRIPT
1. Introduction
In the last twenty years, ionic liquids (ILs) have become an area of intensive scientific investigation.
Constituted exclusively by ions, ILs are organic salts with melting point <100°C, often even liquid
at room temperature (RTILs), regarded as green alternatives to traditional organic volatile solvents
(VOCs) on account of their typical high thermal stability and almost negligible vapor tension. ILs
are also attractive on account of the wide tunability of their physico-chemical and functional
properties through molecular design. They have thus been increasingly employed by scientists of all
branches of chemistry in a polyhedral application range (e.g., biopolymer dissolution and
transformation [1-3], organic synthesis [4], functional material preparation [5], modulators for
fragrance release [6]). Some IL features, like their ionic conductivity (with them acting both as
solvent and supporting electrolyte), wide potential window, high viscosity, high order at a charged
electrode surface extending from many layers, make them particularly attractive for electrochemical
applications, ranging from electrodeposition of inorganic or organic films, to batteries and
capacitors, to analytical and preparative electrochemical applications [7-10].
Although it is possible to formulate biocompatible ILs, recent studies have demonstrated that a
significant number of reported ILs are no less toxic than conventional solvents, at least to the
aquatic environment [11]. Hence, the environmental impact of ILs has gained attention in academia
and industry, and the preparation of ILs from renewable materials has attracted a growing interest in
recent years [12]. Several classes of natural compounds, for instance amino acids [13],
carbohydrates [12], fatty acids [14], and terpenoids [12], have been investigated for the synthesis of
"BioBased" ILs, of more complex structures and functional properties than the commonly used ILs
bearing simple alkyl chains. For example, they include many examples of chiral ionic liquids, CILs.
CILs have been so far mainly investigated as catalysts in asymmetric organic reactions and as chiral
selectors in chromatographic purifications and in spectroscopic analysis [15]. However, studies are
increasingly focusing on other CIL features, such as the relationship between the thermophysical
properties and the alkyl side chain length of the cation [16], the surface activity of a homologous
3

ACCEPTED MANUSCRIPT
series of CILs [17], the arrangement at the liquid/air interface [18], and the chiral conformation of
the achiral counterion induced by the chiral ion [19]. Furthermore, the high local order at the
electrode surface can make CILs better candidate media for chiral electrochemistry and
electroanalysis than chiral solvents or chiral supporting electrolytes [20]. However, surprisingly, the
potentialities of CIL media for enantioselective electrochemistry have remained nearly unexplored
until very recent studies by some of us [21,22].
In this context, a detailed study is proposed of a series of BioBased CILs, consisting of imidazolium
or pyrrolidinium cations with chiral substituents obtained from natural terpenoids (1R)-myrtenol 1a,
(S)-perillyl alcohol 1b, (1R)-nopol 1c, and (S)-citronellol 1d (Fig. 1) and of bistriflimide as counter
anion..
Fig. 1. Natural terpenoids used for the synthesis of the corresponding CILs
The systematic CIL series available (four of them having melting points below 25 °C, and thus
being RTCILs) was investigated by thermogravimetric analysis TGA, differential scanning
calorimetry DSC, optical rotatory dispersion ORD and electronic circular dichroism ECD, and the CIL
electrochemical features and enantiodiscrimination ability studied by cyclic voltammetry CV and
electrochemical impedance spectroscopy EIS. Systematic sequences in the family enabled to
identify correlations between structure and functional properties, of interest for application as media
particularly in electrochemistry and electroanalysis.
Concerning the latter perspective, recently some of us have highlighted the potentialities of chiral
ionic liquids to achieve enantioselective voltammetry, resulting in significant peak potential
differences between the enantiomers of chiral electroactive probes even as low-concentration
4

ACCEPTED MANUSCRIPT
additives in common achiral ionic liquids [21,22]. These outstanding performances were justified
[20,21] considering the exceptionally high local order at the ionic liquid|charged electrode
interphase, recently experimentally demonstrated [10]. Notably, such first studies employed "inhe-
rently chiral" ionic liquids (ICILs). In "inherently chiral" molecules and materials, both chirality
and key functional properties originate from the same structural element, so far typically a tailored
torsion with high racemization barrier in the whole main molecular backbone (in the ICIL case, in
the cation backbone) [21]. In the much more investigated case of chiral electrode surfaces,
inherently chiral selectors showed outstanding enantioselectivity respect to chiral selectors with chi-
rality originating from simple stereocenters external to the main backbone [20].
In this frame it was important and overdue to investigate the enantiodiscrimination performances of
"classical" CILs bearing localized stereocenters comparing them with the ICIL outstanding ones.
This prompted us to perform a series of chiral electroanalysis tests, involving different CIL media,
two significantly different chiral probes, and two transduction modes (CV and EIS).
2. Experimental
2.1 Synthesis
Compounds 2c, 2d, 3a, 6c, 6d, 8c, and 8d (Scheme 1) have been prepared following reported
procedures [18,23].
2.1.1 Synthesis of N-methylpyrrolidinium methanesulphonates: representative procedure:
A mixture of the mesylate 2 (50 mmol) and N-methypyrrolidine (50 mmol) in CH3CN (20 ml) was
refluxed for 72h. The solvent was removed under vacuum obtaining the pure product.
1-{2-[(1R,5S)-6,6-dimethylbicyclo[3.1.1]hept-2-en-2yl]ethyl}-1-methylpyrrolidin-1-ium
methanesulfonate (9c): 16.2 g (49.1 mmol, 98%) white solid; mp: 113-118°C. 1H NMR (250 MHz,
CDCl3) δ 5.37 (d, J = 1.5 Hz, 1H), 3.72 (s, 2H), 3.60 (s, 2H), 3.47 – 3.23 (m, 2H), 3.17 (d, J = 3.7
Hz, 3H), 2.72 – 2.52 (m, 3H), 2.34 (ddd, J = 9.8 Hz, J = 8.6 Hz, J = 5.2 Hz, 3H), 2.17 (s, 6H), 2.07
– 1.90 (m, 2H), 1.32 – 1.13 (m, 3H), 1.05 (dd, J = 8.6, 2.7 Hz, 1H), 0.80 – 0.65 (m, 3H); 13C NMR
5

ACCEPTED MANUSCRIPT
(63 MHz, CDCl3) δ 141.6, 120.8, 64.3, 61.9, 48.4, 45.5, 40.4, 39.6, 38.1, 31.6, 31.3, 31.2, 26.0,
21.6, 21.6, 21.1.
2.1.2 Alkylchloride synthesis from perillyl alcohol:
A solution of methanesulfonyl chloride (50 mmol) in CH2Cl2 (20 ml) was added dropwise at 0°C to
a solution of alcohol 1 (50 mmol) and DIPEA (55 mmol) in CHCl3 (25 ml) during 20 min, then the
reaction mixture was stirred at room temperature for 6h. The reaction was quenched with saturated
aqueous NaHCO3 solution and was extracted with CHCl3 (3 x 20ml). The collected organic phases
were dried over anhydrous Na2SO4, then the solvent was evaporated under reduced pressure giving
the pure product.
(S)-1-(chloromethyl)-4-(prop-1-en-2-yl)cyclohex-1-ene (3b): 8.19 g (48.0 mmol, 96%) yellow
liquid. 1H NMR (250 MHz, CDCl3) δ 5.83 – 5.75 (m, 1H), 4.75 – 4.65 (m, 2H), 3.97 (s, 2H), 2.24 –
13
2.04 (m, 4H), 2.04 – 1.78 (m, 2H), 1.72 – 1.67 (m, 3H), 1.57 – 1.38 (m, 1H); C NMR (63 MHz,
CDCl3) δ 149.2, 134.1, 126.9, 108.9, 50.0, 40.6, 30.6, 27.3, 26.3, 20.7.
2.1.3 Synthesis of N-methylimidazolium chloride. Representative procedure:
A mixture of the allylic chloride 3 (51 mmol) and N-methylimidazole (50 mmol) in CH3CN (20 ml)
was refluxed for 72h. The solvent was removed under vacuum obtaining the pure product.
1-{[(1R,5S)-6,6-dimethylbicyclo[3.1.1]hept-2-en-2-yl]methyl}-3-methylimidazolium (4a): 12.41 g
1
(49.1 mmol, 98%) white solid. H NMR (250 MHz, CDCl3) δ 9.81 (s, 1H), 7.44 (s, 1H), 6.81 (s,
1H), 5.22 (s, 1H), 4.26 (s, 2H), 3.56 (s, 3H), 1.89 – 1.37 (m, 5H), 0.64 (s, 3H), 0.52 (d, J = 8.8 Hz,
1H), 0.14 (s, 3H); 13C NMR (63 MHz, CDCl3) δ 139.5, 136.3, 124.5, 123.2, 120.6, 53.3, 42.2, 39.1,
36.9, 35.5, 30.6, 30.2, 24.7, 20.0.
(S)-1-methyl-3-{[4-(prop-1-en-2-yl)cyclohex-1-en-1-yl]methyl}imidazol-3-ium chloride (4b):
12.13 g (48.0 mmol, 96%) pale yellow liquid. 1H NMR (250 MHz, CDCl3) δ 10.12 (s, 1H), 7.62 (t,
J = 1.8 Hz, 1H), 7.20 (t, J = 1.8 Hz, 1H), 5.76 (d, J = 3.7 Hz, 1H), 4.70 (s, 2H), 4.59-4.50 (m, 2H),
13
3.97 (s, 3H), 2.13 – 1.58 (m, 6H), 1.57 – 1.53 (m, 3H), 1.40 – 1.20 (m, 1H); C NMR (63 MHz,
6

ACCEPTED MANUSCRIPT
CDCl3) δ 147.3, 136.0, 129.3, 128.0, 122.9, 120.5, 107.7, 54.0, 38.7, 35.1, 29.0, 25.5, 24.7, 19.4.
2.1.4 Synthesis of N-methylpyrrolidinium chloride. Representative procedure:
A mixture of the allylic chloride (51 mmol) and N-methylpirrolidine (50 mmol) in CH3CN (20 ml)
was refluxed for 72h. The solvent was removed under vacuum obtaining the pure product.
(S)-1-methyl-1-((4-(prop-1-en-2-yl)cyclohex-1-en-1-yl)methyl)- 1-methylpyrrolidin-1-ium chloride
1
(5b): 12.4 g (48.6 mmol, 97%) white solid. H NMR (250 MHz, CDCl3) δ 6.12 (s, 1H), 4.61 (s,
1H), 4.54 (s, 1H), 4.14 (s, 2H), 3.79 – 3.51 (m, 4H), 3.06 (s, 3H), 2.27 – 1.66 (m, 10H), 1.58 (s,
13
3H), 1.48 – 1.27 (m, 1H); C NMR (63 MHz, CDCl3) δ 148.2, 138.2, 127.0, 109.2, 69.7, 63.9,
63.8, 48.0, 39.5, 30.8, 29.8, 27.0, 21.4, 21.2, 20.6.
2.1.5. Metathesis reactions. Representative procedure:
An aqueous solution of LiTf2N (50 mmol) was added to an aqueous solution of the IL (50 mmol)
and the mixture was stirred at room temperature for 24 h. The insoluble oily material was separated
adding CH2Cl2 and the organic layer was washed with water and brine and dried over anhydrous
Na2SO4. The solvent was evaporated under reduced pressure to give the expected product.
1-{2-[(1R,5S)-6,6-dimethylbicyclo[3.1.1]hept-2-en-2yl]ethyl}-1-methylpyrrolidin-1-ium
bis[(trifluoromethyl) sulfonyl]imide (7c): 24.9 g (48.4 mmol, 97%) pale yellow liquid; [α]D -17,9 [c
1,07 in CHCl3]. 1H NMR (250 MHz, CDCl3) δ 5.40 (s, 1H), 3.49 (td, J = 11.6, 6.9 Hz, 4H), 3.33 –
3.20 (m, 2H), 3.01 (s, 3H), 2.49 – 2.31 (m, 3H), 2.24 (t, J = 9.1 Hz, 6H), 2.15 – 1.93 (m, 2H), 1.23
13
(d, J = 10.9 Hz, 3H), 1.11 (d, J = 9.6 Hz, 1H), 0.78 (s, 3H); C NMR (63 MHz, CDCl3) δ 141.2,
127.5, 122.4, 121.2, 117.3, 112.2, 64.7, 63.0, 48.3, 45.4, 40.4, 38.2, 31.6, 31.3, 31.2, 26.0, 21.5,
21.4, 21.0.
1-{[(1R,5S)-6,6-dimethylbicyclo[3.1.1]hept-2-en-2yl]methyl}-3-methylimidazolium
bis[(trifluoromethyl)sulfonyl]imide (6a): 24.1 g (48.4 mmol, 97%) pale yellow liquid; [α]D -17,8 (c
1,0 in CHCl3). 1H NMR (250 MHz, CDCl3) δ 8.63 (s, 1H), 7.33 (s, 1H), 7.21 (s, 1H), 5.72 (s, 1H),
7

ACCEPTED MANUSCRIPT
4.60 (s, 2H), 3.89 (s, 3H), 2.44 – 2.19 (m, 3H), 2.08 (s, 1H), 1.93 (dd, J = 11.9 Hz, J = 6.8 Hz, 1H),
1.21 (s, 3H), 1.06 (d, J = 8.9 Hz, 1H), 0.71 (s, 3H). 13C NMR (63 MHz, CDCl3) δ 140.2, 136.0,
127.5, 126.0, 123.9, 122.4, 122.2, 117.3, 112.2, 54.7, 43.2, 40.1, 38.1, 36.3, 31.6, 31.3, 25.8, 20.8.
(S)-1-methyl-1-((4-(prop-1-en-2-yl)cyclohex-1-en-1-yl)methyl)-3-methylimidazolium
bis[(trifluoromethyl)sulfonyl]imide (6b): 24.4 g (49.1 mmol, 98%) pale yellow liquid; [α]D -34,5 (c
1,06 in CHCl3). 1H NMR (250 MHz, CDCl3) δ 8.59 (s, 1H), 7.32 (t, J = 1.8 Hz, 1H), 7.23 (t, J = 1.8
Hz, 1H), 5.86 (d, J = 2.8 Hz, 1H), 4.70 – 4.66 (m, 1H), 4.66 – 4.63 (m, 1H), 4.58 (s, 2H), 3.88 (s,
3H), 2.18 – 1.75 (m, 6H), 1.71 – 1.64 (m, 3H), 1.53 – 1.33 (m, 1H); 13C NMR (63 MHz, CDCl3) δ
148.8, 135.9, 130.3, 130.0, 127.4, 123.9, 122.3, 122.1, 117.2, 112.1, 109.1, 55.9, 40.1, 36.2, 30.3,
26.8, 25.8, 20.6.
(S)-1-methyl-1-((4-(prop-1-en-2-yl)cyclohex-1-en-1-yl)methyl)pyrrolidin-1-ium
bis[(trifluoromethyl)sulfonyl]imide (7b): 24.5 g (49.0 mmol, 98%) pale yellow liquid; [α]D -24,3 (c
1,26 in CHCl3). 1H NMR (250 MHz, CDCl3) δ 6.11 (s, 1H), 4.84 – 4.55 (m, 2H), 3.80 (s, 2H), 3.46
(t, J = 8.9 Hz, 4H), 2.93 (s, 3H), 2.43 – 2.00 (m, 9H), 1.83 (dd, J = 5.1 Hz, J = 10.4 Hz, 1H), 1.72
13
(s, 3H), 1.64 – 1.41 (m, 1H); C NMR (63 MHz, CDCl3) δ 148.3, 138.4, 127.3, 126.4, 122.2,
117.1, 111.9, 108.9, 70.5, 64.1, 64.0, 47.6, 39.3, 30.7, 29.4, 26.9, 21.0, 20.8, 20.3.
2.2 ESI Mass
The identity of the new products 6a, 6b, 7b and 7c was confirmed by electrospray mass
spectroscopy using an API3200QTRAP hybrid triple quadrupole/linear ion trap (ABSciex, Foster
City, CA) in positive mode.
COMPOUND 6a ESI-MS: (+ve) 217 m/z [C14H21N2]+
COMPOUND 6b ESI-MS: (+ve) 217 m/z [C14H21N2]+
COMPOUND 7b ESI-MS: (+ve) 220 m/z [C15H26N]+
COMPOUND 7c ESI-MS: (+ve) 234 m/z [C16H28N]+
8

ACCEPTED MANUSCRIPT
2.3 Infrared spectra
Infrared spectra of the new products 6a, 6b, 7b and 7c were recorded using an ATR-FTIR Agilent
660 (Agilent Technologies, Santa Clara, CA, USA). They are reported in the SI.
2.4 ORD and CD spectra
ECD spectra were recorded in acetonitrile solution (concentration about 0.30 mg cm-3) in 0.1 cm
path-length quartz cell at 25°C on a Jasco Model J-700 spectropolarimeter. The spectra are average
computed over three instrumental scans and the intensities are presented in terms of ellipticity
values (mdeg). Specific rotations were measured in acetonitrile solution (concentration 5 mg cm-3)
by a PerkinElmer polarimeter model 241 equipped with Na/Hg lamps. The volume of the cell was 1
cm3 and the optical path was 10 cm. The system was set at a temperature of 20°C.
2.5 TGA characterizations
TGA was performed by TA Instruments Q500 TGA. Dynamic experiments on samples placed in an
open platinum crucible were performed at temperatures from 60 to 600 °C, with a single linear
heating ramp at 10 °C min-1 under a N2 flux (flow = 10 cm3 min-1).
2.6 DSC and TGA+DSC characterizations
Differential Scanning Calorimetry DSC characterizations have been performed by complex
temperature scan cycles between 188 and 473 K (details in the Discussion paragraph), under
constant nitrogen flux (50 cm3 min-1) by a DSC 3 500 STARe system from Mettler-Toledo®
equipped with a HUBER TC100-MT cooler, on a few mg of CILs (weighed on a high-precision
balance) in pinholed standard aluminum crucibles (No. 00026763, 40 mm3). Weight constancy
between room temperature and 473 K has been verified by a combined TGA/DSC 3+ STARe
system from Mettler-Toledo®, in standard alumina crucibles (No. 00024123, 70 mm3), under the
same nitrogen flux. The same instrument was also used to record the TGA pattern of commercial
BMIMTf2N.
9

ACCEPTED MANUSCRIPT
2.7 Electrochemistry
2.7.1 CIL characterization. The seven salts 6a-d, 7b,7c and 8c were electrochemically
characterized by cyclic voltammetry CV at scan rates ranging 0.05-2 V/s using an AutoLab PGStat
potentiostat and a classical three-electrode glass minicell (working volume about 2 cm3). The latter
included as working electrode a glassy carbon GC disk embedded in glass (Metrohm, 1.5 mm
diameter) polished by diamond powder (1µm Aldrich) on a wet cloth (Struers DP-NAP), as counter
electrode a platinum disk, and as reference electrode a saturated aqueous calomel one (SCE)
inserted in a compartment with the working medium ending in a porous frit, to avoid contamination
of the working solution by water and KCl traces. Experiments were run on 0.00075 M solutions in
ACN + 0.1 M tetraethylammonium tetrafluoroborate TEABF4 as the supporting electrolyte,
previously deaerated by nitrogen bubbling. Positive and negative half cycles have been separately
recorded to avoid reciprocal contamination by electron transfer products. The reported potentials
have been normalized vs the formal potential of the intersolvental ferricinium|ferrocene (Fc+|Fc)
reference redox couple, recorded in the same conditions.
2.7.2 CV enantioselectivity tests CV enantioselectivity tests were performed by cyclic voltammetry
at 0.05 V/s scan rate on screen printed electrode (SPE) supports (Dropsens, custom made without
paint, with Au working and counter electrodes and an Ag pseudoreference electrode, resulting in
good reproducibility at constant conditions with the present working protocol). Potentials were
referred to the formal potential of the ferrocene|ferrocinium intersolvental reference redox couple,
obtained by recording its CV pattern in the same conditions of the enantioselection tests. The
experiments were performed either with bulk CILs (6a, 6b) or using solid or liquid family terms
(6a, 6b, 7c) as low-concentration chiral additives in achiral commercial IL 1-butyl-3-
methylimidazolium bis(trifluoromethanesulfonyl)imidate BMIMTf2N (or BMIMTFSI; CAS
174899-83-3; Aldrich ≥ 98%) having the same counteranion. CVs were recorded in open air
conditions, depositing on the working electrode a drop of one of the above chiral media with i)
10

ACCEPTED MANUSCRIPT
0.002 M (R)-(+)- or (S)-(–)-N,N’-dimethyl-1-ferrocenylethylamine (Aldrich, submitted to a further
chromatographic purification step) as the enantiopure probe, already adopted as model one (on
account of its chemical and electrochemical reversibility) by some of us when testing "inherently
chiral" electrode surfaces and media [21,24] ii) in the presence of 0.01 M of either enantiomer of
the chiral probe BT2T4 (2,2′-Bis(2,2′-bithiophene-5-yl)-3,3′-bithianaphthene) [20]. For the sake of
comparison, the tests were also carried out with the same protocol, but in the absence of the chiral
additives. Concentration calibration plots have also been obtained for bulk CILs (6a, 6b), by
recording CV patterns at increasing substrate ((R)-(+)- or (S)-(–)-N,N’-dimethyl-1-
ferrocenylethylamine) concentration in the 0.001–0.004 M range.
Statistical tests for peak potentials were performed by repeatedly recording the CV pattern of each
enantiopure probe, at constant working protocol.
Enantioselective electrooligomerizations with (R)-BT2T4 and (S)-BT2T4 dissolved in bulk CILs (6a,
6b) were carried out, with the same above protocol, in the 0-1.1 V vs Ag+|Ag range by cycling the
oxidative potential for 36 times.
2.7.3 Electrochemical impedance spectroscopy tests EIS spectra were also recorded in selected CIL
media, both to compare their charge transport properties and to achieve complementary information
about their enantiodiscrimination ability by a possible alternative approach. The measurements have
been carried out in two-electrode setup mode, spreading 0.020 cm3 of the sample on a flat
conductivity cell with two concentric graphite electrodes (B&C Electronics SZ 3271 with a cell
constant of 1 cm-1), with no applied potential, the frequency range being 106-0.1 Hz. The tested
samples included (a) CILs 6a, 6b and 6d as such and in the presence of either enantiomer of the
chiral probe BT2T4; (b) achiral IL BMIMTf2N as such and in the presence of either enantiomer of
the chiral probe BT2T4; (c) 0.1 m KCl, for comparison’s sake (in this case the cell was immersed in
the aqueous solution).
11

ACCEPTED MANUSCRIPT
3. Results and discussion
3.1. Synthesis
Imidazolium and pyrrolidinium based ILs were obtained through a two-step sequence, i.e.
transformation of the primary hydroxyl group of the selected terpenoid alcohols into good leaving
groups, either mesylates or chloride, followed by the Menshutkin reaction, using N-
methylimidazole or N-methylpyrrolidine as a nucleophile. (Scheme 1) The isolated mesylate or
chloride ILs have then been transformed into the corresponding bistriflimide salts 6a-d (liquids at
room temperature) and 7b-c (solids at room temperature) by anion metathesis.
Scheme 1. Synthetic approach for the preparation of the terpenoid-based CILs
12

ACCEPTED MANUSCRIPT
Natural alcohols 1a-d used in this investigation can be classified into allylic and non-allylic
alcohols. It is noteworthy that when subjected to the same reaction with methanesulfonyl chloride,
allylic or non-allylic alcohols furnished the chloride or the methansulfonate derivatives,
respectively. This outcome has been observed before [23]. The higher reactivity of the allylic
mesylates probably favor the subsequent in situ substitution reaction leading to the corresponding
chloride derivatives in good yields.
3.2. ORD and CD spectra
The chiroptical properties of the investigated CILs are accounted for by the optical rotatory
dispersion spectra and electronic circular dichroism spectra reported in the Figure at page S17 for
the whole series with the exception of 6d (for which a chiroptical characterization is available in
[18]). Dealing with natural product derivatives, they are available as single enantiomers. In
particular, all of them are levorotatory, and all of them are ECD active, but at very low wavelengths
(<225 nm), which is consistent with the large HOMO-LUMO gaps observed in the CV experiments
(see later on).
3.3. Thermal analysis
3.3.1. TGA
The thermal stability of ILs has often been evaluated using thermogravimetric analysis TGA at a
single linear heating rate in controlled atmosphere. Although the onset temperature (calculated from
the TGA curves obtained from single temperature-ramp experiments) generally overestimates the
long-term thermal stabilities of ILs, this parameter can be used to analyse the relative thermal
stability of a set of ILs [25].
Fig. 2 collects a synopsis of TGA characteristics obtained for the six bistriflimide CILs plus 8c
(methanesulphonate analogue of 7c, with melting point above 100°C) under N2 flux with a heating
rate of 10 °C min-1.
13

ACCEPTED MANUSCRIPT
While commercial butylmethylimidazolium bistriflimide BMIMTf2N (a TGA of which is also
reported for comparison's sake, Fig. 2 top right) features a single, sharp degradation process with
onset at about 430 °C, the four ILs 6a, 6b, 6c and 6d (Fig. 2 left column), only differing from
BMIMTf2N for bearing one of the a,b,c,d natural unsaturated residuals instead of a simple butyl
chain, appear significantly less thermally stable. In fact, they all feature two subsequent degradation
steps, a first one (of smaller area) with onset located in the 210-330 °C range (i.e. much lower than
the onset of BMIMTf2N as well as other common n-alkylmethylsubstituted imidazolium and
pyrrolidinium salts [25,26]), preceding a second one of larger area, located close to the BMIMTf2N
onset. The pattern and location of the first degradation step looks rather similar for imidazolium and
pyrrolidinium ILs having the same natural residual (6b /7b, second row; 6c / 7c, third row).
The former observations point to the first degradation step being connected with the unsaturated
natural residual. In this frame, the onset temperature of the first step appears modulated by the
natural residual, the increasing thermal stability order being 6b,7b from (S)-perillyl alcohol < 6c,7c
from (1R)-nopol < 6a from (1R)-myrtenol < 6d from (S)-citronellol.
Only IL 6d can be classified as moderately stable (300 °C ≤ Tonset < 350 °C), while all the other
ones are definable as least stable (Tonset < 250 °C) or less stable (250 °C ≤ Tonset < 300 °C) [25]. For
cases 6a, 6b and 7b the higher lability could be due to the possible formation of relatively stable
allyl cations (E1 mechanism) or allyl radicals, whereas nopol derivatives 6c, 7c (both featuring a
significantly different, more complex first degradation step) could favour elimination reactions
involving the allylic proton. Instead the degradation mechanisms hypothesized for myrtenol, perillyl
alcohol and nopol derivatives cannot work for 6d, in which the double bond is located at an
adequate distance from the imidazolium moiety, making it the most stable derivative, with the first
degradation step very close and partially merging to the second one.
The anion should also play a dramatic role in triggering thermal degradation, as evidenced by the
comparison of the TGA patterns of 6c and 8c, consisting of the same cation combined with a
14

ACCEPTED MANUSCRIPT
bistriflimide or a methansulphonate anion, respectively. In both cases two degradation steps are
observed, but in the presence of methansulphonate (8c) 85% weight is lost in the first step.
Actually, at least with imidazolium cations, the bistriflimide anion often gives more thermally
stable ILs (> 350 °C) respect to other counteranions, due to its low nucleophilicity and basicity
(high temperatures are required for the decomposition of this anion to more nucleophilic atoms or
groups capable of attacking the cation).
15

ACCEPTED MANUSCRIPT
6a
6b
6c
BMIMTf2N
7b
7c
8c
6d
Fig. 2. TGA (green) and TGA derivative (blue) curves for the investigated CILs. TGA and TGA
derivative curve obtained for butylmethylimidazolium bistriflimide BMIMTf2N are also reported
for comparison.
16

ACCEPTED MANUSCRIPT
3.3.2. DSC
Among the seven compounds tested, only the four imidazolium salts with bistriflimide anions are
liquid at room temperature. For a more detailed characterization of thermal transitions in the whole
series, an exhaustive DSC study has been carried out at 10°C/min in nitrogen flow, according to the
following protocol:
•
•
•
•
•
•
cooling from 25 °C to -85 °C;
heating from -85°C to +200 °C (i.e. before any decomposition onset, see former paragraph);
cooling from 200 °C to -85 °C;
heating from -85°C to +25°C;
cooling from 25 °C to -85 °C;
heating from -85°C to +200 °C.
A synopsis of DSC characteristics is provided in Fig. 3 (further examples are provided in Fig. SI
4.1), including on the left the DSC patterns of the four salts that are liquid at room temperature, and
on the right those of the three salts that are solid at room temperature.
Only in one case (8c, the only IL with methanesulfonate anion in our series) the observed behaviour
is that of a solid salt, with a sharp endothermic melting peak in the heating half cycle at a
reproducible temperature above 100°C (120°C (I cycle)/117 °C (II cycle)), corresponding to a
splitted freezing peak (of area, i.e. enthalpy, globally comparable to the melting one) in the cooling
half cycle at 88-85 °C, accounting for a significant supercooling effect; no other significant
transitions are observed at lower temperatures.
Upon changing the anion from methanesulfonate to bistriflimide the salt becomes a room-
temperature ionic liquid RTIL (6c), with low tendency to crystallization, with a striking change in
its DSC characteristics. In fact only a reversible glass transition (corresponding to much smaller
enthalpies than in the former case; compare the y scales in the plots) can be reproducibly observed
17

ACCEPTED MANUSCRIPT
at about -50°C as an endothermic small peak in all heating half cycles, while in the cooling ones a
comparably small exothermic peak can be observed in the first cycle and apparently depends on the
sample history (it disappears in subsequent cycles, and it is not displayed by a more aged sample as
in some SI 4 examples). An intriguing large exothermic transition (with no associated weight loss,
as checked by TGA) is observed at 120-185 °C, which however disappears after the first heating
cycles although the glass transition remains reproducible.
A similar behaviour is displayed by the remaining three RTILs (all of them also including an
imidazolium cation and a bistriflimide anion): a glass transition endothermic small peak is
reproducibly observed in the heating half cycles, while the corresponding exothermic peak in the
cooling half cycles can be observed or not depending on the temperature cycling and sample
history. The glass transition temperature is approximately the same for 6a, 6b and 6c (including
cyclohexene moieties) while it is significantly lower for the open-chain derivative 6d, consistently
with higher structural disorder.
7b and 7c, differing from RTILs 6b and 6c for the pyrrolidinium cation moiety replacing the
imidazolium one, are solid at room temperature but could still be classified as ILs (not RTILs) since
they both have their melting points below 100°C. In particular,
•
7b has a reproducible thermal behaviour with two subsequent reversible transitions, one at
+85°C (endo)/+65(exo) and one at -24°C (endo)/-31°C(exo); they could correspond to
melting/freezing and solid/solid transitions, respectively;
•
7c, consistently with its featuring a more structure-breaking bicyclic moiety with respect to
7b, has a polymorph behaviour depending on the thermal history, showing a sharp melting
peak at 85°C in the first heating ramp, with no freezing partner in the subsequent cooling
step; a reversible glass transition is instead observed, followed by a melting peak of
significantly smaller area and at a significantly lower temperature.
18

ACCEPTED MANUSCRIPT
6a
7b
6b
6c
7c
8c
6d
Fig. 3. DSC curves for the investigated ILs; other examples obtained on samples more aged and/or
of different weight are reported in Fig. SI 4.1.
19

ACCEPTED MANUSCRIPT
3.4. Electrochemistry
The seven BioBased molecular salts have been electrochemically investigated by cyclic voltam-
metry on glassy carbon GC electrode in acetonitrile ACN with 0.1 M TEABF4 as supporting elec-
trolyte. A synopsis of CV features is reported in Fig. 4, key parameters being collected in Table 1.
First reduction peak systems are very similar in potential and shape for all salts with imidazolium
cations (6a-d, 8c), resembling the one exhibited by BMIMTf2N, reported in SI 5; instead they are
shifted to potentials more negative by 0.2/0.3 V for both salts with pyrrolidinium cations (7b,7c).
Thus, in the four imidazolium salts first reduction must involve the imidazolium site. Comparing
the CV patterns of BMIMTf2N, BMIMBF4, BMIMPF6 (SI 5), the counteranion might be somehow
involved in the cathodic process, too; on the other hand, 8c, with methanesulphonate anion, has a
CV pattern even more similar to the BMIMTf2N one than 6a-d, having Tf2N counteranions. In any
case, the process leads to accumulation of an electroactive product on the electrode (irregular
oxidative peaks at about -0.75 V vs (Fc+/Fc), particularly evident at high scan rates, SI 6). The
possible reduction mechanisms of imidazolium cation, including dimer formation, as well as of
Tf2N anion have been recently discussed by De Vos et al. [27].
On the other hand, unlike simpler BMIM salts (SI 5), first oxidation peaks can be clearly observed
before the background, with peak potentials significantly depending on the natural a,b,c,d scaffold
moiety, with little influence of the imidazolium/pyrrolidinium residuals, in the sequence:
6c, 7c, 8c (1.35-1.41 V) < 6d (1.45 V) < 6a (1.59 V) < 6b,7b (1.76-1.77 V)
Therefore, such first oxidation should involve alkene oxidation. Actually, according to the literature
[28], alkenes can undergo oxidation in such a positive potential range, and their oxidation potentials
can be significantly modulated by the surrounding molecular structure, even by several hundreds of
mV.
Differences between CIL first oxidation and first reduction peak potentials provide an indication
20

ACCEPTED MANUSCRIPT
about their potential windows when used as chiral media or additives for electrochemical
experiments, although it must be remembered that for bulk CILs both negative and positive
backgrounds would be closer.
In the present case, such windows (Fig. 4, top right) are wide on both negative and positive sides,
and their global widths (4.1-4.8 V) are significantly modulated by the molecular structure.
Table 1. CV peak potentials of the investigated salts ~0.00075 M in ACN+0.1 M TEABF4.
EpC vs (Fc+/Fc)
EpA vs (Fc+/Fc)
EpIA− EpIC
-2.74 V
-2.97 V
+1.59 V
4.33 V
6a
6b
7b
6c
-2.77 V
+1.76 V (shoulder)
+1.77 V (shoulder)
+1.41 V
4.53 V
4.82 V
4.22 V
-3.00V
Shoulder before background
-3.05 V
-2.81 V
-3.00V
Background -3.1 V
-2.80 V
+1.37 V
+1.35 V
4.47 V
4.15 V
7c
8c
-2.96 V
-2.83 V
+1.45 V
4.28 V
6d
-2.99V
21

ACCEPTED MANUSCRIPT
6a
7b
7c
8c
6b
6c
6d
Fig. 4. CV patterns obtained at 0.2 V/s for the seven salts ~0.00075 M in ACN + 0.1 M TEABF4
(starting oxidative and reductive half cycles, red and blue; background is also reported in grey).
22

ACCEPTED MANUSCRIPT
3.5. Chiral enantiodiscrimination voltammetry tests
The enantiodiscrimination ability of 6a and 6b (RTILs) and 7c (IL with melting point above room
T) has been tested in CV experiments, at first with (R)-(+)- or (S)-(−)-N,N’-dimethyl-1-ferro-
cenylethylamine, which we usually employ as a model chiral probe, commercially available in both
enantiomers and giving chemically and electrochemically reversible CV peaks at mild potentials.
The three above salts were tested as chiral additives in bulk BMIMTf2N (having the same
counteranion), and the two RTILs also as bulk media. The tests were performed in open air on
commercial disposable screen printed cells, on which a drop of chiral medium with 0.002 M (R)-
(+)- or (S)-( −)-N,N’-dimethyl-1-ferrocenylethylamine was deposited. The same protocol was also
applied to bulk achiral BMIMTf2N, for comparison's sake. Repetitions on new SPE supports were
performed in all cases in order to check the result repeatability.
Selected results are shown in Fig. 5. Statistically significant differences, of the order of several tens
of mV, are observed in all cases for the formal potentials of the ferrocenyl probe enantiomers
(calculated as the average of forward and backward CV peaks, and corresponding to standard
potentials neglecting activity coefficients), with the (R)- enantiomer being activated before the (S)-
one in all cases. We also verified that, with the same protocol, the (R)- and (S)- probe enantiomers
give practically coincident peak potentials in achiral BMIMTf2N in the absence of the chiral
additive (thin lines in Fig. 5). Reasonably, in bulk CIL media the enantiomer peaks are farer away
from the benchmark ones in achiral BMIMTf2N than in the cases in which the same CILs are used
as additives in bulk BMIMTf2N.
Partial linear dynamic range tests were carried out in bulk CILs 6a and 6b, confirming the linear
dependence of peak currents on the concentration of both probe enantiomers in our working range
(Fig. 6), with potential differences slightly increasing with concentration.
To test the general character of the media enantiodiscrimination ability, enantioselectivity tests in
bulk 6a and 6b were extended to the enantiomers of a quite different probe, the thiophene-based
BT2T4 monomer, "inherently chiral" and with free homotopic terminals enabling oxidative regiore-
23

ACCEPTED MANUSCRIPT
0.06 M 6a in bulk BMIMTf2N
bulk 6a
∆E°'/V = 0.037±0.009
Confidence interval tspool√ ((N1+N2)/N1N2)
= 0.002 (95%), 0.004 (99.7%)
∆E°'/V = 0.047±0.007
Confidence interval tspool√ ((N1+N2)/N1N2)
= 0.002 (95%), 0.004 (99.7%)
0.05 M 6b in bulk BMIMTf2N
bulk 6b
∆E°'/V = 0.035±0.002
Confidence interval tspool√ ((N1+N2)/N1N2)
= 0.003 (95%), 0.009 (99.7%)
∆E°'/V = 0.050±0.03
Confidence interval tspool√ ((N1+N2)/N1N2)
= 0.037 (90%), 0.052 (95%)
0.05 M 7c in bulk BMIMTf2N
∆E°'/V = 0.04±0.01
Confidence interval tspool√ ((N1+N2)/N1N2)
= 0.01 (95%)
Fig. 5. Enantiodiscrimination tests with (S)- or (R)- ferrocenyl chiral probes (red and green). Thick
lines: CV patterns in the presence of the chiral selector, either as additive in bulk achiral
BMIMTf2N (left) or as bulk ionic liquid (right). Thin lines: CV patterns in bulk BMIMTf2N without
chiral additive, for comparison.
24

ACCEPTED MANUSCRIPT
0.001 M (S)-Fc
0.001 M (R)-Fc
0.002 M (S)-Fc
0.002 M (R)-Fc
0.004 M (S)-Fc
0.004 M (R)-Fc
6a
0.001 M (S)-Fc
0.001 M (R)-Fc
0.002 M (S)-Fc
0.002 M (R)-Fc
0.004 M (S)-Fc
0.004 M (R)-Fc
6b
Fig. 6. Linear dynamic range tests on SPEs supports for (R)- or (S)- ferrocenyl chiral probes (green
and red lines, respectively) in bulk CILs 6a and 6b.
gular electrooligomerization, yielding the above mentioned enantiopure inherently chiral electroac-
tive films of outstanding enantiodiscrimination ability [24]. As shown in Fig. 7, significant differen-
ces (50-70 mV), even larger than the former ones, are observed for the first oxidation potentials of
the (R)- and (S)-probe enantiomers. (Fig. 7 left) Moreover, performing multiple oxidative CV
cycles, (Fig. 7 right) electrooligomerization of the two enantiopure monomers is obtained, with
significantly different last cycles.
25

ACCEPTED MANUSCRIPT
(S)-BT2T4 in 6a, I cycle
(R)-BT2T4 in 6a, I cycle
(S)-BT2T4 in 6a, XXXVI cycle
(R)-BT2T4 in 6a, XXXVI cycle
∆E1/2 = 0.05 V
6a
(S)-BT2T4 in 6b, I cycle
(R)-BT2T4 in 6b, I cycle
∆E1/2 = 0.07 V
6b
Fig. 7. Enantiodiscrimination tests in in bulk CILs 6a and 6b with (S)- or (R)- BT2T4 inherently
chiral probes (red and green, respectively), undergoing electrooligomerization upon subsequent
oxidative cycles. Left: first cycle; right: XXXVI cycle.
Such results are quite interesting since the highly desirable target of enantiodiscrimination in terms
of potential differences instead of current differences has been so far only seldom obtained in chiral
electroanalysis, notwithstanding many adopted strategies [20]. To our knowledge this is the first
26

ACCEPTED MANUSCRIPT
chiral voltammetry test using CILs as chiral selectors, besides the very recent, striking results
obtained by some of us using ionic liquids or related solid salts with "inherently chiral" dications,
based on bicollidinium [21] or bibenzimidazolium [22] atropisomeric scaffolds with axial C2
symmetry. In that case very large potential differences were observed, of the order of 100-250 mV,
for the enantiomers of chiral probes of different structure and reactivity, even using the ICILs or
related salts as diluted additives in bulk achiral IL.
To explain such enantiodiscrimination performances, we focused on the peculiar structure of the
enantiopure chiral interphase of (R)- or (S)-configuration, at which the electron transfer takes place
involving either (R)- or (S)-chiral probe antipode, resulting in diastereomeric and thus energetically
different conditions. The local structure of an ionic liquid at the interphase with a charged electrode
surface, even in the presence of significant water traces [29], is known to be highly ordered for
many layers [10], somehow resembling a semisolid crystal or liquid crystal. Such electrode|CIL
interphase must be much more effective in transmitting the chiral information than an
electrode|solution classical double layer interphase with chirality implemented in the solvent or
supporting electrolyte [20]. Moreover, since the CIL interphase structure is modulable by the
presence of extra species [30], we assumed that adding to a bulk achiral IL a chiral additive, like the
above ICILs, could result in a local chiral structuring effect similar to the well-known nematic
(achiral) to cholesteric (chiral) structure transition easily induced by chiral additives in bulk liquid
crystals [31]. This is very interesting, since it would allow to obtain enantiodiscrimination also
without having to work in bulk CIL. Moreover, although CIL/probe complexation in bulk solution
was excluded, local specific interactions, e.g. involving heteroatoms as well as (hetero)aromatic
rings, between chiral additive and chiral probe at the electrode/solution interphase cannot be
excluded. Finally, very recent magnetoelectrochemistry experiments by some of us suggest to
evaluate also the possibility of spin-dependent potential modulation at interphases with molecular
spin filter features.
Such explanations could also apply to both the previous ICIL and the present CIL case; however, by
27

ACCEPTED MANUSCRIPT
analogy with the chiral film examples [20,32], the inherently chiral additives should be particularly
efficient on account of (a) the chirality source coinciding with the whole main cation backbone,
rather than in localized stereocentres, (b) the high regioregularity granted by the cation C2 axial
symmetry, which should enhance its structure ordering ability. It was however important and
overdue to verify whether similar effects, albeit possibly smaller, could be observed using chiral,
rather than inherently chiral, ionic liquids as chiral selectors, particularly concentrating on chiral
cations and bistriflimide anions (used to obtain the above ICILs). The present results provide a first
answer, confirming the outstanding efficiency of the inherent chirality respect to non inherent
chirality strategy, and providing a further clue for the development of a reliable interpretative
scheme for electrochemical enantiodiscrimination in chiral ionic liquid media.
3.6. Chiral enantiodiscrimination EIS tests
The features of CILs 6a, 6b and 6d were also investigated by electrochemical impedance
spectroscopy, in two-electrode setup mode, spreading a drop of the CIL sample on a flat cell with
two concentric graphite electrodes.
The CILs were investigated both as such, to compare their charge transport properties, and with
addition of either (S)- or (R)- enantiomer of the BT2T4 monomer, to test EIS as alternative approach
to achieve complementary information about the media enantiodiscrimination ability.
Fig. 8 collects the Nyquist and Bode EIS spectra obtained in the absence of polarization for CIL 6d,
for which we could test both the enantiopure (−) antipode and the (±) racemate. As a comparison,
EIS spectra are also provided of achiral IL BMIMTf2N, lacking the chiral biobased building block.
Since no electron transfer process takes place in the operating conditions, the IL Nyquist plots can
be interpreted following the exhaustive treatment recently proposed by Pilon and coworkers for EIS
spectra of electric double layer capacitor EDLC electrodes and devices [33]. Accordingly to such
model, the diameter of the first semicircle corresponds to bulk electrolyte resistance, the first
28