Ionic liquids as modulators of fragrance release in consumer goods

Article abstract of DOI:10.1039/C6NJ01626J

When organic esters or alcohols were dissolved in each of three novel ionic liquids (which have no effective vapour pressure), the vapour-liquid equilibria (as measured by infrared spectroscopy of the gas phase) revealed significant positive deviation from Raoult's law for a wide range of perfume raw materials. The addition of water amplified the repulsive effect of the ionic liquid matrix, and this was exemplified by a series of ternary phase diagrams.

Full text of DOI:10.1039/C6NJ01626J

NJC
View Article Online
View Journal
PAPER
Ionic liquids as modulators of fragrance release in
consumer goods
Cite this:DOI: 10.1039/c6nj01626j
a
b
a
Federico M. Ferrero Vallana, Ricardo P. Girling, H. Q. Nimal Gunaratne,
b
b
a
Lynette A. M. Holland,* Pauline M. Mcnamee, Kenneth R. Seddon,*
Jonathan R. Stonehouse and Oreste Todini
b
c
Received (in Nottingham, UK)
2
3rd May 2016,
When organic esters or alcohols were dissolved in each of three novel ionic liquids (which have no
effective vapour pressure), the vapour–liquid equilibria (as measured by infrared spectroscopy of the gas
phase) revealed significant positive deviation from Raoult’s law for a wide range of perfume raw materials.
The addition of water amplified the repulsive effect of the ionic liquid matrix, and this was exemplified by a
series of ternary phase diagrams.
Accepted 1st October 2016
DOI: 10.1039/c6nj01626j
www.rsc.org/njc
functional groups. The evaluation of the fragrance performance
in consumer goods is a function of three empirical parameters:
Introduction
Ionic liquids are a class of salts, consisting typically of a large long longevity, substantivity, and diffusivity. These three factors
asymmetrical organic cation and either an organic or inorganic refer to the property of a fragrance to be perceivable on a surface
anion. These can be selected and synthesised with a wide after several hours, the ability to smell that fragrance even after a
variety of structures depending on the specific technology to drying process (i.e. in laundry application), and its capacity for
1
which they will be applied. One intriguing new potential being perceived at a further distance from its source during a
8
application for these materials is controlled fragrance release, reasonably long time period (as in an air fresheners).
especially in household products. For example, ionic liquids
2
Most perfumes start off their life as a liquid mixture for which
have been used as ‘‘fixatives’’ with fragrance compositions to perfumers do not know the exact composition of the headspace
lower the rate of evaporation of a perfume component, or to (defined as ‘‘the air or empty space left above the contents in a
3
9
impart increased stability/longevity of the fragrance. Recently, sealed container’’). In order to gain a greater understanding of
ionic liquids have also been used as pro-fragrances, where a the nature of the vapour phase, and its relationship with the
perfume raw material has been appended covalently to either composition of the liquid phase, it is necessary to investigate
4
the cation or the anion, and for the synthesis of floral notes the thermodynamic equilibria which define the phases. In other
such as d-damascone, which is widely use in the detergent words it is necessary to measure the vapour–liquid equilibria
5
6
industry. Elsewhere, Forsyth et al. investigated the utilisation (VLE) and liquid–liquid equilibria (LLE) of these rather complex
of ionic liquids for the synthesis of the Lily-of-the-Valley mixtures. Headspace analysis is required to provide unambiguous
s
fragrance {b-Lilial ; 3-(4-t-butylphenyl)-2-methylpropanal}.
analytical data. In this context, volatility is a key concept which
Fragrances are added to many successful consumer goods in can be defined as the ‘‘readiness to vaporise or evaporate,
order to impart a long-lasting, pleasant, and product-appropriate tendency to be readily diffused, or dissipated in the atmo-
7
9
odour. The perception of this odour depends upon the concen- sphere, especially at ordinary temperatures’’, and according
tration and nature of the fragrance molecule released into the to the functionalities of the molecule this process happens
air. The design of selective and efficient delivery systems has to at different rates. According to this principle, fragrances can
be able to control the release of highly volatile fragrances be classified as highly volatile (which have vapour pressures
in functional perfumery products, and to increase the stability greater than about 0.13 mbar at 25 1C), moderately volatile
of perfume raw materials (PRMs) with unstable chemical (between about 0.040 mbar to about 0.13 mbar), and poorly
volatile (less than about 0.040 mbar). Based on the different
amount of perfume raw materials in the vapour phase, the
character of the fragrance can be composed of multiple notes.
For example, highly volatile molecules tend to smell citrusy,
green, light and fresh, and they may be noticeable for only a few
minutes after being applied to a substrate. In contrast, the
moderately volatile PRMs are characterised by floral or fruity
a
b
c
10
The QUILL Centre, The Queen’s University of Belfast, Stranmillis Road, Belfast,
BT9 5AG, UK. E-mail: quill@qub.ac.uk
Procter & Gamble London Innovation Centre, Whitehall Lane, Egham, Surrey,
TW20 9NW, UK
Procter & Gamble Brussels Innovation Centre, Temselaan 100, Strombeek-Bever,
Brussels, B-1853, Belgium
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
NJC
notes, which may be detectable for up to several hours. The
poorly volatile compounds are typically sweet, musky or woody,
and can last for several days. This variety of evaporation rates is
not suitable for many fragrance compositions, and can limit
the development of new perfumes. For example, the design of a
fragrance with PRMs that evaporate in three hours after applica-
tion has to be restricted to the highly volatile molecules characters.
Creation of fragrance profiles is dominated by the moderately
or poorly volatile perfume raw material characters. Thus, it is
advantageous to be able to create fragrance compositions that
span a wider range of fragrance characters, whilst still having
all or substantially all of the scents evaporated within a given
time frame.
The purpose of this current work is to describe the use of
ionic liquids (which have no effective vapour pressure) as novel
platforms for dissolving and releasing perfume raw materials,
and to model their evaporation profile. The current work suggest
1
1
a novel idea applicable to perfume product engineering, and
the structures have been selected with the target of human
friendly media for consumer goods applications.
Experimental
Materials
Fig. 1 Structures of materials used in this work: (N-ethyl-2-(2-methoxy-
ethoxy)-N,N-dimethylethanaminium) 6-methyl-3,4-dihydro-1,2,3-oxathiazin-
4
-one 2,2-dioxide, [N1 1 2(2O2O1)][ace]; (N-ethyl-2-(2-methoxyethoxy)-
Ionic liquid precursors. The three new ionic liquids, namely
N,N-dimethylethanaminium) 1,4-bis(2-ethylhexoxy)-1,4-dioxobutane-2-
(N-ethyl-2-(2-methoxyethoxy)-N,N-dimethylethanaminium) 6-methyl- sulfonate, [N112(2O2O1)][aot]; and N-benzyl-N,N-dimethyloctan-1-aminium
3
,4-dihydro-1,2,3-oxathiazin-4-one 2,2-dioxide, [N112(2O2O1)][ace], 1,4-bis(2-ethylhexoxy)-1,4-dioxobutane-2-sulfonate, [N118Bz][aot]; (1-cyclo-
hexyl-2-methylpropan-2-yl) butanoate, chmpb; and 2,2-dimethyl-3-(3-
(N-ethyl-2-(2-methoxyethoxy)-N,N-dimethylethanaminium) 1,4-bis(2-
methylphenyl)propan-1-ol, dmmpp.
ethylhexoxy)-1,4-dioxobutane-2-sulfonate, [N112(2O2O1)][aot], and
N-benzyl-N,N-dimethyloctan-1-aminium 1,4-bis(2-ethylhexoxy)-1,4-
dioxobutane-2-sulfonate, [N1 1 8 Bz][aot], were synthesised (vide
infra) using the following reagents (ex. Aldrich, 99%) without
further purification: 2-(2-methoxyethoxy)ethanol, thionyl chloride,
pyridine, N,N-dimethylethylamine, N,N-dimethyloctylamine,
benzyl chloride, sodium bis(2-ethylhexyl) sulfosuccinate, and
potassium 6-methyl-1,2,3-oxathiazin-4(3H)-one 2,2-dioxide. Dichloro-
methane and methanol were dried using standard methods
prior to use.
Thermodynamics study. The ionic liquids used in this work
(
see Fig. 1) are: IL
and IL , [N118Bz][aot]. Other materials used include (1-cyclohexyl-
-methylpropan-2-yl) butanoate (chmpb), and 2,2-dimethyl-3-
3-methylphenyl)propan-1-ol (dmmpp).
A B
, [N112(2O2O1)][ace]; IL , [N112(2O2O1)][aot];
C
2
(
Methods
A B C
The ionic liquids IL , IL , and IL were synthesised according
to the schemes reported in Fig. 2 and 3, and characterised
1
using H NMR spectroscopy, TGA, elemental analysis, and
electrospray (ethanenitrile) mass spectrometry. The ionic liquids
IL , IL , IL and the molecules dmmpp and chmpb were studied
A
B
C
by examining the vapour–liquid equilibria (VLE) and liquid–
liquid equilibria (LLE).
Synthesis of IL
a) Synthesis of 1-chloro-2-(2-methoxyethoxy)ethane. A solution
of 2-(2-methoxyethoxy)ethanol (50.00 g, 0.416 mol) and pyridine the tertiary amine, and (c) and (d) anion metathesis.
A B
and IL
Fig. 2 Synthetic scheme for the preparation of [N1 1 2(2O2O1)][ace] and
(
[N1 1 2(2O2O1)][aot]: (a) synthesis of the chloroalkane, (b) quaternisation of
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016

View Article Online
NJC
Paper
in vacuo at 50 1C for 1 d (13.32 g, 83% of yield). Microanalysis:
calcd C, 58.26; H, 9.95; N, 2.34; S, 5.36: found C, 57.01; H,
+
1
M
0.3; N, 2.27; S, 4.70 ES-MS (CH CN): calcd M 176.1651,
3
ꢁ
ꢁ
+
161.9861: found M 176.1555, M 161.9668.
(d) Synthesis of [N1 1 2(2O2O1)][aot]. The same procedure was
followed as for [N1 1 2(2O2O1)][ace], except that [N1 1 2(2O2O1)]Cl
10.00 g, 0.0472 mol) sodium docusate (21.00 g, 0.0472 mol)
(
were used, yielding a pale yellow liquid (24.12 g, 85%). Micro-
analysis: calcd C, 58.26; H, 9.95; N, 2.34 S, 5.36: found C, 57.61;
+
H, 9.88; N, 2.41; S, 5.26 ES-MS (CH CN): calcd M 176.1651,
M
3
ꢁ
ꢁ
+
421.2260, found: M 176.1628, M 421.2107.
Syntheses of ILc
a) Synthesis of N-benzyl-N,N-dimethyloctan-1-aminium chloride.
(
A solution of the benzyl chloride (10.00 g, 0.0790 mol) and
N,N-dimethyloctylamine (12.42 g, 0.0790 mol) was added to a
screw cap glass tube. (The stirring speed was set at 500 rpm and
the temperature was set at 50 1C). The reaction mixture was left
in the sealed tube for 7 days before being stopped. The obtained
orange mixture was washed with ethyl acetate, cyclohexane and
concentrated via rotatory evaporator to yield a hygroscopic white
solid, which was subsequently dried under high vacuum at 40 1C
for 3 d. The final chloride ionic liquid was obtained in (24.12 g,
Fig. 3 Synthetic scheme for the preparation of [N1 1 8 Bz][aot]: (a) quater-
nisation of the tertiary amine, and (b) anion metathesis.
3
(
32.92 g, 0.416 mol) in dichloromethane (200 cm ) was added to a
three-necked round-bottom flask fitted with a condenser under a
dinitrogen flush. A solution of thionyl chloride (49.51 g, 0.416 mol)
in dichloromethane was then added dropwise over a period of 2 h
to the stirred mixture via a pressure equalising funnel at room
temperature. Once the addition was complete, the reaction mixture
was then heated under reflux for 12 h. The reaction mixture was
then cooled to room temperature, and washed with saturated
8
7
5%). Microanalysis: calcd C, 71.93; H, 10.65; N, 4.93: found C,
+
3
0.70; H, 10.99; N, 4.77; ES-MS (CH CN): calcd M 248.2378,
+
found: M 248.2375.
3
aqueous sodium hydrogencarbonate (200 cm ), and then magne-
(b) Synthesis of [N1 1 8 Bz][aot]. Sodium docusate (15.66 g,
sium sulfate (5 g) was added to the organic phase. This mixture was
then filtered and the filtrate was concentrated using a rotovaporator
in vacuo at 40 1C to leave a pale yellow liquid. A simple distillation at
0
.0352 mol) was added to a solution of the chloride ionic liquid
1 1 8 Bz][Cl] (10.00 g, 0.0352 mol) in dichloromethane (15 cm ).
3
[N
The reaction mixture was left stirring at room temperature for
d before being placed in the centrifuge. The potassium
7
0 1C in vacuo produced a colourless liquid (56.02 g; 97% yield),
2
12
which was used immediately in step (b).
chloride by-product was fileted off and dichloromethane was
concentrated to afford a viscous pale liquid as the desire
product, which was then dried under high vacuum at 50 1C for
(
b) Synthesis of (N-ethyl-2-(2-methoxyethoxy)-N,N-dimethyl-
ethanaminium)chloride. A solution of the distilled 1-chloro-2-
2-methoxyethoxy)ethane (20.00 g, 0.144 mol), N,N-dimethyl-
1
d. The [N1 1 8 Bz][aot] was obtained in (21.56 g, 91%). Micro-
(
analysis: calcd C, 66.35; H, 10.08; N, 2.09 S, 4.79: found C, 66.55;
H, 10.15; N, 2.19; S, 4.94 ES-MS (CH CN): calcd M 248.2378,
3
M
ethylamine (10.55 g, 0.144 mol; newly opened bottle) and dry
+
3
methanol (50 cm ) was added in air to a screw cap glass tube.
ꢁ
+
ꢁ
421.2260, found: M 248.2366, M 421.2606.
The mixture was then stirred (700 rpm) at 50 1C for 15 days. The
mixture was cooled to room temperature and transferred to a
rotovaporator, and the methanol and the excess of amine were
removed under the reduced pressure. The orange residue was then
1
Instrumental. H NMR spectra were recorded on a Bruker
Avance III 400 MHz spectrometer. ESMS-mass spectroscopy
measurements were carried out on a Waters LCT Premier
instrument with an Advion TriVersa Nano Mate injection system
3
3
washed with ethyl ethanoate (20 cm ꢀ 3), cyclohexane (20 cm , two
(cone voltage 50 V, source 120 1C). Both positive and negative
times) and finally was concentrated using a rotovaporator in vacuo at
ions were detected. Samples were injected as dilute solutions in
ethanenitrile. DSC-scans (for melting point measurements) were
obtained using a TA DSC Q2000 model with a TA Refrigerated
Cooling System 90 (RCS). For each sample, three scans were run
40 1C to leave a hygroscopic orange solid, which was subsequently
dried in vacuo at 40 1C for 3 d (28.85 g, 93% of yield). Microanalysis:
calcd C, 51.05; H, 10.47; N, 6.62: found C, 50.24; H, 10.6; N, 6.55;
+
+
ES-MS (CH CN): calcd M 176.1651; found M 176.1624.
3
ꢁ1
with scan rates of 5 1C min . Microanalysis: carbon, hydrogen,
c) Synthesis of [N1 1 2(2O2O1)][ace]. To a solution of the nitrogen and sulfur contents (wt%) were determined using a
N1 1 2(2O2O1)]Cl (10.00 g, 0.0472 mol) in dichloromethane Perkin-Elmer Series II CHNS/O 2400 CHN Elemental Analyser,
(
[
3
(20 cm ), was added potassium acesulfamate (9.50 g, 0.0472 mol), which provided values within 0.3 wt% error. Water content was
and the mixture was stirred at room temperature for 2 d, whence determined using a Cou-Lo Aquamax Karl-Fisher Moisture Meter
it was centrifuged (3.5 rpm for 20 min) to remove the solid. (compact version 08.05). Ionic liquids were tested in duplicate
The solvent was removed using a rotovaporator in vacuo at 40 1C immediately after drying under high vacuum for 24 h at 50 1C.
to leave a pale yellow liquid, which was subsequently dried The water content was less than 60 ppm for all samples.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
Decomposition temperature measurements were performed
NJC
using a TGA/DSC1 thermogravimetric analyser from Mettler-
Toledo, Inc. The samples were measured in alumina crucibles,
ꢁ
1
at a heating rate of 5 K min under a dinitrogen atmosphere.
The onset of the weight loss in each thermogram was used as a
measure of the decomposition temperature (point at 5 wt%
loss of the sample). The melting points and glass transition
temperatures of the synthesised compounds were measured
by DSC, using a TA Instruments Modulated DSC Q 2000
V24.4 Build 116 with a refrigerated cooling system RCS 90
capable of controlling the temperature down to 220 K. Dry
dinitrogen gas was purged through the DSC cell at a flow rate of
3
ꢁ1
Fig. 4 The gas phase IR spectrum (15 scans at 1 scan per 10 s from 4000
to 1000 cm ) for chmpb at 30 1C.
ca. 20 cm min
.
ꢁ1
Determination of VLE. Calibration procedure: All the FT-IR
measurements reported were recorded using a PerkinElmer
Spectrum 100 spectrometer, fitted with a Cyclonet C5 series
gas cell of volume 1.33 l. The cell was thermostatically con-
trolled with an 8 m tuneable path length, and was linked to a
vacuum pump (Edwards RV8).
r
where C is the ratio between the absorbance of the chosen
characteristic stretching mode, in the case of chmpb nCO, in the
ionic liquid mixture (Amixture) over the absorbance of the same
stretching mode for the neat chmpb (Aneat).
Typically, prior to use, the cell was evacuated several times,
alternating with dry dinitrogen flushes, until the infrared
spectrum of the gas phase revealed the absence of any undesir-
able species. The sample was introduced into the cell and the
evacuation process was started until the pressure dropped
to 0.08 bar. The heating mantle was lowered over the cell,
and the system was left equilibrating at a defined temperature.
A spectrum was recorded periodically in order to check when
the system had reached equilibrium, whence the absorbance
was constant over time.
Determination of kinetics of H O uptake and saturation
2
point. The water content of the IL , IL , and IL has been
A
B
C
quantified using a Cou-Lo Aquamax Karl-Fisher Moisture Meter
compact version 08.05). The experiment consists of sampling
(
each ionic liquid (1 g) left in an open vial during the first 12 h in
order to establish a kinetic of absorption. The saturation points
for these mixtures have been reached after 24 h of air exposure.
Determination of LLE. The cloud point of a material was used to
measure the onset of liquid–liquid immiscibility. Cloud points were
determined at atmospheric pressure using visual detection of the
phase demixing (naked eye observation of turbidity followed by
phase separation). Samples of the solutions were prepared in a
For each point of the calibration line, a defined quantity of
chmpb was weighed. After recording the background, the
weighed sample was introduced into the cell. The cell was then
sealed, evacuated down to 0.08 bar, and then heated up to a
temperature (150 1C) which ensured the complete evaporation
of the sample for a defined equilibration time. Then the
spectrum was recorded (15 scans at 10 spectra per second from
13
closed Pyrex-glass vial with a stirrer inside. Ternary phase diagrams
were constructed from the cloud points of a mixture of an ionic
liquid phase (with a defined water content of either 0.55 wt% or
8
A A B A C
.00 wt%) – IL , (0.1IL + 0.9IL ),† or (0.1IL + 0.9IL )† – and
two organic molecules chmpb and dmmpp (see Fig. 1).
ꢁ
1
4
000 to 1000 cm ). The reproducibility was checked by repeat-
COSMO-RS for VLE prediction. The calculation of the activity
coefficient for mixtures of multi-component systems involved
two main steps: (a) the calculation of the lowest energy con-
formers of the molecules and ionic liquids using COSMO-conf,
and (b) calculation of the average activity coefficients of these
conformers of the multinary system of ionic liquids and the
organic materials. The strength and nature of the interaction
between the ionic liquids and the components involved is
indicated by the calculated Raoult’s plot from COSMO-therm,
and estimating the area of the deviation from the ideal straight
line. In the particular case involving ionic liquids, a pseudo-binary
approach was used in the COSMO-therm program (Version C3.0
Release 14.01) to calculate the LLE and VLE of a mixture
composed of an ionic liquid containing some water, with the
cation and anion of the ionic liquid input as equimolar
amounts of separate ‘‘compounds’’. The chemical potentials
were calculated for the system (anion, cation and water), with
ing the spectral accumulation three times.
Single component vapour pressure measurement. In order
to measure the concentration of chmpb in the vapour phase
above the liquid sample, a defined and recorded quantity of the
material was weighed into a vial and introduced into the FT-IR
cell. The system was evacuated and thermostatically main-
tained at a defined temperature in order to reach equilibrium,
and at this point the IR spectrum was recorded (see Fig. 4).
The two peaks considered for the quantification are those at
ꢁ1
ꢁ1
1
746 cm (due to the CO stretching) and at 2980 cm (due to
the CH stretching).
Multicomponent systems: The vapour pressure of the organic
A B
material chmpb in admixture with an ionic liquid (IL , IL or
IL ) was measured using the same procedure reported for the
C
single component. Here the vapour pressure has been expressed
in terms of the relative concentration C , as defined in eqn (1).
r
Amixture
†
Within this manuscript, the composition of ionic liquids mixtures is specified
Cr ¼
(1)
Aneat
by weight fraction unless otherwise specified.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016

View Article Online
NJC
Paper
1
H-NMR spectra confirmed the 1 : 1 molar ratio between the
anion and the cation and the absence of organic impurities.
1
1-Chloro-2-(2-methoxyethoxy)ethane.
H NMR (d /ppm,
H
4
00 MHz, CDCl ): 3.40 (s, 3H); 3.57 (m, 2H); 3.66 (m, 4H);
3
3
.76 (m, 2H).
1
[
N
1 1 2(2O2O1)]Cl. H NMR (d
H
/ppm, 400 MHz, CDCl
3
):
);
3
2
3
.98 (m, 2H, OCH
H, J = 4 Hz, NCH
.42 (s, 6H, (NCH
1 1 2(2O2O1)][ace], IL
CDCl ): 5.42 (s, 1H, CH[ace]); 3.94 (m, 2H, OCH ); 3.71
2
); 3.94 (m, 2H, OCH
); 3.67 (m, 2H, OCH
); 3.36 (s, 3H, OCH
2
); 3.80–3.79 (quartet,
); 3.52 (m, 2H, OCH
2
2
2
3
)
2
3
); 1.42 (t, 3H, CH
3
).
1
[
N
A
.
H NMR (d /ppm, 400 MHz,
H
Fig. 5 Glass slide perfume raw material as a VLE system.
Table 1 Odour intensity scale
3
2
(
m, 2H, OCH ); 3.64–3.62 (m, 4H, (OCH ) ); 3.51 (m, 2H,
2 2 2
NCH ); 3.34 (s, 3H, OCH ); 3.23 (s, 6H, (NCH ) ); 2.00 (s, 3H,
2
3
3 2
(
CH
3
) [ace]); 1.39 (t, 3H, (CH
3
)). T
g
= ꢁ67 1C.
. H NMR (d /ppm, 400 MHz, CDCl
d 4.12 (m, 1H, (CH) [aot]), 4.10–3.91 (m, 6H, (NCH ), 3.75
s, 2H, (CH ) [aot]), 3.65 (m, 4H, (OCH ), 3.54 (s, 2H, NCH ),
.38 (s, 3H, OCH ), 3.25 (s, 7H, (NCH ), 3.12 (m, 1H, (CH)
aot]), 1.74 (s, 2H, (CH ) [aot]), 1.65 (s, 1H, (CH) [aot]), 1.54
Score
Fragrance intensity
Score
Fragrance intensity
1
[
N
1 1 2(2O2O1)][aot], IL
B
H
3
)
0
1
2
Not detectable
Weak
Moderate
3
4
5
Strong
Very strong
Overpowering
3 2
)
(
3
[
2
2
)
)
3 2
2
2
3
2
the chemical potential of the ionic liquids as the sum of (s, 1H(CH) [aot]), 1.50–1.22 (m, 20H(CH ) [aot]), 0.89 (m, 12H,
2
the chemical potentials of both the cation and anion. The (CH ) [aot]). T = ꢁ81 1C.
2
6
g
1
COSMO-RS calculations and basis set parameterisation were
[N
1 8 Bz H 3
]Cl. H NMR (d /ppm, 400 MHz, CDCl ) d 7.70–7.64
1
1
4
performed at the BP/TZVP level (Turbomole, DFT/COSMO (m, 2H, H2,6), 7.50–7.40 (m, 3H, (ArCH) ), 5.09 (s, 2H,
3
1
5
with the BP functional and TZVP basis set), using the fully (NCH Ph)), 3.56–3.48 (m, 2H, NCH ), 3.32 (s, 6H, (NCH ) ),
2
2
3 2
optimised geometries at the same level of theory and parameter 2.02 (m, 2H, (CH ) ), 1.29 (m, J = 20.5, 13.0 Hz, 10H(CH ) ), 0.88
2
1
2 5
file BP TZVP C21 0025.
3 m
(t, J = 6.8 Hz, 3H, (CH )). T = 62 1C.
1
Olfactory tests. In order to show the effect of the ionic liquids on
[N1 1 8 Bz][aot] IL . H NMR (d /ppm, 400 MHz, CDCl ) d 7.58
C H 3
the perception of fragrance profile in a fragrance composition, test (s, 2H, H2,6), 7.44 (s, 3H, (Ar CH) ), 4.72 (s, 2H, (CH -Ar)), 4.16
3
2
compositions were made, and given to panellists of professional (m, 1H, (CH) [aot]), 3.97 (m, 4H, N(CH ) ), 3.31 (m, 3H, (CH )),
1
2 2
3
noses to evaluate.
3 2 1
3.15 (s, 7H, N(CH ) , CH), 1.60 (s, 1H, (CH) [aot]), 1.51 (s, 1H,
At the testing facility, samples (50 mL) of the fragrance (CH) [aot]), 1.45–1.17 (m, 26H, (CH ) [aot]), 0.93–0.79 (m, 15H,
1
2
compositions or the controls are applied to glass slides and (CH ) [aot]). T = ꢁ27 1C.
2
g
placed on a hot plate at 32 1C to represent skin temperature (see
Fig. 5 for a schematic). Expert panellists were asked to assess VLE measurements
the perceived fragrance profile (intensity and/or character)
Calibration line and vapour pressure of single component
from each pair of samples, i.e., that of the test composition
vs. the corresponding control, at time 0 and later time points
(
chmpb). Infrared spectra of chmpb (see Fig. 4) were recorded
over the temperature range 25–40 1C. The absorbances of
two specific stretching modes of chmpb ( n~ CO = 1746 cm
n~ CH = 2980 cm ) were selected, and plotted against its concen-
tration in the vapour phase (see Fig. 6) and the data are
reported in Table 2.
(e.g., 1, 3, 6, 8 and up to 24 h post application) as the fragrance
ꢁ1
;
profile evolves. The panellists were asked to give a score on a
scale of 0 to 5 for perceived fragrance intensity according to the
odour intensity scale set out in Table 1.
ꢁ1
From the slope of the Beer–Lambert plot, eqn (2), reported in
1
8
Fig. 6, and using the ideal gas law, eqn (3), the vapour pressure
of the standard material has been calculated according to eqn (4)
for each data set.
Result and discussion
Ionic liquid synthesis and characterisation
1
9
A = elc
(2)
(3)
The ionic liquids IL
A B C
, IL and IL (see Fig. 1) were selected
taking into account very preliminary and limited toxicology
1
6
PV = nRT
information. For this reason, new-to-the world ionic liquids
were designed to fit these criteria, and these are described in
detail here. Although a similar study with [C mim][NTf ] could
A
P ¼ elRT
(4)
4
2
have been performed, it would not have been applicable to
1
7
consumer goods based on available toxicity data. Clearly, for where e = molar extinction coefficient, A = absorbance, l = path
this study, a high degree of purity was required. The H-NMR length (in cm), c = concentration (in mol l ), P = pressure,
see below) and ES-MS data confirmed the identity and purity V = volume, n = number of moles, R = gas constant and
1
ꢁ1
(
of these novel liquids. In particular, the integration of the T = temperature.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
NJC
0
i
where, P
5 1C, and w
The data reported for the system [N1 1 2(2O2O1)][ace] and
is the vapour pressure of the pure component at
i
is its mole fraction.
2
chmpb show a pronounced positive deviation from ideality for
this mixture. In particular, at low mole fraction (0.05) of chmpb,
this behaviour suggests a strong mutual repulsion between the
involatile ionic liquid and the volatile organic. The calculated
values (COSMO-therm) for this mixture anticipate the strong
repulsion observed in the experimental data, and also predict
the maximum effect occurring at a mole fraction about 0.2.
The second ionic liquid tested, IL , has the same cationic
B
structure, but the acesulfamate anion has been replaced with
docusate anion, in order to increase the solubility of the hydro-
phobic organic molecules: VLE data are reported in Fig. 8. Both
Fig. 6 Beer–Lambert plots for chmpb. The data were fitted by
linear regression and forced to go through the origin. For n~ CO (blue line),
ꢁ1
ꢁ1
2
gradient = 95 300 l mol cm , R = 0.9693; for n~ CH (red line), gradient =
2 900 l mol cm with R2 = 0.9751. The standard deviation in A for the empirically and computationally, the repulsive effect here is less
ꢁ1
ꢁ1
4
n~ CO is 0.0421, and for the n~ CH is 0.0152. The errors reflected in the
standard deviations arise from the precision of the balance for weighing
the samples.
A
than for IL , and there is a closer match between the theoretical
prediction and the experimental observation.
In order to create a combination between the better solubility
properties of the docusate anion and the greater repulsive effect
of [N1 1 2(2O2O1)][ace], a mixture of the two ionic liquids has
Table 2 Calculated vapour pressure from the infrared absorbance of
been studied. The mole ratio used was IL : IL = 1 : 1 and the
ꢁ1
ꢁ1
A
B
chmpb ( n~ CO = 1746 cm and n~ CH = 2980 cm ) as a function of its gas
phase concentration
a
VLE data are given in Fig. 9. It is evident that the deviation from
Raoult’s law for chmpb is a combination of the two curves shown
in Fig. 7 and 8, demonstrating consistency between the experi-
mental and theoretical data.
The fourth mixture tested was the ionic liquid [N1 1 8 Bz][aot],
IL , which has a docusate anion and a different type of cation,
C
T/1C
A
CH
PCH/kPa
A
CO
PCO/kPa
2
3
3
4
5
0
5
0
0.0084
0.0191
0.0270
0.0515
0.4846
1.1205
1.6101
3.1209
0.0187
0.0429
0.0600
0.1172
0.4861
1.1338
1.6119
3.1998
containing both a phenyl and an eight carbon chain and
chmpb. The VLE plot (see Fig. 10) shows a reduced repulsive
a
A and P are defined in eqn (2)–(4).
effect compared to those obtained for IL
A B
and IL . This is most
The data reported in Table 2 clearly established the extinc- likely due to the presence of the aromatic ring, which will
tion coefficients for the two vibrational modes, n~ CO and n~ CH
as 95 300 l mol cm and 42 900 l mol cm , respectively.
,
induce p–p stacking of the cations within the liquid phase.
ꢁ
1
ꢁ1
ꢁ1
ꢁ1
The calculated VLE data obtained using COSMO-therm are a
These were used to estimating the vapour pressure of chmpb at qualitatively useful data set, and are able to predict the nature
different temperatures. The two sets of calculated pressures of the deviation from ideality, as reported for other organic
2
2
match within an average error of 0.15 kPa. This is most likely molecules dissolved in ionic liquids. However, the calculations
due to the difficulty in controlling temperature inside the cell are not quantitatively accurate in predicting the strength of the
whilst the thermostat is on the heating jacket.
interactions between the solutes and ionic liquids, probably
Vapour pressure of chmpb in multicomponent mixtures. due to the presence of long range ion–ion interactions.
The VLE of organic molecules in ionic liquids are essential data
for describing the thermodynamics of these new mixtures. In the
literature, the interactions between organic solvents and ionic
20
liquids have been mostly evaluated using static methods, which
are useful and applicable for the study of organic molecules with
relatively high vapour pressure. However, in this work, PRMs with
19
low volatility were of interest, and so a newly developed method
was used. The VLE data are expressed as deviations from Raoult’s
law, and the observations are consistent with the nature of the
interactions between organic molecules and ionic liquids, which
21
have already been reported for fluorinated ionic liquids.
The results of the study of the VLE of the systems composed
of four different ionic liquid mixtures with chmpb are reported
in Fig. 7–10. All data deviate from the ideal linear Raoult’s
1
8
plot, as represented by eqn (5).
Fig. 7 Vapour–liquid equilibria data of [N1 1 2(2O2O1)][ace] and chmpb
mixtures at 30 1C in blue, with the COSMO-therm calculations at 30 1C
represented in black and the ideal Raoult’s plot in green.
0
P
i
= P
i
w
i
(5)
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016

View Article Online
NJC
Paper
Fig. 8 Vapour–liquid equilibria of [N11 2(2O2O1)][aot] and chmpb mixtures
at 45 1C blue, COSMO-therm data in black at 45 1C and the ideal linear plot
in green.
Fig. 11 Water uptake of ionic liquids, measured by Karl-Fischer titration,
as a function of time: in blue diamond IL
triangle the mixture (0.9IL + 0.1IL ) and in purple circle the mixture (0.9IL
0.1IL ). The inset represents a simplified expansion of the absorption
A B
, in red squares IL , in green
A
C
A
+
B
between 0–300 min.
air (see Fig. 11). The water uptake of room temperature ionic
2
3
liquids has been widely investigated, and the kinetics of
2
4
absorption have been already reported for a few systems.
The amount of water absorbed is, inter alia, a function of the
cationic and anionic structures, and the time of exposure.
This, of course, will be the realistic situation of any application
of a consumer product. In all cases the ionic liquids absorb waters
from the atmosphere, initially approximately proportionate to
their exposure time (see inset on Fig. 11), but each as a different
Fig. 9 Vapour–liquid equilibria of
a
1 : 1 ionic liquid mixture of rate. After 24 h, all the ionic liquids appeared to have reached
[
N
1 1 2(2O2O1)][ace] and [N1 1 2(2O2O1)][aot] with chmpb at 45 1C in blue; equilibrium, and all showed a significant water content.
COSMO-therm data in black and the ideal linear plot in green.
The water uptake of the IL and IL
A
, IL
B
C
and two mixtures
(
0.9IL + 0.1IL and 0.9IL + 0.1IL ) has been studied titrating
A
B
A
C
the water content in time and the results are the kinetic of the
water absorption and as final value the saturation point has
been reported occurring after 24 h of atmospheric exposure.
Fortuitously, the presence of water has a moderating effect on
the vapour pressure of the mixture. This is demonstrated by the
deliberate addition of precise amount of water to solutions of
chmpb in IL
IL (see following section).
Liquid–liquid equilibria measurements. The interactions of
ionic liquids containing dissolved organic materials with water
A B C A B
, IL and IL , as well as to mixtures of IL , IL and
C
25
have been studied previously, and phase separation was observed
to occur. An insolubility gap is defined as the ‘‘area within the
coexistence curve of an isobaric phase diagram (temperature vs.
composition) or an isothermal phase diagram (pressure vs.
Fig. 10 Vapour–liquid equilibria of the mixture [N1 18Bz][aot] and chmpb at
4
2
6
5 1C in blue, COSMO-therm data in black and the ideal linear plot in green. composition)’’ and, the physicochemical observations are
consistent with this phenomenon. The strength of the
interactions involved are mainly driven by the structures of
However, for establishing trends and predicting qualitative the ionic liquids and the organic molecules.
behaviour, this method has proved excellent. The LLE data for IL , IL and IL and their mixtures reveals a
A
B
C
Kinetics of water uptake and saturation. The ionic liquids region in the ternary phase diagram, delimited by the points of
reported in Fig. 1 and their mixtures, due to their structures, the binodal line (defined by the black and red data series of cloud
are hygroscopic and so tend to absorb water if left in the open points), where the two organic species (chmpb and dmmpp) are
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
NJC
A B
Fig. 14 Ternary phase diagram of the ionic liquid mixture (0.1IL + 0.9IL )
A
Fig. 12 Ternary phase diagram of IL [N1 1 2(2O2O1)][ace] with dmmpp
with dmmpp and chmpb. The black series of data defines the binodal line
of a mixture with 0.2 wtH2O%. The red series of data defines the binodal line
of a mixture with 8 wtH2O%.
and chmpb. The black series of data defines the binodal line of a mixture
with 0.2 wtH O%. The red series of data defines the binodal line of a mixture
2
with 8 wtH O%.
2
The enhancement in the solubility gap between the wet
mixture (8 wtH O%) and the dry one (0.2 wtH O%) is mainly
due to the effect of the cation. In Fig. 13, the solubility gap for a
possible real dry mixture (represented by black dots) has an
acceptable composition range, usable in a homogeneous
solution of ionic liquids and organic components mixtures
insoluble in the ionic liquid mixtures. From the data reported
in the first ternary phase diagram (Fig. 12) for the components
2
2
IL , chmpb and dmmpb (recorded as mole fraction on the three
A
A
axes); it is clear that IL is the key component responsible for
the water uptake. For example during the first 12 h, when it
reached the 8 wtH O%, the insolubility gap defined by the red
2
(VLE data previously reported). At the same time, the repulsive
A
data series is increased compared with the dried IL mixture
2
3
effect is represented by the positive deviation from Raoult’s law
reported in Fig. 7–10.
(black data series).
In order to increase the solubility of chmpb in IL , 10 wt% of
A
A B
Olfactory tests. The ionic liquids mixtures IL and IL were
one of the other two ionic liquids (IL or IL ) has been added to
B
C
tested in accordance with the protocol described above, and the
olfactive profile, strength and character, of the PRMs in an open
system was evaluated (see Table 3). All formulations contain
the mixture, and the results are reported in Fig. 13 and 14.
Again, in these cases, the water uptake increases the area of the
insolubility region.
5
wt% of the chmpb, which imparts a character which varies
from a fruity prune note when very concentrated, through harsh
green as it becomes less concentrated and to a sweeter green
note when at lower concentration. In formulation A, the PRM is
solubilised in a traditional perfumery solvent, dipropylene glycol
(dpg), and a very strong fragrance strength is perceived from
the start that only falls by 1 unit after three hours, and is still
perceived as moderate in strength for up to 24 hours. The
character also retains a fairly constant harsh green note for the
first 5 hours. In formulation B, the same amount of the PRM is
present in the ionic liquid mixture. In this case, the initial
strength is greater by 1 unit compared to formulation A. This
initial over-powering strength is maintained for the first 3 hours
and drops by only 1 unit over five hours, clearly demonstrating
the increased level of the PRM in the air as a result of the
repulsion effect of the ionic liquid. After 24 hours, the odour
from the ionic liquid solution of chmpb is much weaker,
demonstrating the repulsion effect of the ionic liquid where
the perfume material has been exhausted.
A C
Fig. 13 Ternary phase diagram of the ionic liquid mixture (0.1IL + 0.9IL )
COSMO-RS predictions. COSMO-RS (conductor-like screening
with dmmpp and chmpb. The black series of data defines the binodal line
of a mixture with 0.2 wtH2O%. The red series of data defines the binodal line
of a mixture with 8 wtH2O%.
2
7
model for real solvents), proposed by Klampt, allows a theore-
tical description of a system composed of ionic liquids (ionic)
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016

View Article Online
NJC
Paper
Table 3 Results of olfactory test on two samples: sample A is a mixture of dipropylene glycol (dpg) and (chmpb); in sample B, the dpg was replaced with
the ionic liquid mixture 52.5 : 42.5 wt% [N1 1 2(2O2O1)][ace] (IL
A
) and [N1 1 2(2O2O1)][aot] (IL
B
)
Time after application
Formulation
10 min
1 h
3 h
3.0
5 h
24 h
(A) 5 wt% chmpb in 95 wt% DPG
4.0
4.0
3.0
2.0
Harsh green
Harsh green
Harsh green
Harsh green
Sweeter green
(
4
B) 5 wt% chmpb in 52.5 wt% IL
2.5 wt% IL
A
&
5.0
5.0
5.0
4.0
1.5
B
Very fruity, prune
Very fruity, prune
Very fruity, prune
Very fruity, prune
Very weak, fruity
and organic molecules (molecular). This methodology, once Table 4 Raoult’s deviation plot area (A
d
) calculated using a Simpson
integration method. The data were reported in the first column for the
dry mixture of each perfume raw material listed here with the ionic liquid
28
evaluated against empirical data, will allow the determination
of phase diagrams and VLE data for multicomponent mixtures
far more easily and rapidly than any experimental programme.
For this reason, three-component mixtures containing one
organic species, one ionic liquid, and water were examine both
[N1 12(2O2O1)][ace], and in the second column for the wet mixture (same
2
composition but with 10 wt% H O)
Organic materials
A
d
(dry)
d 2
A (with 10% H O)
computationally and empirically, in order to evaluate the Adoxal
Ambercore
0.86
0.50
0.97
0.80
0.64
1.14
0.81
1.06
0.66
0.66
0.75
0.97
0.87
0.72
0.70
0.72
0.76
0.75
0.78
0.78
0.74
0.71
0.77
0.95
0.68
0.77
0.71
0.83
2.26
1.53
2.87
2.06
1.59
3.59
2.20
2.92
1.63
1.69
1.82
2.65
2.32
1.75
1.79
1.69
1.98
1.85
2.15
2.09
1.89
2.05
2.42
2.68
1.55
2.42
1.97
2.24
validity and usefulness of the COSMO-RS predictions.
The empirical vapour–liquid equilibria data for IL , IL , and
Ambrocenide
a-Amyl cinnamaldehyde
A
B
C
IL with chmpb (see Fig. 8, 9 and 11, respectively) are reported, Apritone
Butylundecilenate
Cyclohexadecenone
D-Muscenone
along with the predictions from the COSMO-therm calcula-
tions. Within the errors of the experimental data (which in
some cases are considerable owing to the low vapour pressure Exaltenone
Fixolide
Frutene
Galaxolide
of chmpb), the agreement is about 20%. More significantly, the
calculated data show the same trend as the experimental data,
and reproduce both the sign of the deviation from linearity in Geranyltiglate
Glycolierral
Hexalon
Hexyl cinnamicaldehyde
Iso-E-SuperR
Linalyl cinnamate
Muscenone
the Raoult’s plots and the shape and position of the observed
maxima. In other words the predictions have value in that they
reproduce reality qualitatively, but not quantitatively.
Having established the general utility of the model and its
application to a large set of organic fragrances (see Table 4), the
Muscogene
COSMO-RS model has been used to classify the strength of the Nectaryl
Nerolidol
Norlimbanol
OkoumalR
interaction between perfume raw material and the ionic liquid.
Manipulation of a large database of molecules, required for
studying commercial fragrance products, was performed in the Pharone
Timberol
Vertofix
Z-Ambrettolide
past using large sets of experimental data connected with
the arbitrary split of molecular structures in their functional
2
9
group constituents. The experimental data available for the
system previously described are not sufficient for using a group
contribution method such as UNIFAC, and for this reason a area values for a selected list of hydrophobic perfume raw
quantum chemical calculation model COSMO-RS (Conductor materials in the ionic liquid [N1 1 2(2O2O1)][ace]. The activity
2
7
like Screening Model for Real Solvents) has been applied. The coefficients of the chmpb, calculated using COSMO-RS, suggest
proof of principle for this controlled fragrance release system has that the presence of water enhances the values, probably
3
0
been demonstrated using a mixture of a perfume raw material, a because of the chaotropic effect of water in this system.
co-solvent, and one of three ionic liquids, but is clear that a real
mixture would involve an increased number of components.
The LLE data previously reported defined the effect of the
water uptake in the solubility of the chmpb in the ionic liquid
Conclusions
mixtures. Here we demonstrate that a greater insolubility A new way to release organic material into the gas phase with
region in the ternary phase diagram (see Fig. 12–14) is reflected an enhanced vapour pressure has been described, the organic
by an increased deviation from Raoult’s law. Computationally, material exhibiting positive deviations from Raoult’s law. The
the strength of the ‘‘salting out’’ effect, due to the ionic liquid’s system involves dissolving the organic in an ionic liquid, which
water uptake, can be represented as the area of deviation from itself exhibits negligible vapour pressure. Moreover, utilising
28
Raoult’s plot calculated using the Simpson’s method (see Table 4).
the custom design, which is a fundamental precept of ionic
The data reported show the increase in the Raoult’s deviation liquid chemistry, the novel ionic liquids used in this study were
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
NJC
specifically designed to be toxicologically safe, and detailed
synthetic procedures have been reported.
8 T. Stora, S. Escher and A. Morris, Chimia, 2001, 55, 406–412.
9 Definition of ‘‘head space’’ as given by the OED, Oxford
University Press, Oxford, 2016.
In a study of the liquid–liquid equilibria of (1-cyclohexyl-2-
methylpropan-2-yl) butanoate in all four ionic liquids, it was 10 C. S. Sell, Chemistry and the Sense of Smell, John Wiley &
surprising to discover that the addition of water (which will be
Sons, Hoboken, New Jersey, 2nd edn, 2014.
omnipresent in a practical application of this new phenomenon) 11 V. G. Mata, P. B. Gomes and A. E. Rodrigues, AIChE J., 2005,
enhanced the vapour pressure of the ester. This observation
51, 2834–2852.
suggests that the enhanced volatility of fragrances is not just a 12 D. A. Gedefaw, Y. Zhou, Z. Ma, Z. Genene, S. Hellstr o¨ m,
function of their intrinsic volatility, but also depends on the
strength of the chaotropic interaction with the ionic liquid, and
indeed with the organic component itself.
Given the large number of known and predictable ionic
liquids, it is not feasible to screen all the possible options of both
F. Zhang, W. Mammo, O. Ingan ¨a s and M. R. Andersson,
Polym. Int., 2014, 63, 22–30.
13 J. R. Trindade, Z. P. Visak, M. Blesic, I. M. Marrucho, J. A. P.
Coutinho, J. N. Canongia Lopes and L. P. N. Rebelo,
J. Phys. Chem., 2007, 111, 4737–4741.
solute and solvent. For this reason, a computational approach 14 (a) R. Ahlrichs, M. B ¨a r, M. H ¨a ser, H. Horn and C. K o¨ lmel,
based on a solvation model (COSMO-RS) was investigated in order
to compute the nature and the strength of the interaction between
the solute and the solvent. This rapid and easily accessible
Chem. Phys. Lett., 1989, 162, 165–169; (b) A. Sch ¨a fer,
A. Klamt, D. Sattel, J. C. Lohrenz and F. Eckert, Phys. Chem.
Chem. Phys., 2000, 2, 2187–2193.
methodology revealed a remarkable qualitative agreement with 15 A. Sch ¨a fer, C. Huber and R. Ahlrichs, J. Chem. Phys., 1994,
the experimental data in both the direction and the maximum
100, 5829–5835.
effect for the Raoult’s law deviation, but was less successful in 16 L. A. M. Holland, O. Todini, D. M. Eike, J. M. Velazquez
predicting its magnitude. Moreover, the methodology was simply
applicable to multicomponent systems, providing invaluable
guidance to minimise the number of experimental determina-
tion required to find optimal conditions.
Mendoza, S. A. Tozer, P. M. Mcnamee, J. R. Stonehouse,
W. E. Staite, H. C. R. Fovargue, J. A. Gregory, K. R. Seddon,
H. Q. N. Gunaratne, A. V. Puga, J. Estager, F.-L. Wu,
S. D. Devine, M. Blesic and F. M. Ferrero Vallana, WO Pat.,
2016049390 A1, 2016.
1
7 M. Petkovic, K. R. Seddon, L. P. N. Rebelo and C. S. Pereira,
Chem. Soc. Rev., 2011, 40, 1383–1403.
Acknowledgements
F. M. F. V. wishes to thank the EPSRC and Procter & Gamble for 18 P. W. Atkins and J. De Paula, Elements of Physical Chemistry,
the award of a doctoral studentship.
Oxford University Press, Oxford, 2nd edn, 2013.
9 F. M. Ferrero Vallana, L. A. M. Holland and K. R. Seddon,
Aust. J. Chem., 2016, DOI: 10.1071/CH16151.
0 A. Heintz, T. V. Vasiltsova, J. Safarov, E. Bich and
S. P. Verevkin, J. Chem. Eng. Data, 2006, 51, 648–655.
1
2
Notes and references
1
M. Freemantle, An Introduction to Ionic Liquids, Royal Society
of Chemistry, Cambridge, 2010.
21 M. B. Shiflett, M. A. Harmer, C. P. Junk and A. Yokozeki,
2
3
P. Davey, Perfum. Flavor., 2008, 33, 34–35.
Fluid Phase Equilib., 2006, 242, 220–232.
F.-M. Petrat, F. Schmidt, B. Stutzel and G. Kohler, US Pat., 22 M. G. Freire, L. M. Santos, I. M. Marrucho and J. A. Coutinho,
0060166856 A1, 2003.
Fluid Phase Equilib., 2007, 255, 167–178.
(a) M. Blesic, H. Q. N. Gunaratne, P. Nockemann, 23 K. R. Seddon, A. Stark and M. J. Torres, Pure Appl. Chem.,
P. McCarron and K. R. Seddon, RSC Adv., 2013, 3, 329–333;
2000, 72, 2275–2287.
b) H. Q. N. Gunaratne, P. Nockemann and K. R. Seddon, 24 C. D. Tran, S. H. D. P. Lacerda and D. Oliveira, Appl.
2
4
(
Chem. Commun., 2015, 51, 4455–4457.
Spectrosc., 2003, 57, 152–157.
5
(a) D. M. Bourgeois, G. S. Miracle, P. J. Porter, E. Boros, 25 M. G. Freire, A. F. M. Claudio, J. M. Araujo, J. A. Coutinho,
K. R. Seddon and H. Q. N. Gunaratne, US Pat., 8093432 B2,
012; (b) Y. G. Aouad, D. M. Bourgeois, G. S. Miracle,
I. M. Marrucho, J. N. C. Lopes and L. P. N. Rebelo, Chem.
Soc. Rev., 2012, 41, 4966–4995.
2
J. Wevers, H. Q. N. Gunaratne, K. R. Seddon, E. Boros, 26 W. J. Work, K. Horie, M. Hess and R. F. T. Stepto, Pure Appl.
S. P. Katdare and T. Belhocine, US Pat., 7414158 B2, 2008.
Chem., 2004, 76, 1985.
6
7
S. A. Forsyth, H. Q. N. Gunaratne, C. Hardacre, A. McKeown, 27 A. Klamt, J. Phys. Chem., 1995, 99, 2224–2235.
D. W. Rooney and K. R. Seddon, J. Mol. Catal. A: Chem., 28 C. W. Clenshaw and A. R. Curtis, Num. Math., 1960, 2, 197–205.
2
005, 231, 61–66.
29 S. I. Sandler, Chemical and Engineering Thermodynamics,
John Wiley, New York, 3rd edn, 1998.
British Perfumery: A Fragrant History, ed. K. Glendenning and
M.-C. Dillon, Brithish Society of Perfumers, Frome, 2013.
30 H. Zhao, J. Chem. Technol. Biotechnol., 2006, 81, 877–891.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016