Bronsted acids in ionic liquids: How acidity depends on the liquid structure

Article abstract of DOI:10.1039/c4cp03217a

Gutmann Acceptor Number (AN) values have been determined for Bronsted acid-ionic liquid mixtures, over a wide compositional range. Four systems of general formula [C₂mim][A]-HA (A^- = bistriflamide, [NTf₂]^-; triflate, [OTf]^-; mesylate, [OMs]^-; or acetate, [OAc]^-, [C₂mim]⁺ = 1-ethyl-3-methylimidazolium cation) were studied. A library of Bronsted acidic systems of varying acidity was constructed and the AN parameter was found to be a convenient approach for quantifying their acidity. HOAc, HOMs and HOTf, when dissolved in ionic liquids, were found to associate with the respective anions to form hydrogen-bonded anionic clusters, [A(HA)_x]^-. In contrast, HNTf₂ was solubilised as a discrete, undissociated molecule. AN values were sensitive to the presence of anionic clusters; acidity could be buffered to a particular AN by binding the solubilised acid in the anionic cluster form. Overall, a simple way to manipulate and quantify the Bronsted acidity of acid-ionic liquid mixtures was demonstrated, and measured AN values were related to liquid speciation. This journal is

Full text of DOI:10.1039/c4cp03217a

PCCP
View Article Online
View Journal | View Issue
PAPER
Brønsted acids in ionic liquids: how acidity
depends on the liquid structure
Cite this: Phys.Chem.Chem.Phys.,
2
014, 16, 23233
a
a
a
a
Jade A. McCune, Peizhao He, Marina Petkovic, Fergal Coleman,
ab
a
a
Julien Estager, John D. Holbrey, Kenneth R. Seddon and
a
Małgorzata Swadzba-Kwasny*
´
´
Gutmann Acceptor Number (AN) values have been determined for Brønsted acid–ionic liquid mixtures,
ꢀ
over a wide compositional range. Four systems of general formula [C
2
mim][A]–HA (A = bistriflamide,
ꢀ ꢀ ꢀ ꢀ
+
[
NTf ] ; triflate, [OTf] ; mesylate, [OMs] ; or acetate, [OAc] , [C mim] = 1-ethyl-3-methylimidazolium
2
2
cation) were studied. A library of Brønsted acidic systems of varying acidity was constructed and the AN
parameter was found to be a convenient approach for quantifying their acidity. HOAc, HOMs and HOTf,
when dissolved in ionic liquids, were found to associate with the respective anions to form hydrogen-
ꢀ
Received 20th July 2014,
bonded anionic clusters, [A(HA)
x
] . In contrast, HNTf
2
was solubilised as a discrete, undissociated molecule.
Accepted 11th September 2014
AN values were sensitive to the presence of anionic clusters; acidity could be buffered to a particular AN by
binding the solubilised acid in the anionic cluster form. Overall, a simple way to manipulate and quantify
the Brønsted acidity of acid–ionic liquid mixtures was demonstrated, and measured AN values were related
to liquid speciation.
DOI: 10.1039/c4cp03217a
www.rsc.org/pccp
Table 1 A list of ion abbreviations used in this contribution
Introduction
Solutions of acids in ionic liquids
Cation
+
1
1
-Ethyl-3-methylimidazolium
-Methylpyrrolidinium
[C
[Hmpyr]
2
mim]
Solutions of Brønsted acids in ionic liquids are of interest for a
number of applications: as ‘liquefied’ pharmaceutically active
+
1
2
drugs, anti-fungal agents, components in conductive compo- Anion
3
site materials, and – most commonly – Brønsted acidic catalysts
for a very wide range of organic reactions. At a fundamental Ethanoate (acetate)
ꢀ
Halide
X
(X = Cl or Br)
4
,5
ꢀ
[OAc]
[tfa]
[OMs]
[OTf]
ꢀ
Trifluoroethanoate (trifluoroacetate)
Methanesulfonate (mesylate)
Trifluoromethanesulfonate (triflate)
level, these solutions are interesting as complex fluids with a
dense network of hydrogen bonds.
ꢀ
ꢀ
ꢀ
Structural studies on systems sharing the same anion/ Bis{(trifluoromethyl)sulfonyl}amide (bistriflamide)
conjugate base have shown association of complex, H-bonded
[NTf2]
ꢀ
6
anionic clusters, [A(HA) ] . These are small oligomeric units,
x
1
0
comprising typically 2–4 anions and acids, associated by strong charged species. Similarly, ionic liquids formed by proton
hydrogen bonds, and persistent enough to be detected by transfer between an acid and a base, with the acid in excess,
various spectroscopic techniques. Organic salts and ionic have been shown to form complex anions, for example
ꢀ
11
ꢀ 6
ꢀ 5
liquids with halide anions have been shown to form dimeric [F(HF) ] , [tfa(Htfa)] , [HSO (H SO ) ] , and a wide range
x
4
2
4 x
1,12,13
0
ꢀ
[
Xꢁ ꢁ ꢁHꢁ ꢁ ꢁX ] anions when brought into contact with an appro- of carboxylate–carboxylic acid clusters.
Poorly ionic mix-
ꢀ
priate HX, such as [Clꢁ ꢁ ꢁHꢁ ꢁ ꢁCl] , which is stable even under tures of 1-methylpyrrolidine and acetic acid were shown to
7
ꢀ
13
reduced pressure, as well as other (less persistent under contain polymeric [OAc(HOAc)x] anions. Such clusters are
ꢀ
8
ꢀ
ꢀ
vacuum) oligomeric species: [Cl(HCl)
2
] , [HBr
2
] , [Br(HBr)
not specific to an ionic liquid environment; they have been
mim][OTf] (see observed in aqueous solutions, non-polar solvents, and are
Table 1 for the list of abbreviations used in this work) were often found in crystal structures.
2
]
ꢀ
9
14
15
and [ClHBr] . Solutions of triflic acid in [C
2
1
6
reported to contain a dense network of H-bonds between all
The acidity of bistriflamidic acid
a
b
Hydrogen bis{(trifluoromethyl)sulfonyl}amide, HNTf
an acidic N–H functionality, in contrast to the more common
O–H function found in many acids. Pure HNTf is a low-melting
2
, contains
QUILL, The Queen’s University of Belfast, Belfast, BT9 5AG, UK.
E-mail: m.swadzba.kwasny@qub.ac.uk
CERTECH asbl, Rue Jules Bordet, Zone Industrielle C, B-7180 Seneffe, Belgium
2
This journal is ©the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 23233--23243 | 23233

View Article Online
Paper
PCCP
Fig. 2 Interactions of the spectroscopic probe, triethylphosphine oxide
(
tepo), with a Lewis acid (LA), left, and with a Brønsted acid (HB), right.
2 2
Fig. 1 Crystal structure of HNTf , showing strands of H-bonded HNTf
units (H-bonds – turquoise; C – grey, S – yellow, O – red, F – lime, N – purple
18
groups. Finally, the AN scale is a quantitative measure of acidity
applicable to both Lewis and Brønsted acids.
and H – white).
3
1
AN values are determined by P NMR spectroscopy, making
1
7
31
solid (49–50 1C), which crystallises in a characteristic ribbon- use of small quantities of the weakly basic P NMR probe
1
8
like structure (Fig. 1). In the crystal structure of HNTf
NTf unit is in the trans-position, which is calculated to be the medium under investigation. Changes in the P NMR
favourable by 2.3 kJ mol can chemical shift are induced through formation of acid–base
2
, the molecule, triethylphosphine oxide (tepo), which is dissolved in
3
1
2
ꢀ
1 19
ꢀ
2
.
However, in salts, [NTf
]
crystallise in either a cis- or trans-conformation (depending on adducts with Lewis acids, or partial protonation by Brønsted
the cation), and in the liquid phase (solutions or ionic liquids) acids (Fig. 2).
it tends to assume both cis- and trans-conformations in a
AN measurements have been used to study Lewis acidity of
2
0
29
dynamic equilibrium.
HNTf
gas phase, HNTf
In glacial acetic acid solution, HOTf was measured to be was demonstrated to be among key factors affecting catalytic
chlorometallate ionic liquids, and to investigate a wide range
3
0
2
is generally recognised to be a very strong acid. In the of ‘typical’ ionic liquids.
2
1
31
5
2
was found to be a stronger acid than HOTf.
Acidity, expressed on both Hammett and Guttmann scales,
1
7
2 a
stronger than HNTf (pK = 4.2 vs. 7.8), but both acids were performance of Brønsted acidic ionic liquids.
classified as very strong. In contrast, when equimolar amounts
of water and HNTf2 were dissolved in a non-coordinating
Aim of this work
solvent, such as CCl4 or benzene, HNTf did not protonate
2
The principal objective was to study the acidity of acid–ionic
liquid mixtures as a function of their composition, investigating
examples of both weak and strong acids. It was aimed to
understand the relationship between the acidity of such systems,
i.e. AN values, and their speciation, i.e. the presence of anionic
H-bonded clusters. Quantification of acidity for different acids,
at a range of compositions, will lead to establishment of a library
of systems with a fine-tuned acidity, in analogy to the library of
1
5
water, behaving like a weak acid. This was in stark contrast to
other acids, such as HOTf or H SO . Furthermore, HNTf
dissolved in anhydrous non-polar solvents did not form
H-bonded clusters, [HNTf , but was solubilised in the mole-
2
4
2
2 x
]
1
5
cular form (again, in contrast to HOTf). Stoyanov et al. argued
ꢀ
that [NTf
2
]
is such a very weak base/hydrogen bond acceptor,
that it is energetically more favourable for HNTf to be solvated
2
ꢀ
by solvent molecules, than to form the cluster, [NTf (HNTf )] .
2
2
32
Lewis acidic ionic liquids, previously established by our group.
This approach should provide more insight into the selection of
acidic ionic liquid systems as homogenous catalysts.
Acidity in protic ionic liquids
Many techniques have been used to measure the acidity of ionic
liquids, and equally, many different scales have been used to
express the results. The approach used routinely for aqueous
2
2
Experimental
Materials
solutions poses technical difficulties in non-aqueous media.
Establishing pH or pK in ionic liquid media is possible, albeit
a
2
3
challenging. Hammett acidities reported vary widely depend-
Unless otherwise stated, all materials were purchased from
Sigma-Aldrich and used as received. Triethylphosphine oxide
was purchased from Sigma-Aldrich, opened and stored in a
2
4,25
ing on the measurement procedure used,
and the measure-
ments are extremely sensitive to impurities, as they rely on
26
UV-vis probes. A range of methodologies, associated difficulties,
and best recommended practice have been expertly discussed by
nitrogen-filled glovebox (MBraun LabMaster dp; o0.1 ppm O
2
2
and H O). Ethyl trifluoromethanesulfonate, 99%, was purchased
24,25
Johnson and co-workers.
A noteworthy addition to this strand
2
in glass ampoules packed under argon. [C mim]Br was prepared
by alkylation of 1-methylimidazole with bromoethane according
to a standard literature procedure.
of research is the use of an NMR probe to measure Hammett
acidity, very recently introduced in ionic liquids by Welton and
27
co-workers.
2
8
In this work, a Guttmann Acceptor Number (AN) approach
has been employed, which benefits from straightforward NMR-
Syntheses
based methodology, and in the literature good correlations are In all cases, these are improved modifications of known
3
3–36
found for measurements performed independently by different synthetic procedures.
2
3234 | Phys. Chem. Chem. Phys., 2014, 16, 23233--23243
This journal is © the Owner Societies 2014

View Article Online
PCCP
Paper
3
3
HNTf
2
.
2
Li[NTf ] (1.00 mol eq.) and sulfuric acid, (3.00 mol eq.) Finally, the synthesised ionic liquid was dried under high
were placed in a cylindrical Kugelrohr flask and fixed to a Kugelrohr vacuum (70 1C, overnight). The purity and structure of the
1
13
distillation apparatus. High vacuum was applied and the tem- product were confirmed using H and C NMR spectroscopy.
3
6
perature of the oven was elevated slowly until the first signs of
[C mim][OAc]. [C mim][HSO ] (1.0 mol eq.) was poured
2 2 4
sublimation were observed (B50 1C). After the sublimation, white into a round-bottomed flask equipped with a stirring bar and
crystals were gathered in the collector flask, which was cooled dissolved in hot deionised water (ca. 95 1C). Strontium hydroxide
with ice throughout the experiment. The hygroscopic, fuming octahydrate (1.2 mol eq.) was weighed into a conical flask and
sublimate was melted using a heat-gun and very quickly trans- dissolved in hot deionised water (ca. 95 1C). The [C mim][HSO ]
2 4
ferred into a bottle with a sealable, corrosion-resistant cap. The solution was quantitatively transferred to the same conical flask,
product was obtained as a white, low-melting solid, typically in and the cloudy mixture was stirred vigorously, whilst allowing to
7
5–80% yield.
cool to ambient temperature. The white precipitate (strontium
typical synthesis, concentrated hydrogensulfate) was removed via double filtration under
3
4
[
C mim][NTf ].
In
a
2
2
aqueous solutions of [C mim]Br (1.0 mol eq.) and LiNTf
(
reduced pressure (a P3 glass sintered funnel with a 0.5 cm silica
1.2 mol eq.) were poured into a round-bottomed flask equipped gel layer), washed with deionised water (ca. 20 1C) and discarded.
with a stirring bar. The flask was closed with a stopper and the The colourless solution of [C mim][OH] was neutralised to the
mixture stirred vigorously (30 min). Subsequently, the top (aqueous) expected pH of [C mim][OAc] aqueous solution, pH = 8.4, by the
layer was decanted, and the bottom (ionic liquid) layer was washed addition of acetic acid (1.0 mol eq.). Water was removed using a
three times with an aqueous solution of LiNTf and ten times with rotary evaporator (80 1C, 4 h). To the crude ionic liquid,
2
2
2
2
2
deionised water. The remaining solvents were removed using a an equal volume of the dichloromethane/acetonitrile mixture
rotary evaporator (80 1C, 1 h) and then high vacuum (80 1C, over- (1 : 2 vol ratio) was added, resulting in a small amount of fine
ꢀ
2
night, 10 mbar). The purity and structure of the product were white precipitate, which was again removed by filtration under
1
13
19
confirmed using H, C, and F NMR spectroscopy. Efficiency reduced pressure. Finally, solvent was removed using the rotary
of anion exchange was confirmed by elemental analysis (Li o evaporator
0
.005 wt%; Br = 0.13 wt%).
(1 h, 40 1C), and then the ionic liquid was dried under high
3
4
2
[C mim][OTf]. 1-Methylimidazole (1.0 mol eq.) was poured vacuum (50 1C, 2 days). The purity and structure of the product
1
13
into a two-necked round-bottomed flask equipped with a were confirmed using H and C NMR spectroscopy.
stirring bar, an equal volume of dichloromethane was added [C mim][A]–HA. In a typical experiment, ionic liquid (ca. 3 g)
2
to the same flask, and the solution was stirred vigorously in an was weighed out into a sample vial equipped with a stirring bar.
ice bath (0 1C) until cooled. An ampoule of ethyl trifluoro- An appropriate amount of acid was weighed in the fumehood,
methanesulfonate (1.1 mol eq.) was opened in a glovebox and directly into the sample vial containing the ionic liquid, using a
2
the content was transferred immediately into a pressure- Pasteur pipette (or a spatula for HNTf ), to achieve the desired
equalising dropping funnel, which was closed with a stopper, ratio of acid to ionic liquid. Actual masses of both reactants
taken out of the glovebox, and mounted on the flask containing were recorded. The vial was closed with a plastic cap and stirred
the cooled solution. The whole setup was connected to a vigorously (ambient temperature, ca. 30 min) to ensure homo-
Schlenk line and purged with argon. Ethyl trifluoromethane- geneity. Argon was bubbled through this solution (ambient
sulfonate was added dropwise into the vigorously stirred reac- temperature, ca. 30 min), which was subsequently placed in the
tion mixture (ca. 20 min). Afterwards, the reaction was carried glovebox.
out overnight, gradually reaching room temperature. The pro-
duct (room temperature ionic liquid) was washed four times
Analytical methods
with ethyl acetate, and the remaining solvent was removed
AN measurements. Sample preparation was carried out in a
using a rotary evaporator (1 h, 40 1C). Finally, the synthesised dry, nitrogen-filled glovebox. In a typical experiment, a sample
ionic liquid was dried under high vacuum (70 1C, overnight). (ca. 1 g) of each composition was weighed into a sample vial
3
The purity and structure of the product were confirmed using (10 cm ) equipped with a PTFE-coated magnetic stirring bar.
1
13
19
H, C, and F NMR spectroscopy.
The vial was left on the balance, and a small, accurately
3
5
[
C mim][OMs].
2
1-Methylimidazole (1.0 mol eq.) was measured, amount of triethylphosphine oxide (tepo) was added
+
poured into a two-necked round-bottomed flask equipped with (1–5 mol% per mol of the [C
2
mim] cation). The sample vial was
a stirring bar and stirred vigorously in an oil bath (40 1C). Ethyl then closed with a plastic cap and the mixture was stirred
methanesulfonate (1.2 mol eq.) was transferred into a pressure- overnight to ensure complete dissolution of tepo. Subsequently,
equalising dropping funnel, which was closed with a stopper the liquid was loaded into an NMR tube (5 mm, borosilicate
6
and mounted on the round-bottomed flask; the whole setup glass), containing a sealed capillary with d -dimethylsulfoxide
was connected to a Schlenk line and purged with argon. Ethyl (as an external lock). The tube was closed with a standard cap,
methanesulfonate was added dropwise into the vigorously sealed with parafilm, and removed from the glovebox immedi-
stirred reaction mixture (ca. 20 min). After the reaction (over- ately prior to measurement. For each ionic liquid–acid mixture
night, 40 1C), the product (room temperature ionic liquid) was studied, three samples with different small quantities of tepo
3
1
washed four times with ethyl acetate, and the remaining were prepared, to allow extrapolation of the P NMR chemical
solvent was removed using a rotary evaporator (1 h, 40 1C). shift value to infinite dilution, dinf
.
This journal is ©the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 23233--23243 | 23235

View Article Online
Paper
PCCP
31
P NMR spectra were acquired at 121.452 MHz using a Bruker
In this work, the starting point is an ionic liquid,
DRX 300 spectrometer. Phosphoric acid, 85% solution in water, [C mim][A], which is reacted with an acid, HA. To facilitate
2
was used as an external reference. All samples were measured at comparison with previous contributions, we perform a Gedanken
2
7 1C. Moreover, three solutions of tepo in hexane (ca. 5, 10 and experiment assuming that the ionic liquid was prepared by
3
7
15 mol%) were prepared, and then measured at 27 1C. For each ionic addition of an acid to a carbene (pK
aH O
= ca. 27), as shown
2
31
liquid, the P NMR chemical shift for the infinite dilution of tepo, in eqn (2).
31
dinf, was determined by extrapolation from the P NMR chemical
shifts measured at different tepo concentrations. The chemical shift
of tepo in hexane, extrapolated to infinite dilution, dinf hex, was used
as a reference (dinf hex = 0 ppm). The AN values for all samples were
(
2a)
28
calculated from the following formula: AN = 2.348ꢁd .
inf
1
13
H and C NMR spectroscopy. In the glovebox, neat ionic
liquids were loaded into NMR tubes (5 mm, borosilicate glass)
containing sealed capillaries with d -dimethylsulfoxide (as an
6
(2b)
external lock). The tubes were closed with a standard cap,
sealed with parafilm, and removed from the glovebox immedi-
ately prior to measurement. All spectra were measured at 27 1C,
using a Bruker DRX 300 spectrometer.
Following this logic, the composition of neat ionic liquid,
[
C mim][A], is described by w = 0.50, that of solutions of acids
2 HA
is characterised by 0.50 o w o 1.00, and that of neat acid as
HA
w
HA = 1.00.
Results and discussion
Experimental plan
The acidity of acid–ionic liquid mixtures
Measurements and observed trends. The spectroscopic
probe, tepo, was used in very low concentrations to minimise
the interference with equilibria in the studied systems. Recorded
Choice of acids. Four ionic liquid–acid systems of general
formula [C mim][A]–HA (A = NTf , OTf, OMs, or OAc) were prepared,
2 2
with varying molar ratios of acids, wHA. All samples were mobile
liquids, from colourless to pale yellow. For the [C mim][NTf ]–
31
P NMR signals of the protonated probe were found to be slightly
concentration-dependent, as exemplified in Fig. 3. To eliminate
2
2
31
this factor, the P NMR shifts were extrapolated to infinite
dilution of tepo, dinf, using linear regression. The dinf values were
then used to calculate AN. The AN values, along with errors derived
from the linear extrapolation, are listed in Table 2. Noteworthily,
error appears to be largest for very low concentration of acids,
where even small amounts of the probe are significant compared
to the amount of introduced acid. In some cases, AN values
measured for samples of the same system, with the same acid
concentration, albeit prepared at different times from different
batches of ionic liquids, differ by more than the assigned
HNTf system, homogenous solutions were formed for w
r
2
HNTf2
0.80; for higher acid concentrations, the mixture formed a biphasic
liquid at elevated temperatures (fine droplets suspended in bulk
liquid), and the whole sample solidified over several minutes upon
cooling, starting from the suspended phase.
The acids were chosen to represent a wide acidity spectrum.
Acetic acid is a weak carboxylic acid, known to form strong
anionic clusters. HOMs and HOTf are very strong sulfonic acids,
commonly used in catalysis. The only structural difference
between them is perfluorination of the methyl group in the latter
case, which increases the acidity of HOTf with respect to HOMs.
experimental error (viz. [C
2 2 2
mim][NTf –HNTf ], wHNTf2 = 0.60,
where DAN = 1.6). This likely derives from minute variations in
The interesting character of the HNTf
2
system has been discussed
ꢀ
above; furthermore, [NTf
2
] -based ionic liquids are hydrophobic,
and hence may be of particular interest for catalytic applications.
Each system shared the same anion and conjugate base of
the acid; for two different bases in the system, the proton would
associate with the stronger base, as already reported in the
2
6
literature. The same publication demonstrated that cations
have negligible influence on the acidity.
Definition of molar ratio of acid, vHA. Previous studies on
H-bonded clusters in ionic liquids have focussed on proton
transfer systems, where their composition was expressed as a
molar ratio of acid, wHA (eqn (1)).
1
/2HA + B $ 1/2[HB ][A] wHA = 0.33
(1a)
(1b)
(1c)
(1d)
2
HA + B $ [HB][A]
HA + B $ [HB][A(HA)]
HA + B $ [HB][A(HA)2] wHA = 0.75
wHA = 0.50
31
Fig. 3 Experimental P NMR chemical shifts for tepo as a function of tepo
2
w
HA = 0.67
concentration in selected [C mim][OAc]–HOAc compositions, expressed as
2
wHOAc. Experimental data were extrapolated to infinite dilution, dinf/ppm, using
3
linear regression.
2
3236 | Phys. Chem. Chem. Phys., 2014, 16, 23233--23243
This journal is © the Owner Societies 2014

View Article Online
PCCP
Paper
Table 2 AN values for different compositions for four ionic liquid–acid systems of a general formula [C
2
mim][A]–HA, where wHA is the molar ratio of acid
used theoretically (see eqn (3))
A = NTf
2
A = OTf
A = OMs
A = OAc
w
HA
AN
28.3 ꢂ 1.4
w
HA
AN
33.4 ꢂ 0.4
w
HA
AN
w
HA
AN
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
.50
.51
.51
.52
.53
.56
.59
.60
.60
.62
.62
.64
.64
.66
.67
.71
.71
.75
.77
.80
.83
.86
.00
0.50
0.51
0.51
0.52
0.53
0.54
0.56
0.56
0.60
0.60
0.64
0.64
0.67
0.67
0.71
0.75
0.80
0.86
0.92
1.00
0.50
0.51
0.51
0.52
0.56
0.57
0.59
0.60
0.60
0.60
0.61
0.64
0.64
0.65
0.67
0.71
0.71
0.75
0.78
0.80
0.86
0.92
1.00
25.4 ꢂ 0.4
83.4 ꢂ 3.1
86.6 ꢂ 3.1
89.5 ꢂ 0.9
95.3 ꢂ 0.3
95.8 ꢂ 0.3
96.1 ꢂ 0.2
96.0 ꢂ 0.2
94.9 ꢂ 0.2
96.3 ꢂ 0.1
95.9 ꢂ 0.1
95.3 ꢂ 0.1
95.9 ꢂ 0.1
96.1 ꢂ 0.1
98.3 ꢂ 0.1
105.5 ꢂ 0.1
105.8 ꢂ 0.1
111.0 ꢂ 0.1
115.4 ꢂ 0.1
116.5 ꢂ 0.1
121.2 ꢂ 0.1
124.5 ꢂ 0.1
0.50
0.53
0.56
0.59
0.60
0.61
0.63
0.64
0.66
0.69
0.71
0.75
0.80
0.91
1.00
16.9 ꢂ 0.1
18.9 ꢂ 0.1
19.5 ꢂ 0.1
19.9 ꢂ 0.3
21.5 ꢂ 0.1
21.6 ꢂ 0.1
22.9 ꢂ 1.4
21.0 ꢂ 0.1
22.6 ꢂ 0.1
26.2 ꢂ 0.2
29.6 ꢂ 0.3
35.3 ꢂ 0.1
39.1 ꢂ 0.3
47.3 ꢂ 0.4
124.4 ꢂ 1.7
125.5 ꢂ 1.7
123.5 ꢂ 0.2
124.9 ꢂ 0.1
123.3 ꢂ 0.1
124.1 ꢂ 0.1
124.9 ꢂ 0.1
123.3 ꢂ 0.1
124.3 ꢂ 0.1
124.5 ꢂ 0.1
123.5 ꢂ 0.1
124.3 ꢂ 0.1
124.8 ꢂ 0.1
123.3 ꢂ 0.1
123.5 ꢂ 0.1
125.1 ꢂ 0.1
123.9 ꢂ 0.1
124.0 ꢂ 0.1
124.4 ꢂ 0.1
—
118.6 ꢂ 0.3
118.7 ꢂ 0.1
118.0 ꢂ 0.1
119.4 ꢂ 0.1
117.6 ꢂ 0.1
117.6 ꢂ 0.1
117.6 ꢂ 0.1
117.2 ꢂ 0.1
117.2 ꢂ 0.1
117.3 ꢂ 0.1
117.3 ꢂ 0.1
118.9 ꢂ 0.5
118.8 ꢂ 0.5
121.9 ꢂ 0.1
123.2 ꢂ 0.1
125.3 ꢂ 0.1
127.7 ꢂ 0.1
129.3 ꢂ 0.1
129.1 ꢂ 0.1
28
52.9
—
—
28
126.1
acid concentration. In no case is the increase in AN value
directly proportional to wHA
.
In the acetate system (Fig. 4, ), upon the addition of acid to the
neat ionic liquid, there is a gradual, linear increase in AN up to
w
HOAc = 0.67; the acidity is buffered to a nearly constant AN value by
ꢀ
the formation of the [OAc(HOAc)] complex, which is held together
by strong hydrogen bonds. Moving towards higher acid content, AN
values increase more rapidly up to wHOAc = 0.75, as dimers are
ꢀ
gradually replaced by [(OAc)(HOAc)
2
] . Finally, between wHOAc = 0.75
and 1.00, AN increases linearly; it can be envisaged that the
structure of the liquid gradually shifts from that of an ionic liquid
to that of a pure acid. Fitted linear regressions are given in eqn (3).
2
AN = 34.0ꢁwHOAc + 0.5; R = 0.856; 0.50 o wHOAc o 0.67
(3a)
2
AN = 135.8ꢁwHOAc ꢀ 67.1; R = 0.995; 0.67 r wHOAc r 0.75
3b)
(
2
AN = 71.3ꢁwHOAc ꢀ 18.1; R = 0.997; 0.75 r wHOAc o 1.00
3c)
(
Fig. 4 AN values in four systems of a general formula [C
2
mim][A]–HA,
where A = OAc ( ), OMs ( ), OTf (’) and NTf ). The composition is given
2
(
This interpretation is in line with previous structural studies on
similar systems, in particular with work on 1-methylpyrrolidine–
acetic acid mixtures by MacFarlane and co-workers.
in mol ratio of acid to 1-ethyl-3-methylimidazole carbene, wHA. The lines are
fitted by linear regression. Dashed line at wHA = 0.67 is a guidline for the eye.
13
In the mesylate system (Fig. 4, ), addition of a small
amount of acid results in a rapid increase in acidity, which
adventitious water content, which occurred despite adopting a
uniform procedure of sample handling and preparation.
To facilitate meaningful analysis of the data, AN values for
the four investigated systems were plotted against the composi-
tion, wHA (Fig. 4).
then remains constant (AN = 95.8 ꢂ 0.5) for 0.52 4 wHOMs
4
0
.65. Between wHOMs = 0.65 and 0.78, AN increases linearly,
following eqn (4), and subsequently the behaviour deviates
from linear, approaching the acidity of pure acid.
All the neat ionic liquids have low acidity; measured values
are in close agreement with those reported in the literature.
As anticipated, in all cases AN values increase with increasing
2
3
0
AN = 144.7ꢁwHOMs + 2.7; R = 0.993; 0.65 r wHOMs r 0.78
(4)
This journal is ©the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 23233--23243 | 23237

View Article Online
Paper
In the triflate system (Fig. 4, ’), even the smallest quantity
PCCP
Table 3 Distances between hydrogen and oxygen atoms partaking in H-
ꢀ
of acid caused an immense increase in acidity, which was bonding in [Aꢁ ꢁ ꢁH–A] dimeric clusters, based on averaged data from
4
0
buffered at AN = 118 ꢂ 0.8 for 0.50 4 wHOTf Z 0.67. This can CSD
ꢀ
be rationalised by the formation of [OTf(HOTf)] , which had
a
a
a
Cluster
O–H /Å
Hꢁ ꢁ ꢁO /Å
Oꢁ ꢁ ꢁO/Å
+O–Hꢁ ꢁ ꢁO /1
1
0
15
been found in both ionic liquids and in molecular solvents.
Afterwards, AN values increased linearly up to wHOTf = 0.86
eqn (5)), suggesting formation of weakly-bound oligonuclear
ꢀ
[
[
[
AcOꢁ ꢁ ꢁH-OAc]
1.23 ꢂ 0.01 1.23 ꢂ 0.01 2.46 ꢂ 0.01 178 ꢂ 3
ꢀ
MsOꢁ ꢁ ꢁH-OMs] 1.03 ꢂ 0.10 1.45 ꢂ 0.10 2.50 ꢂ 0.03 172 ꢂ 2
ꢀ
(
TfOꢁ ꢁ ꢁH-OTf]
0.78 ꢂ 0.12 1.73 ꢂ 0.07 2.49 ꢂ 0.05 171 ꢂ 7
1
0
clusters, and subsequently, deviating from linearity, they
reached the acidity of pure acid.
a
Data averaged only for detected protons.
2
AN = 49.1ꢁwHOTf + 86.3; R = 0.974; 0.67 r wHOTf r 0.86
itself. This explains why, typically, this system is not catalyti-
(5)
3
9
cally active.
From a practical viewpoint, when handling solutions of
HOTf and HNTf in ionic liquids, there is a safety benefit as
How can these trends be justified in terms of dynamic
equilibria? In the region of 0.50 o wHA o 0.67, buffering to a
constant AN value was observed. This buffering occurs accord-
ing to eqn (6); for the case where x = 0 to 1, and the equilibrium
is strongly shifted to the right.
2
the solutions do not fume like the corresponding neat acids,
and are consequentially easier to handle.
Acidity vs. cluster structure. To aid understanding of
ꢀ
ꢀ
H-bonded clusters, [Aꢁ ꢁ ꢁH–A] , in the liquid phase, crystal
[A(HA)
x
]
+ HA $ [A(HA)x+1]; x = 0 to 6
(6)
structures of salts featuring such clusters were analysed. The
4
0
Interestingly, although the acidities of neat HOTf and HOMs entire Cambridge Crystal Structure Database was searched
are quite similar (DAN = ANHOTf ꢀ AN = 3), the mesylate using the search term ‘‘Oꢁ ꢁ ꢁHꢁ ꢁ ꢁO’’, where ‘‘ꢁ ꢁ ꢁ’’ was defined as
HOMs
ꢀ
ꢀ
system is buffered to a much lower acidity (DAN 4 20), as ‘any bond’. All structures with [AcOꢁ ꢁ ꢁH-OAc] , [MsOꢁ ꢁ ꢁH-OMs]
ꢀ
ꢀ
shown in Fig. 4. Since [OMs] is a better H-bond acceptor than or [TfOꢁ ꢁ ꢁH-OTf] units were identified; no analogues featuring
ꢀ
ꢀ
ꢀ
[
OTf] , protons are more strongly bound within [OMs(HOMs)]
[NTf
the hydrogen-bonding protons had been found explicitly in the
2
] were found. From these structures, the subset in which
clusters, which in turn leads to lower AN values.
For 0.67 o wHA o 1.00, the structure of the liquid changes crystallographic data were selected. Within this selection, dis-
ꢀ
from acid dissolved in ionic liquid, to ionic liquid dissolved in tances between hydrogen and oxygen atoms in the [Aꢁ ꢁ ꢁH–A]
acid. Acetic acid is a very structured medium, with a strong clusters were averaged for acetate (8 structures), mesylate
ꢀ
network of H-bonds. It is visible that acetates form [A(HA)]
(3 structures, proton detected for 2) and triflate clusters
ꢀ
(up to wHA = 0.67) then [A(HA)
2
]
(up to wHA = 0.75), and then (5 structures, proton detected for 3), as shown in Table 3.
acid becomes the continuous medium (eqn (6) is shifted to the
In examining these clusters, the more symmetrical the
1
3,38
left for wHA 4 0.75).
HOMs and HOTf are less structured, as O–H–O unit is, the stronger are the internal H-bonds, and the
their conjugate bases are weaker H-bond acceptors. Therefore, more stable is the cluster. Thus, binuclear acetate clusters,
the ‘ionic’ structure prevails longer: up to wHOMs D 0.78 in the where the two O–H bonds have identical length (1.23 ꢂ 0.01 Å;
mesylate system, and up to wHOTf D 0.85 in the triflate system. very strong H-bond), are very stable. Low acidity of the proton
ꢀ
Past these wHA values, eqn (6) is shifted to the left.
in [OAc(HOAc)] derives directly from these strong H-bonds. In
ꢀ
In the bistriflamide system (Fig. 4, ), as in the triflate one, mesylate complexes, [OMs(HOMs)] is asymmetric (rO–H = 1.03
the addition of a small amount of acid leads to a spike in and rOꢁ ꢁ ꢁH = 1.45 Å). HOMs is a much stronger acid, and
acidity, up to a very high level of AN = 124.3 ꢂ 0.8. However, in mesylate is a poorer H-bond acceptor, and so the acid retains
contrast to other systems, this value remains constant through- more of its identity as a molecular species. This trend is
out the measurable composition range 0.05 o wHNTf2 r 0.80. continued with the [OTf(HOTf)] cluster. Triflate is a weak base
This suggests that there are no clusters formed in this range, and an extremely poor H-bond acceptor; therefore the O–H
ꢀ
ꢀ
and that HNTf is dissolved in its molecular form, as found in bond in [OTf(HOTf)] is very short, shorter than in water
2
the work on acid solutions in non-polar solvents by Stoyanov (0.957 Å in the gas phase), and Oꢁ ꢁ ꢁH is of medium strength.
1
5
et al. For 0.80 o w
o 1.00, the mixtures formed biphasic Schematic structures of the three dimers are shown in Fig. 5.
HNTf2
liquids at elevated temperature and solidified readily (forming
The AN studies indicated that there is no association of
solid with small amount of encapsulated liquid) when cooled to HNTf
2
with anions in the [C
2
mim][NTf
2
]–HNTf
2
system. This is
ꢀ
ambient temperature.
supported by the absence of [NTf
2
(HNTf
2
)] clusters in the CSD
Implications. Solutions of HOTf or HNTf
liquids are both extremely acidic, with the HNTf
2
in analogous ionic database, and by studies on HNTf
2
solutions in non-polar
] is the poorest H-bond acceptor studied here;
1
5
ꢀ
2
2
system being solvents. [NTf
slightly more so. This is in agreement with the suggestion of in addition, there is a steric hindrance, as well as fluorine–
Johnson and co-workers, that gas-phase acidities of neat acids fluorine and oxygen–oxygen repulsion, hindering the formation
might be more indicative of acidities of ionic liquid solutions of N–Hꢁ ꢁ ꢁN bonds, as shown in Fig. 5.
30
2
4
than the pK
a
values of the corresponding acids in water.
This analysis ties in the empirical basicity of the anions
Solutions of acetic acid in the corresponding ionic liquid with their structural consequences and resulting acidity of their
had extremely low acidity, much lower than even the acetic acid solutions.
2
3238 | Phys. Chem. Chem. Phys., 2014, 16, 23233--23243
This journal is © the Owner Societies 2014

View Article Online
PCCP
Paper
system were the most sensitive to compositional changes (Fig. 6),
and those in the [C
affected (Fig. 9). The most deshielded protons were found in
C mim][OTf]–HOTf system (Fig. 8). Spectra for [C mim][OMs]–
2 2 2
mim][NTf ]–HNTf system were the least
[
2
2
HOMs are shown in Fig. 7.
Interactions of imidazolium cations with anions in ionic
3
4,41
liquids are well discussed elsewhere.
The chemical shifts of
protons in the imidazolium ring depend on the anion (down-
field shift due to H-bonding) and on the concentration (upfield
shift due to ring stacking). The proton in the C2 position is the
most acidic, hence it is the most sensitive to changes in H-bond
1
accepting ability of the anion. The H NMR chemical shifts for
the C2 protons in the four systems are plotted as a function of
composition, wHA, in Fig. 10.
The most basic anions form the strongest H-bonds. There-
fore, in the neat acetate ionic liquid, the dC2–H signal is strongly
shifted downfield (10.165 ppm), and it is found more upfield
for the other ionic liquids, in the order of decreasing basicity of
ꢀ
ꢀ
ꢀ
the anion: [OMs] (9.015 ppm), [OTf] (8.370 ppm) and [NTf2]
8.053 ppm). Upon addition of an acid, when H-bonded clusters
are formed, the strength of the cation–anion interactions
(
ꢀ
Fig. 5 Proposed structures of anionic dimers, [A(HA)] , in systems of a
general formula [C
2
mim][A]–HA, where A = OAc, OMs and OTf, compared decreases in favour of acid–anion interactions (Fig. 11).
ꢀ
with the discrete nature of HNTf2 and [NTf
system.
2
]
in the [C
2 2
mim][NTf ]–HNTf
2
This is true for acetate, mesylate and triflate systems, which
all follow the same trend, reflecting the formation of [A(HA) ]
2
ꢀ
clusters, and then further upfield shift of the dC2–H signal as
ꢀ
oligomeric [A(HA)
x
]
clusters are formed. In other words,
1
H NMR spectroscopy
ꢀ
complex [A(HA) ] anions are weaker bases than corresponding
H NMR spectra were measured for four liquid systems at ambient [A] , due to charge distribution of a single negative charge over
temperature, each over a range of wHA values. As expected, a greater number of electronegative centres. The [NTf2] anion
x
1
ꢀ
ꢀ
chemical shifts of the acidic protons in the [C mim][OAc]–HOAc is essentially unable to associate with HNTf , so the d
C2–H
2
2
1
Fig. 6 H NMR spectra (400.112 MHz, 27 1C, neat, external d
figure.
6
-DMSO lock) of the [C
2
mim][OAc]–HOAc system; the wHOAc values are marked in the
Phys. Chem. Chem. Phys., 2014, 16, 23233--23243 | 23239
This journal is ©the Owner Societies 2014

View Article Online
Paper
PCCP
1
Fig. 7 H NMR spectra (400.112 MHz, 27 1C, neat, external d
6
-DMSO lock) of the [C
2
mim][OMs]–HOMs system; the wHOMs values are marked in the
figure.
1
Fig. 8 H NMR spectra (400.112 MHz, 27 1C, neat, external d
6
2
-DMSO lock) of the [C mim][OTf]–HOTf system; the wHOTf values are marked in the figure.
Magnification of a downfield part of a spectrum, showing low-intensity signals for compositions of lower acid content.
2
3240 | Phys. Chem. Chem. Phys., 2014, 16, 23233--23243
This journal is © the Owner Societies 2014

View Article Online
PCCP
Paper
1
Fig. 9 H NMR spectra (400.112 MHz, 27 1C, neat, external d
6
-DMSO lock) of the [C
2
mim][NTf
2
]–HNTf
2
system; the wHNTf are marked in the figure.
2
Fig. 11 H-bonding in a neat ionic liquid (wHA = 0.50) and in an equimolar
mixture of ionic liquid and acid, sharing the same conjugate base
2
(wHA = 0.67), exemplified for the [C mim][OAc]–HAc system.
which may be assumed to be the chemical shift for an effec-
+
tively non H-bonded C2 proton in a neat [C mim] ionic liquid.
2
1
ꢀ
H NMR chemical shifts for the acidic protons in [A(HA)x]
Fig. 10 Changes in chemical shift of C2–H proton, dC2–H, as a function of complexes, dAꢁ ꢁ ꢁHꢁ ꢁ ꢁA, are plotted as a function of composition,
composition, wHA, for four systems of a general formula [C
2
mim][A]–HA,
w
HA, in Fig. 12.
where: A = OAc ( ), OMs ( ), OTf (’), or NTf ( ). Dashed curves are only
2
As the composition changes, the observed signal is
guidelines for the eye.
an average of several acidic proton signals, derived from
H-bonded clusters in dynamic equilibria (eqn (6)). This has
signal is constant up to wHNTf2 D 0.70 (two moles of acid per been demonstrated experimentally in low-temperature NMR
3
8
one mol of ionic liquid). Above this concentration, it is possible studies on the pyridine–acetic acid system by Golubev et al.,
ꢀ
ꢀ
that the acid and anions begin to agglomerate via weak where signals corresponding to [OAc(HOAc)] , [OAc(HOAc) ] ,
2
N–Hꢁ ꢁ ꢁO bonding (Fig. 3), which leads to further weakening and to acid dimers [(HOAc)
2
], were observed.
of cation–anion interaction. Analogous behaviour was observed for acetate, mesylate and
For both triflate and bistriflamide systems, at high triflate systems. With increasing acid content, the dAꢁꢁꢁHꢁꢁꢁA signals
acid concentrations, the dC2–H signals achieve identical values, shift downfield, to plateau around wHA = 0.67. This corresponds to a
This journal is ©the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 23233--23243 | 23241

View Article Online
Paper
PCCP
Brønsted acids dissolved in ionic liquids typically form
ꢀ
anionic H-bonded clusters, [A(HA)
x
] . AN measurements are
sensitive to the structure of the liquids, and reflect how acidity
changes with changing liquid structure. The value of x depends
ꢀ
on the basicity of A in the system; for acetate x = 1 or 2, for
mesylate or triflate x = 1, and for bistriflamide x = 0 (the clusters
do not form).
Unlike acids with stronger conjugate bases, HNTf
2
does aggre-
] -based ionic liquids, but is solubi-
lised in a molecular form. At low concentrations, [C mim][NTf ]–
HNTf system is more acidic than the [C mim][OTf]–HOTf system,
ꢀ
gate with the anions in [NTf
2
2
2
2
2
but at very high HOTf concentration the order is reversed
see Fig. 4).
(
1
H NMR chemical shifts of the acidic protons, dAꢁ ꢁ ꢁHꢁ ꢁ ꢁA, do
not correspond to the strength of Brønsted acidity in the system,
but rather inform on the relative strength of H-bonding between
acid and anions in the anionic clusters.
In summary, a simple way to manipulate and quantify
Brønsted acidity of acid–ionic liquid mixtures was demon-
strated, and measured AN values were related to the liquid
speciation, which in turn provided structural and quantitative
underpin for the selection of the Brønsted acidic ionic liquid
system for a catalytic application.
Fig. 12 Changes in chemical shift of Aꢁ ꢁ ꢁHꢁ ꢁ ꢁA protons, dAꢁ ꢁ ꢁHꢁ ꢁ ꢁA, as a
function of composition, wHA, for four systems of a general formula
[C
2
mim][A]–HA, where: A = OAc ( ), OMs ( ), OTf (’), or NTf
2
( ).
Dashed curves are only guidelines for the eye.
ꢀ
maximum in the concentration of [A(HA)] . It is known that,
ꢀ
amongst H-bonded complexes, [A(HA)] dimers have the strongest Acknowledgements
HꢁꢁꢁO, therefore the weakest O–H bonds, and in consequence such
38
The authors acknowledge QUILL and its Industrial Advisory
Board, as well as the EPSRC (Portfolio Partnership Scheme,
Grant EP/D029538/1), for support. Dr P. Nockemann is
acknowledged for useful discussion about crystallography,
and Mr R. W. Murphy for help with NMR studies.
protons are the most deshielded. Moving towards higher wHA
values, dAꢁꢁꢁHꢁꢁꢁA signals shift upfield. The gradual decrease in
dAꢁꢁꢁHꢁꢁꢁA in the acetate system corresponds to the formation of
discrete trinuclear clusters. Further addition of acid leads to a
solution of ionic liquid in acid. Although the existence of higher-
15
order clusters was postulated also in triflate systems, data in
Fig. 12 show no evidence for larger clusters in either the mesylate or
triflate system. It appears that, as the acid content increases beyond
Notes and references
w
HA = 0.67, the equilibrium shown in eqn (6) shifts to the left,
1 (a) K. Bica and R. D. Rogers, Chem. Commun., 2010, 46, 1215;
(b) J. Stoimenovski, D. R. MacFarlane, K. Bica and R. D.
Rogers, Pharm. Res., 2010, 27, 521; (c) J. Stoimenovski, P. M.
Dean, E. I. Izgorodina and D. R. MacFarlane, Faraday Discuss.,
liberating free acid. This behaviour lies on the opposite spectrum to
known behaviour of the [cation]F–HF system, where polynuclear
clusters were recorded.
The d values recorded for the [C mim][NTf ]–HNTf
2
system (0.51 o w r 0.80) remain constant, and are shifted
relatively far upfield when compared to the other studied
42
Aꢁ ꢁ ꢁHꢁ ꢁ ꢁA
2
2
2011, 154, 335.
2 L. Myles, R. Gore, M. ˇS pul ´a k, N. Gathergood and S. J.
Connon, Green Chem., 2010, 12, 1157.
HA
systems. This corresponds to molecular HNTf
ionic liquid.
2
solvated in the
3 E. Margaretta, C. Olmeda and L. Yu, J. Appl. Polym. Sci.,
2012, 127, 2453.
It is worth reiterating that, as clearly evidenced by presented
data, the chemical shift of the acidic proton does not corre-
spond to its ability to protonate a base.
4 (a) K. Sasaki, H. Nagai, S. Matsumura and K. Toshima,
Tetrahedron Lett., 2003, 44, 5605; (b) M. P. Atkins, M. J. Earle,
K. R. Seddon, M. Swad ´z ba-Kwa ´s ny and L. Vanoye, Chem.
Commun., 2010, 46, 1745.
5
K. Matuszek, A. Chrobok, F. Coleman, K. R. Seddon and
M. Swad ´z ba-Kwa ´s ny, Green Chem., 2014, 16, 3463.
M. Yoshizawa, W. Xu and C. A. Angell, J. Am. Chem. Soc.,
2003, 125, 15411.
Conclusions
6
It has been demonstrated that Acceptor Number is a valid and
reliable measure of the Brønsted acidity in ionic liquids, from
very weak to very strong acids. A library of ionic liquid–acid
solutions, characterised by a wide spectrum of acidity, has been
developed.
7 T. A. Zawodzinski and R. A. Osteryoung, Inorg. Chem., 1988,
27, 4383.
8 J. Campbell and K. Johnson, Inorg. Chem., 1993, 32, 3809.
9 G. Driver and K. E. Johnson, Green Chem., 2003, 5, 163.
2
3242 | Phys. Chem. Chem. Phys., 2014, 16, 23233--23243
This journal is © the Owner Societies 2014

View Article Online
PCCP
Paper
1
0 C. M. Burba, N. M. Rocher and R. Frech, J. Phys. Chem. B,
009, 113, 11453.
1 G. A. Olah, T. Mathew, A. Goeppert, B. T o¨ r o¨ k, I. Bucsi,
X.-Y. Li, Q. Wang, E. R. Marinez, P. Batamack, R. Aniszfeld
and G. K. S. Prakash, J. Am. Chem. Soc., 2005, 127, 5964.
2 F. Kohler, H. Atrops, H. Kalali, E. Liebermann, E. Wilhelm,
and M. Swad ´z ba-Kwa ´s ny, Dalton Trans., 2010, 39, 11375;
(c) J. Estager, P. Nockemann, K. R. Seddon, M. Swad ´z ba-
Kwa ´s ny and S. Tyrrell, Inorg. Chem., 2011, 50, 5258;
(d) M. Currie, J. Estager, P. Licence, S. Men, P. Nockemann,
K. R. Seddon, M. Swad ´z ba-Kwa ´s ny and C. Terrade, Inorg.
Chem., 2013, 52, 1710.
2
1
1
1
F. Ratkovics and T. Salamon, J. Phys. Chem., 1981, 85, 2520. 30 (a) M. Schmeisser, P. Illner, R. Puchta, A. Zahl and R. van
3 K. M. Johansson, E. I. Izgorodina, M. Forsyth, D. R. MacFarlane
and K. R. Seddon, Phys. Chem. Chem. Phys., 2008, 10, 2972.
4 R. D. Gillard and R. Ugo, J. Chem. Soc. A, 1966, 549.
5 E. S. Stoyanov, K.-C. Kim and C. A. Reed, J. Phys. Chem. A,
Eldik, Chem. – Eur. J., 2012, 18, 10969; (b) M. Holzweber,
R. Lungwitz, D. Doerfler, S. Spange, M. Koel, H. Hutter and
W. Linert, Chem. – Eur. J., 2012, 19, 288; (c) V. Gutmann, The
Donor-Acceptor Approach to Molecular Interactions, Plenum
Press, New York, 1978.
1
1
2004, 108, 9310.
1
6 T. Ramnial, S. A. Taylor, M. L. Bender, B. Gorodetsky, 31 L. Magna, J. Bild ´e , H. Olivier-Bourbigou, T. Robert and
P. T. K. Lee, D. A. Dickie, B. M. McCollum, C. C. Pye, B. Gilbert, Oil Gas Sci. Technol., 2009, 64, 669.
C. J. Walsby and J. A. C. Clyburne, J. Org. Chem., 2008, 32 (a) J. Estager, A. A. Oliferenko, K. R. Seddon and M. Swad ´z ba-
7
3, 801.
7 J. Foropoulos Jr and D. D. DesMarteau, Inorg. Chem., 1984,
3, 3720.
Kwa ´s ny, Dalton Trans., 2010, 39, 11375; (b) J. Estager,
P. Nockemann, K. R. Seddon, M. Swad ´z ba-Kwa ´s ny and
S. Tyrrell, Inorg. Chem., 2011, 50, 5258; (c) M. Currie,
J. Estager, P. Licence, S. Men, P. Nockemann, K. R. Seddon,
M. Swad ´z ba-Kwa ´s ny and C. Terrade, Inorg. Chem., 2013,
52, 1710.
1
1
2
8 (a) A. Haas, C. Klare, P. Betz, J. Bruckmann, C. Kr u¨ ger, Y.-H.
Tsay and F. Aubke, Inorg. Chem., 1996, 35, 1918; (b) Z. Zak,
A. Ruzicka and C. Michot, Z. Kristallogr., 1998, 213, 217.
1
9 M. Herstedt, M. Smirnov, P. Johansson, M. Chami, J. Grondin, 33 L. Conte, G. Gambaretto, G. Caporiccio, F. Alessandrini and
L. Servant and J. C. Lassegues, J. Raman Spectrosc., 2005, 36, 762. S. Passerini, J. Fluorine Chem., 2004, 125, 243.
0 K. Fujii, Y. Soejima, Y. Kyoshoin, S. Fukuda, R. Kanzaki, 34 P. Bonh oˆ te, A.-P. Dias, N. Papageorgiou, M. Armand,
2
Y. Umebayashi, T. Yamaguchi, S.-I. Ishiguro and T. Takamuku,
J. Phys. Chem. B, 2008, 112, 4329.
K. Kalyanasundaram and M. Gr ¨a tzel, Inorg. Chem., 1996,
35, 1168.
2
1 I. A. Koppel, R. W. Taft, F. Anvia, S.-Z. Zhu, L.-Q. Hu, 35 M. Blesic, M. Swad ´z ba-Kwa ´s ny, J. D. Holbrey, J. N. Canongia
K.-S. Sung, D. D. DesMarteau, L. M. Yagupolskii and
Y. L. Yagupolskii, J. Am. Chem. Soc., 1994, 116, 3047.
Lopes, K. R. Seddon and L. P. N. Rebelo, Phys. Chem. Chem.
Phys., 2009, 11, 4260.
22 I. A. Koppel, R. W. Taft, F. Anvia, S.-Z. Zhu, L.-Q. Hu, 36 J. L. Ferguson, J. D. Holbrey, S. Ng, N. V. Plechkova,
K.-S. Sung, D. D. DesMarteau, L. M. Yagupolskii and
Y. L. Yagupolskii, J. Am. Chem. Soc., 1994, 116, 3047.
K. R. Seddon, A. A. Tomaszowska and D. F. Wassell, Pure
Appl. Chem., 2012, 84, 723.
2
2
2
2
2
2
2
3 B. C. Thompson, O. Winther-Jensen, B. Winther-Jensen and 37 A. M. Magill, K. J. Cavell and B. F. Yates, J. Am. Chem. Soc.,
D. R. MacFarlane, Anal. Chem., 2013, 85, 3521. 2004, 126, 8717.
4 K. E. Johnson, R. M. Pagni and J. Bartmess, Monatsh. Chem., 38 N. S. Golubev, S. N. Smirnov, V. A. Gindin, G. S. Denisov,
2
007, 138, 1077.
H. Benedict and H.-H. Limbach, J. Am. Chem. Soc., 1994,
116, 12055.
39 C. Chiappe and S. Rajamani, Eur. J. Org. Chem., 2011, 5517.
5 L. M. Mihichuk, G. W. Driver and K. E. Johnson, Chem-
PhysChem, 2011, 12, 1622.
6 T. Robert, L. Magna, H. Olivier-Bourbigou and B. Gilbert, 40 F. H. Allen, Acta Crystallogr., Sect. B: Struct. Sci., 2002,
J. Electrochem. Soc., 2009, 156, F115. 58, 388.
7 J. Gr ¨a svik, J. P. Hallett, T. Q. To and T. Welton, Chem. 41 A. Elaiwi, P. B. Hitchcock, K. R. Seddon, N. Srinivasan,
Commun., 2014, 50, 7258.
8 V. Gutmann, The Donor-Acceptor Approach to Molecular
Interactions, Plenum Press, New York, 1978.
Y.-M. Tan, T. Welton and J. A. Zora, J. Chem. Soc., Dalton
Trans., 1995, 3467.
42 G. A. Olah, T. Mathew, A. Goeppert, B. T o¨ r o¨ k, I. Bucsi,
X.-Y. Li, Q. Wang, E. R. Marinez, P. Batamack, R. Aniszfeld
and G. K. S. Prakash, J. Am. Chem. Soc., 2005, 127, 5969.
9 (a) T. A. Zawodzinski Jr and R. A. Osteryoung, Inorg. Chem.,
1989, 28, 1710; (b) J. Estager, A. A. Oliferenko, K. R. Seddon
This journal is ©the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 23233--23243 | 23243