Hydrolysis study of fluoroorganic and cyano-based ionic liquid anions - Consequences for operational safety and environmental stability

Article abstract of DOI:10.1039/c2gc35855g

The hydrolytic stability of ionic liquid anions is a key property with regard to their technical applicability and environmental stability. From a technical point of view hydrolytic processes may lead to reduced durability, diminished technical performance and reduced operational safety in that corrosive and/or toxic hydrolysis products are formed. On the other hand, susceptibility to hydrolytic processes is advantageous where environmental stability and persistency are concerned, since hydrolysis is the most important abiotic degradation pathway in the environment. We investigated the hydrolytic stability of the most common ionic liquid anions, dicyanimide [N(CN) ₂]^-, tricyanmethanide [C(CN)₃]^-, tetracyanidoboranate [B(CN)₄]^-, bis(trifluoromethylsulphonyl)imide [(CF₃SO₂) ₂N]^-, trifluorotris(pentafluoroethyl)phosphate [(C ₂F₅)₃PF₃]^- and 1,1,2,2-tetrafluoroethanesulphonic acid [H(C₂F₄)SO ₃]^-, as a function of pH (1, 7, 9 and 13) and temperature. The results show that there was no difference in hydrolytic stability as recorded for 1-ethyl-3-methylimidazolium (IM12) or for the alkali cations. All the anions were stable under neutral and slightly basic conditions (half-lives at 25 °C ?1 year). In strongly acidic and basic solutions, however, B(CN)₄^-, (CF₃SO₂)₂N ^-, (C₂F₅)₃PF₃^- and H(C₂F₄)SO₃^- were hydrolytically stable, whereas N(CN)₂^- and C(CN)₃^- were not. The kinetics of hydrolysis were recorded and Arrhenius plots were generated for the latter two anions. In addition, their hydrolysis pathways and the resulting products were identified via mass spectrometry. The cytotoxicity of hydrolysed IL solutions towards the mammalian cell line IPC-81 and the identified hydrolysis products (pure compounds) was investigated for a first estimate of their toxicological properties.

Full text of DOI:10.1039/c2gc35855g

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Green Chemistry
Cite this: Green Chem., 2012, 14, 2474
www.rsc.org/greenchem
PAPER
Hydrolysis study of fluoroorganic and cyano-based ionic liquid
anions – consequences for operational safety and environmental stability
a,b
b
b
c
b
Stephanie Steudte, Jennifer Neumann, Ulrike Bottin-Weber, Michael Diedenhofen, Jürgen Arning,
a
b
Piotr Stepnowski and Stefan Stolte*
Received 1st June 2012, Accepted 24th July 2012
DOI: 10.1039/c2gc35855g
The hydrolytic stability of ionic liquid anions is a key property with regard to their technical applicability
and environmental stability. From a technical point of view hydrolytic processes may lead to reduced
durability, diminished technical performance and reduced operational safety in that corrosive and/or toxic
hydrolysis products are formed. On the other hand, susceptibility to hydrolytic processes is advantageous
where environmental stability and persistency are concerned, since hydrolysis is the most important
abiotic degradation pathway in the environment. We investigated the hydrolytic stability of the most
−
−
common ionic liquid anions, dicyanimide [N(CN) ] , tricyanmethanide [C(CN) ] , tetracyanidoboranate
2
3
−
−
[
[
B(CN) ] , bis(trifluoromethylsulphonyl)imide [(CF SO ) N] , trifluorotris(pentafluoroethyl)phosphate
4 3 2 2
(C F ) PF ] and 1,1,2,2-tetrafluoroethanesulphonic acid [H(C F )SO ] , as a function of pH (1, 7, 9
−
−
2
5 3
3
2
4
3
and 13) and temperature. The results show that there was no difference in hydrolytic stability as recorded
for 1-ethyl-3-methylimidazolium (IM12) or for the alkali cations. All the anions were stable under neutral
and slightly basic conditions (half-lives at 25 °C ⇒1 year). In strongly acidic and basic solutions,
−
−
−
−
however, B(CN)4 , (CF SO ) N , (C F ) PF and H(C F )SO were hydrolytically stable, whereas
3
2 2
2
5 3
3
2
4
3
N(CN)2− and C(CN)3− were not. The kinetics of hydrolysis were recorded and Arrhenius plots were
generated for the latter two anions. In addition, their hydrolysis pathways and the resulting products were
identified via mass spectrometry. The cytotoxicity of hydrolysed IL solutions towards the mammalian cell
line IPC-81 and the identified hydrolysis products (pure compounds) was investigated for a first estimate
of their toxicological properties.
perfluorinated compounds like tetrafluoroborate (BF4−),
Introduction
−
hexafluorophosphate (PF6 ), bis(trifluoromethylsulphonyl)imide
−
Interest in ionic liquids (ILs), salts with low melting points
[
[
(CF SO ) N ], or trifluorotris(pentafluoroethyl)phosphate
3
2 2
−
(<100 °C), has risen in recent years. Depending on the appropri-
(C F ) PF ]. The highly fluorinated anions in particular
2
5 3
3
ate combination of cations and anions, they can be thermally and
electrochemically highly stable, be liquid over a wide range of
temperatures, and have negligible vapour pressure. Several appli-
cations and potential implementations of ILs have been
are of exceptional interest owing to their hydrophobicity, low
12,13
viscosity and broad electrochemical window.
Recently,
the super-acidic anion 1,1,2,2-tetrafluoroethanesulphonic acid
−
(
H(C F )SO ) attracted attention because it can be applied in
2 4 3
1
,2
3,4
14
described, e.g. in aluminium plating,
gas compression,
CO2 absorption or as a catalyst in alkylation, etherification
5
,6
7,8
15
dye-sensitized solar cells,
lubrication formulations
and
and isomerization reactions. Also, cyano-based anions such
9
–11
as dicyanimide (N(CN)2−),
16
tricyanmethanide (C(CN)3
− 17
)
separation techniques.
The cations used are mainly organic
and tetracyanidoboranate (B(CN)4−
)
18
form ILs with wide
N-heterocyclic entities like imidazolium, pyridinium, pyrrolidi-
nium and morpholinium, or aliphatic ammonium or phos-
phonium cations. They are usually asymmetrically substituted
with bulky alkyl side chains of different length and/or function-
ality. The anions cover a wide range of types, from halides, alkyl
sulphates, sulphonates, phosphates and phosphonates to
electrochemical windows and low viscosities. ILs based on
these anions therefore offer more efficient mass transport and
higher conductivity, which make them particularly interesting for
electrochemical applications. In particular, B(CN)4− based ILs
have a high potential for applications in dye-sensitized solar
19
cells.
a
The considerable thermal and electrochemical stabilities of
certain ILs are additional reasons for their versatility. From a
technical point of view, hydrolytic stabilities are equally impor-
tant, because hydrolytic processes in aqueous media or caused
by air moisture may lead to reduced durability (degradation),
diminished technical performance (change of properties,
Department of Environmental Analytics, Faculty of Chemistry,
University of Gdańsk, Sobieskiego 18/19, 80-952 Gdańsk, Poland
b
Department of Sustainable Chemistry, Centre for Environmental
Research and Sustainable Technology, University of Bremen, Leobener
Str. UFT, 28359 Bremen, Germany. E-mail: sstolte@uni-bremen.de
c
COSMOlogic GmbH & Co. KG, Burscheider Strasse 515, 51381
Leverkusen, Germany
2474 | Green Chem., 2012, 14, 2474–2483
This journal is © The Royal Society of Chemistry 2012

View Online
formation of corrosive degradation products) and impaired oper-
ational safety in that corrosive and/or toxic hydrolysis products
are formed. On the other hand, susceptibility to hydrolytic pro-
cesses is advantageous in terms of environmental stability and
persistence, since hydrolysis represents the most important
abiotic degradation pathway in the environment. Several IL
−
−
−
anions, such as AlCl4 , PF6 and BF , are known to react with
4
water, leading to the formation of HCl or HF, as the case may
2
0
−
be. It was found that the potential of PF6 to be hydrolysed is
strongly influenced by the type of cation it is attached to.
Whereas KPF is not significantly hydrolysed either in acidic
6
2
1
(
pH < 1), neutral and basic (pH > 12) solutions, LiPF was
6
unstable in propylene carbonate–dimethyl carbonate–water solu-
Fig. 1 Structures of (a) the 1-ethyl-3-methylimidazolium (IM12)
cation and investigated anions as inorganic salts: (b) sodium dicyanimide
−
2−
tions; here, HF, PO F and PO3F were identified as hydroly-
2
2
22
sis products. Studies of primarily imidazolium-based ILs
2 3
(NaN(CN) ), (c) potassium tricyanmethanide (KC(CN) ), (d) potassium
−
showed that BF4 was rapidly hydrolysed in aqueous solutions
tetracyanidoboranate (KB(CN) ), (e) lithium bis(trifluoromethylsulpho-
4
−
even at room temperature. In contrast, PF was more stable
nyl)imide (Li(CF SO ) N), (f) potassium trifluorotris(pentafluoroethyl)-
6
−
3
2 2
under neutral conditions, although PO2F2 was formed as a
phosphate (K(C F ) PF ), (g) potassium 1,1,2,2-tetrafluoroethanesul-
2
5 3
3
product of hydrolysis under acidic or high temperature con-
ditions. The side chain of the cation influences the course of
hydrolysis for both anions. By changing the cation from 1-butyl-
phonic acid (KH(C
2
F
4 3
)SO ).
postulated from an MS identification of the hydrolysis products.
In the context of environmental and operational safety and for a
preliminary hazard assessment, native ILs, hydrolysed solutions
and isolated hydrolysis products were tested for cytotoxicity
using the promyelocytic leukaemia rat cell line IPC-81.
This comprehensive study is addressed to users of ILs in
different fields of applications in order to expand knowledge of
the properties of ILs, as well as their behaviour in their appli-
cations and in the environment.
3
-methylimidazolium (IM14) to 1-methyl-3-octylimidazolium
(IM18) the formation of hydrolysis products was increased,
2
3
perhaps due to decreasing cation–anion interactions. Swatloski
and co-workers were able to isolate the white crystals that were
formed during the synthesis and purification of IM14 PF . X-ray
studies identified IM14 F⋯H O as a hydrolysis product. By
replacing three fluorides in PF6 with perfluorinated alkyl chains
such as C F , the hydrolytic stability could be increased, since
no detectable HF formation was observed after 5 h in boiling
6
24
2
−
2
5
12
water.
Using
a
rapid colorimetric assay, Baker and
Baker tracked the hydrolysis of IM14 BF , IM14 PF ,
4
6
Material and methods
IM14 (CF SO ) N and 1-hexyl-3-methylimidazolium (IM16)
3
2 2
(
C F ) PF by means of the pH-dependent colour change of a
2 5 3 3
Chemicals
−
reporter dye. Whereas at 25 °C only BF4 showed a pH
decrease, this was demonstrated at 50 °C for PF after 24 h and
after one week for (C F ) PF
−
Lithium bis(trifluoromethylsulphonyl)imide (Li(CF SO ) N),
sodium dicyanimide (NaN(CN) ), potassium tetracyanidobora-
nate (KB(CN) ), potassium trifluorotris(pentafluoroethyl)phos-
phate (K(C F ) PF ), potassium tricyanmethanide (KC(CN) ),
1-ethyl-3-methylimidazolium tricyanmethanide (IM12 C(CN)3),
1-ethyl-3-methylimidazolium dicyanimide (IM12 N(CN)2),
1-ethyl-3-methylimidazolium tetracyanidoboranate (IM12 B(CN)4),
1-ethyl-3-methylimidazolium bis(trifluoromethylsulphonyl)imide
(IM12 (CF SO ) N), 1-ethyl-3-methylimidazolium trifluorotris-
(pentafluoroethyl)phosphate (IM12 (C F ) PF ) were kindly
provided by Merck KGaA (Darmstadt, Germany). Potassium
1,1,2,2-tetrafluoroethanesulphonic acid (KH(C F )SO ) and
1-ethyl-3-methylimidazolium 1,1,2,2-tetrafluoroethanesulphonic
acid (IM12 H(C F )SO ) were obtained from IoLiTec Ionic
Liquids Technologies GmbH (Heilbronn, Germany).
6
3
2 2
−
− 25
and (CF SO ) N . Even
2
5 3
3
3
2 2
2
−
−
though the hydrolysis of BF4 and PF6 has been thoroughly
investigated, there are few detailed studies investigating the
hydrolytic processes undergone by the most common IL anions
4
2
5 3
3
3
like N(CN)2−, C(CN) , B(CN)4 , (CF SO ) N , (C F ) PF
−
−
−
−
3
3
2 2
2
5 3
3
−
and H(C F )SO (Fig. 1). Recently, Neumann et al. showed
that none of the anions (except H(C F )SO , which has not yet
been investigated) is either aerobically or anaerobically bio-
degradable, which may have unforeseeable consequences for the
environment in terms of persistence. Therefore, an investi-
gation of the hydrolytic degradation of the different IL anion
species in aquatic systems at the pH values normally found in
the environment is crucial for the hazard assessment of ionic
liquids.
2
4
3
−
2
4
3
3
2 2
2
5 3
3
2
6
2
4
3
2
4
3
In order to evaluate the hydrolytic stability of IL anions under
environmental conditions as well as in their applications, tests
were performed according to OECD guideline 111. The hydroly-
sis at different pH values (pH 1, 7, 9, and 13) of these anions as
Hydrochloric acid, anhydrous sodium carbonate and sodium
bicarbonate were purchased from the Sigma-Aldrich Corporation
(Deisenhofen, Germany), acetonitrile (HPLC grade) was
obtained from VWR International (Darmstadt, Germany),
potassium chloride from Merck KGaA (Darmstadt, Germany),
sodium hydroxide from Across Organics (Geel, Belgium),
potassium dihydrogen phosphate from Carl Roth GmbH & Co.
KG (Karlsruhe, Germany) and boric acid from Riedel de Haen
GmbH (Seelze, Germany).
+
+
+
inorganic salts (Li , Na or K ) and combined with the 1-ethyl-
-methylimidazolium (IM12) cation is determined by means of
3
ion chromatographic measurements. Half-lives, activation ener-
gies and frequency factors can be estimated from kinetic studies
at different temperatures. Mechanisms of hydrolysis are
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 2474–2483 | 2475

View Online
Hydrolysis
where t is the time, c is the concentration of the substance at
t
time t and c is the initial concentration. The half-life of the reac-
tion (t0.5) is given by
0
The hydrolysis tests were performed according to OECD guideline
2
7
111. The buffer with pH 1.0 was prepared using 47.5 mL 0.2 M
HCl, 25 mL 0.2 M KCl and made up to 100 mL with distilled
water. 29.63 mL 0.1 M NaOH and 50 mL 0.1 M KH PO diluted
ln 2
t0:5 ¼
ð2Þ
2
4
k
with distilled water to 100 mL were used to obtain the buffer with
pH 7.0. To prepare the buffer with pH 9.0, distilled water was
added to a mixture of 21.30 mL 0.1 M NaOH and 50 mL 0.1 M
H BO in 0.1 M KCl in a 100 mL flask. A mixture of 132 mL
The correlation of temperature (T) and k can be expressed by
the Arrhenius equation:
E
RT
A
3
4
k ¼ Aeꢀ
ð3Þ
0
.2 M NaOH and 50 mL 0.2 M KCl was used to prepare the
buffer with pH 13. The pH values of the buffers were checked
and adjusted using small amounts of conc. HCl or NaOH. For the
preliminary test four replicates of 10 mL solutions containing
where R is the gas constant, A the frequency factor of the reac-
tion and E the activation energy of the reaction.
A
By determining these parameters the rate constant and there-
fore the half-life of the reaction can be calculated for any
temperature.
1
00 μM IL and the inorganic salt were prepared in all buffers. The
first replicate was measured by ion chromatography (a Metrohm
8
2
81 Compact IC system with an online eluent degasser, a
0 μL injection loop, a Metrosep A supp ion exchange column
(dimensions – 50 × 4.0 mm ID and 5 μm mean particle size)
Identification of degradation pathways
coupled with a Metrosep A Supp 4/5 Guard and a Metrosep
RP Guard, a self-regenerating Suppressor Module (MSM) and a
In order to determine free cyanide, a cyanide cuvette test
LCK 315 (Hach Lange GmbH, Düsseldorf, Germany) was
applied according to the procedure given by the supplier.
CO -suppressor (MCS),
a
conductometric detector and
2
Metrohm software (MagICNet version 1.1 compact), all purchased
from Metrohm, Herisau, Switzerland) to obtain the initial concen-
tration of IL anions using the method reported in ref. 28. The
eluent was 3.2 mM Na CO , 1.0 mM NaHCO and varying per-
−
The CN content was determined for samples from preliminary
tests of cyano-based anions that underwent hydrolysis. 2 mM
solutions of the IL in buffer (pH 1 or 13) were hydrolysed
according to results of kinetic studies and measured by mass
spectrometry. The pH of alkaline samples was adjusted with
conc. HCl prior to MS measurements. The mass spectrometer
used was an electrospray ionization MS equipped with an ion
trap detector (Bruker-Daltonic GmbH, Bremen, Germany).
Mass spectra were acquired in negative and positive ion
modes in the m/z 50–300 scan range. The ESI source conditions
were set to a capillary voltage of 3500 V, the drying gas flow
2
3
3
centages of acetonitrile, depending on the anion investigated. The
−1
flow rate was adjusted to 0.7 mL min . The limits of detection
and quantification were determined to be <0.10 μM and 0.30 μM
28
respectively. For quantification of the IM12 cation a VWR
Hitachi HPLC-UV system was used containing the L2130 HTA-
pump, L2130 degasser, L2200 autosampler, L2300 column oven,
L2450 diode array-detector and the EZChrom Elite software. As
the stationary phase a hydrophilic interaction liquid chromato-
graphy column (HILIC, Multospher 100 Si 5 μm, 125 × 4.6 mm)
with a guard column purchased from CS–Chromatographie
Service GmbH (Langerwehe, Germany) was used with a mobile
phase comprised of 65 v/v% methanol (HPLC grade) and 35 v/v%
aqueous K HPO solution (40 mM). The system was operated
−1
rate to 5 L min , the drying gas temperature to 300 °C and the
nebulizer to 70 psi.
Toxicity of hydrolysis products
2
4
−1
at a flow rate of 1 mL min and 40 °C oven temperature and a
detection wavelength of 211 nm was used for quantification of
IM12 with limits of detection and quantification of 8 and 22 μM.
The other three replicates were stored in headspace vials and
kept closed under an argon atmosphere for five days at 323 ± 0.5 K
in a water bath (GFL mbH, Burgwedel, Germany). Afterwards,
the final IL concentration of samples was measured again by ion
chromatography or HPLC. If IL hydrolysis was less than 10%,
the substance was considered to be hydrolytically stable with a
half-life of >1 year at 298 K and no further tests were carried
out. If hydrolysis exceeded 10%, advanced kinetic tests were
performed: 100 μM of the substance were again dissolved in the
buffer, and stored at three different temperatures (283, 298, 310,
The cytotoxicity of the hydrolysed solutions and identified
hydrolysis products (pure compounds) was determined using
29
the promyelocytic leukaemia rat cell line IPC-81. Cultures of
IPC-81 were grown in RPMI medium (with L-glutamine, without
NaHCO3, supplemented with 1% penicillin–streptomycin
and 1% glutamine, pH 7) with 10% horse serum at 37 °C
(5% CO2).
The cytotoxicity assay was carried out according to Ranke
30
et al. Cell viability was measured using a colorimetric assay
for 96-well plates with 2-(4-iodophenyl)-3-(4-nitrophenyl)-
5-(2,4-disulphophenyl)-2H-tetrazolium monosodium salt (WST-1)
reagent. Each plate contained blanks (no cells), controls (no toxi-
cants), and the test substance in a 1 : 1 dilution series. Stocks of
identified hydrolysis products were prepared in a culture medium
with max. 0.5% dimethylsulphoxide (DMSO) to improve the
solubility of the substances. This DMSO concentration has been
proven not to be cytotoxic (data not shown). A 300 mM solution
of the inorganic salt in buffer (pH 1) was kept at a constant
temperature for a fixed time, corresponding to the results
obtained from the kinetic studies, to prepare approximately 50
3
23 or 343 K). The concentration was determined using ion
chromatography and checked at regular intervals. Assuming
pseudo-first order reaction kinetics, the rate constant (k) can be
obtained by linear regression according to eqn (1):
ꢀ
ꢁ
ct
ꢀ
ln
¼ kt
ð1Þ
c0
2476 | Green Chem., 2012, 14, 2474–2483
This journal is © The Royal Society of Chemistry 2012

View Online
and 100% hydrolysed solutions. The solution was further
neutralized with a small amount of conc. NaOH to pH 7. For the
test, IPC-81 cells at a concentration of 15 × 10 cells mL
the anion and the 1-ethyl-3-methylimidazolium (IM12) cation
were determined by ion chromatographic and HPLC measure-
ments, respectively. No hydrolysis was found for the IM12
cation in all tested buffer systems (data not shown). The recorded
hydrolysis rates (in %) for the anions with inorganic and 1-ethyl-
3-methylimidazolium (IM12) cations are shown in Fig. 2. In
general, there was no significant difference in anion stability
between the inorganic salt and the corresponding IL. The type of
cation therefore seems to play a minor role in the hydrolysis of
the anion. Under environmental conditions (pH 7 and 9), all the
anions tested were degraded by <10%. In strongly acidic (pH 1)
5
−1
(in RPMI with 8% foetal calf serum) were incubated for 44 h in
9
6-well plates in the presence of the test substance and for an
additional 4 h in the presence of WST-1 reagent. Cell viability,
assessed as the ability to reduce WST-1, was observed photo-
metrically at 450 nm in a microplate reader (MRX, Dynatech
Laboratories, Chantilly, USA). The cytotoxicity of the com-
pounds was expressed as the percentage of cell viability
measured as WST-1 reduction compared to controls. Each dose–
response curve was recorded for at least 9 parallel dilution series
on three different 96-well plates. Positive controls with carbenda-
zim were checked at regular intervals.
−
−
or basic (pH 13) solutions only B(CN) , (CF SO ) N ,
4
3
2 2
−
−
(C F ) PF
and H(C F )SO
were hydrolytically stable.
2
5 3
3
2
4
3
A high rate of hydrolysis was found for IM12 N(CN) at pH 1
2
(100%) and at pH 13 (81%). The same applied to IM12 C(CN)3,
but degradation was greater at pH 13 (100%) than at pH 1
(54%).
Computational calculations
The pK predictions were performed using a linear free energy
a
relationship (LFER) based on the free energy differences
Kinetics
between the ions and the protonated forms at infinite dilution in
3
1
water at 25 °C,
as implemented in the COSMOtherm
In order to calculate the rate constant (k) for IM12 N(CN) and
IM12 C(CN)3 the course of hydrolysis was tracked (at least
2
32
program. The free energies of the species in water, used in
this model, are obtained from the COSMO-RS method.
32
6
points between 10% and 90% degradation) at three different
COSMO-RS is a quantum chemically based statistical thermo-
dynamics procedure, which enables the calculation of the chemi-
cal potential and related properties of molecules in mixtures.
The underlying quantum chemical calculations were per-
formed at the density functional theory (DFT) level, utilizing
temperatures (283, 298, 310, 323 or 343 K) at pH 1 and pH 13.
The results are plotted according to eqn (1) (Fig. 3). As
expected and indicated by the linear correlation, the reaction
follows pseudo-first-order kinetics. The slope of the straight line
obtained by linear regression represents the rate constant of the
hydrolysis.
3
3
the conductor-like screening model (COSMO),
the BP
34–36
functional
and a triple-ζ valence polarized basis set (TZVP).
tures were fully optimized. The TURBOMOLE program
with RI (resolution of identity) approximation
The results and the half-lives calculated according to eqn (2)
are summarized in Table 1. At pH 1 and 283 K the half-life of
3
7,38
All struc-
N(CN)2− is about 10.5 h (3.75 × 10 ± 983 s), at 298 K it
4
39
package was used for this task.
3
−
decreases to <2 h (6.49 × 10 ± 35.1 s) and at 310 K N(CN)2 is
rapidly hydrolysed to 50% within approximately 20 min (1.27 ×
3
1
0 ± 8.88 s). Hydrolysis at pH 13 is slower by comparison. The
Results
rate constant for the same temperature (310 K) decreases from
−
4
−6 −1
−6
5
.44 × 10 ± 3.79 × 10
s
at pH 1 to 2.16 × 10 ± 5.69 ×
Preliminary tests
−9 −1
1
0
s
at pH 13. Owing to the reduced rate constants, the half-
life of N(CN)2 in basic media at 310 K, 323 K and 343 K rises
to ca. 4 days (3.20 × 10 ± 841 s), 1.5 days (1.21 × 10 ± 259 s)
and 14 h (4.92 × 10 ± 91.8 s) respectively.
−
Preliminary tests for the hydrolytic stability of the inorganic salts
and ILs were performed at 323 K for 5 days in different buffer
systems (pH 1, 7, 9, 13). The initial and final concentrations of
5
5
4
Fig. 2 Results of preliminary hydrolysis tests for different anions (N(CN) −, C(CN) −, B(CN) − −) as
−
−
2 3 4 3 2 2 2 5 3 3 2 4 3
, (CF SO ) N , (C F ) PF and H(C F )SO
alkali salts (left) and IM12 salts (right) at different pH values (1, 7, 9 and 13) (n.d. = not determined).
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 2474–2483 | 2477

View Online
Fig. 3 Hydrolysis kinetics at different temperatures of (a) IM12 N(CN)
IM12 C(CN) at pH 13.
2
at pH 1; (b) IM12 N(CN)
2
at pH 13; (c) IM12 C(CN)
3
at pH 1; and (d)
3
Table 1 Reaction rate constants (k) and half-lives (t0.5) for the pseudo-
st-order hydrolysis kinetics of the IL anions investigated (cation: IM12)
at different pH and temperatures
Using the experimentally determined rate constant at different
1
temperatures the activation energies (E ) and frequency factors
A
(
A) of the hydrolysis reactions are accessible (Table 2) via the
pH T [K] k [s−
1
]
t0.5 [s]
Arrhenius equation (eqn (3)). Even though the hydrolysis of
N(CN)2− at pH 1 was the fastest (highest rate constant at 310 K),
the calculated activation energy was the highest (99.9 ±
Anion
−
−5
−7
−7
−6
−9
−8
−8
−9
−7
−7
−9
−8
−7
4
N(CN)
2
1
1
1
1
283
1.85 × 10 ± 4.85 × 10
1.07 × 10− ± 5.77 × 10
3.75 × 10 ± 983
4
3
2
3
98
10
6.49 × 10 ± 35.1
−1
−4
3
0.5 kJ mol ). The opposite observation was made for the reac-
5.44 × 10 ± 3.79 × 10
2.16 × 10− ± 5.69 × 10
5.71 × 10− ± 1.22 × 10
1.41 × 10− ± 2.63 × 10
1.27 × 10 ± 8.88
tion of N(CN)2− in alkaline media. Here, the activation energy
6
5
3
3
310
3.20 × 10 ± 841
6
5
−1
3
3
23
43
1.21 × 10 ± 259
was the lowest (48.6 ± 0.1 kJ mol ), but also the frequency
5
4
4.92 × 10 ± 91.8
−1
−
−7
5
factor (359.2 ± 10.9 s ) of the reaction was small. Generally,
C(CN)
3
310
7.94 × 10− ± 8.06 × 10
8.73 × 10 ± 8855
6
5
the frequency factors found for the reactions under acidic con-
ditions were higher than those recorded at pH 13.
3
3
23
43
3.89 × 10 ± 1.09 × 10
1.78 × 10 ± 4966
3.04 × 10− ± 2.03 × 10
1.92 × 10− ± 9.46 × 10
5.83 × 10− ± 3.20 × 10
2.08 × 10− ± 1.08 × 10
6
2.28 × 10 ± 152
4
6
5
298
3.60 × 10 ± 1772
6
5
3
3
10
23
1.19 × 10 ± 652
5
4
3.34 × 10 ± 173
Degradation pathways
C(CN) − behaves in the opposite manner. In an acidic
3
To identify the degradation products of the anions a cuvette
test for free cyanide anions was performed. None of the
samples tested, N(CN)2− and C(CN)3− in either of the buffer
−7
environment the hydrolysis rate constant is 7.94 × 10 ± 8.06 ×
−
9 −1
1
1
0
0
s
(310 K), whereas at pH 13 its value is higher (5.83 ×
−6
−8 −1
−
± 3.20 × 10 s ). The half-lives for hydrolysis at pH 1 are
systems (pH 1 and 13), showed any significant CN content
5
5
−1
ca. 10 days (8.73 × 10 ± 8855 s), 2 days (1.78 × 10 ± 4966 s)
and 6 h (2.28 × 10 ± 152 s) at 310 K, 323 K and 343 K respec-
(<0.1 mg L ) after 5 days at 323 K.
4
During the MS measurements, ions with an m/z− = 84 were
tively. As indicated by the greater rate constants, the half-lives in
additionally detected in the negative ion mode in samples of
5
−
alkaline solutions fall to approximately 4 days (3.60 × 10 ±
N(CN)2 in both buffer systems (pH 1 and 13). Samples of the
5
C(CN)3− anion showed anions with m/z− = 65, 108 and 126.
1
772 s) at 298 K, 33 h (1.19 × 10 ± 652 s) at 310 K and 9 h
4
(3.34 × 10 ± 173 s) at 323 K.
It is known that nitriles can be hydrolysed by either acid- or
2478 | Green Chem., 2012, 14, 2474–2483
This journal is © The Royal Society of Chemistry 2012

View Online
Table 2 Equations resulting from linear regression analysis, obtained by plotting −ln(k) against T− , and the coefficient of determination (R ), as
1
2
well as the activation energy (E
A
) and frequency factor (A) of the Arrhenius equation (eqn (3)), for all the investigated IL anions (IM12 as counter
cation) at different pH
2
[kJ mol−1
A [s−1
]
Anion
pH
1
Equation of linear regression
R
E
A
]
−
13
12
9
N(CN)
2
y = 12 017 × −31.2
y = 5839 × −5.88
y = 11 746 × −23.8
y = 9159 × −17.5
0.994
0.987
0.999
0.999
99.9 ± 0.5
48.6 ± 0.1
97.7 ± 0.3
76.2 ± 0.2
3.5 × 10 ± 6.9 × 10
359.2 ± 10.9
2.2 × 10 ± 2.6 × 10
1
1
1
3
−
10
C(CN)
3
7
6
3
4.1 × 10 ± 3.7 × 10
Fig. 4 Postulated hydrolysis degradation pathway for the N(CN) −
anion: the framed anions were detected in the negative ion mode of MS-measure-
2
ments; the assumed degradation pathway is highlighted in the grey hatched box.
base-initiated reactions to form carboxamides, which will further
hydrolyse in a second step to carboxylic acids. By applying this
knowledge to the mass-to-charge ratios in the MS spectra,
possible degradation pathways for N(CN)2 and C(CN)3− are
shown in Fig. 4 and 5 respectively. For both anions only parts of
the proposed degradation mechanism could be demonstrated by
the MS measurements (framed anions). The part of the pathway
that is assumed on the basis of theoretical considerations is
shown in a grey hatched box.
hydrolysis products, in vitro cytotoxicity studies were performed
using the rat leukaemia cell line IPC-81. The half-maximal effec-
tive concentrations (EC ) obtained are summarized in Table 3
50
−
(first column). Some of the identified hydrolysis products are
common chemicals with high production volumes and further
available toxicological data, which are also shown in Table 3.
In general the degradation products and IL anions tested as
inorganic salts or combined with IM12 cations exhibited a low
acute cytotoxicity. The EC50 values of both anions are lower for
the IM12 salts than for the inorganic salts. This could be due to
40
the known cytotoxic effect of the IM12 cation. As the results
Toxicity of hydrolysis products
show, N(CN)2− has a lower cytotoxic potential than C(CN) .
−
3
In order to evaluate the acute toxicological behaviour of the
The identified hydrolysis products of each anion exhibited discon-
IL and inorganic salts as well as that of selected identified
tinuous cytotoxic behaviour. Whereas no EC50 was recorded for
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 2474–2483 | 2479

View Online
Fig. 5 Postulated hydrolysis degradation pathway for the C(CN) −
anion: the framed anions were detected in the negative ion mode of MS-measure-
3
ments; the assumed degradation pathway is highlighted in the grey hatched box.
the cyanourea sodium salt, carbamylurea (biuret), urea, 2-cyano-
acetamide and acetamide up to the highest tested concentration of
as inorganic salts, in order to exclude effects caused by the IL
cation. The starting concentration of each solution was set at
100%. The resulting concentrations, obtained by 1 : 1 dilution,
were then calculated as a % from this relative concentration. The
hydrolysed solutions of both anions turned out to be less toxic
than the native anions. The respective EC50 values for the
N(CN)2− anion are 50%, 72% and 68% (relative to the starting
concentration of each solution) for the 0%, 50% and 100%
hydrolysed solutions. The differences are significantly greater for
the C(CN)3− anion: whereas an approximately 1 : 8 times diluted
solution of the non-hydrolysed solution shows 50% cell viability
−1
−1
1
0 mM (591–1290 mg L ), cyanamide (EC50 = 238 mg L )
−1
and malononitrile (EC50 = 132 mg L ) displayed elevated acute
cytotoxicity. The latter two compounds were also found to be
toxic towards rats after oral application. The other compounds
were classified as irritants (carbamylurea and urea) or harmful
(cyanourea sodium salt, 2-cyanoacetamide, acetonitrile and acet-
amide) according to their effects after oral application to rodents.
The cytotoxic behaviour of a non-hydrolysed, 50% and 100%
hydrolysed solution is illustrated in Fig. 6 for both anions tested
2480 | Green Chem., 2012, 14, 2474–2483
This journal is © The Royal Society of Chemistry 2012

View Online
Table 3 Results of cytotoxicity experiments with IPC-81 cells and from literature for N(CN) −
ILs. Additionally, data for their identified hydrolysis products are shown
and C(CN) −
anions as inorganic salts and native
2
3
EC50 IPC-81 (standard
LD50 oral (rat)
LD50 dermal (rat)
LC50 Daphnia magna
[mg L ]
deviation) [mg L−
1
]
[mg kg
−1
]
[mg kg
−1
]
−1
Hazard symbol
Substance
a
NaN(CN)
IM12 N(CN)
2
1250
1170 (1030–1320)
>1220
a
2
Cyanourea sodium
Harmful
Irritant
b
salt
Carbamylurea
>1110
b
(
biuret)
b
Cyanamide
238 (230–253)
>974
125
8471
84
8200
3.2 (48 h)
3910 (48 h); >10 000 (24 h)
Toxic
Irritant
b
Urea
a
KC(CN)
3
718 (555–979)
258 (234–285)
132 (121–145)
a
IM12 C(CN)
Malononitrile
3
b
14
340
Toxic, dangerous for the
environment
Harmful
Harmful
Harmful
b
c
2
-Cyanoacetamide
>1290
n.d.
>591
1680
2460
7000
b
d
Acetonitrile
Acetamide
2000
3600 (48 h)
b
a
b
c
d
From chemical database http://www.il-eco.uft.uni-bremen.de/. From MSDS. Not rat, but mouse. Not rat, but rabbit.
the linear free energy relationship (LFER) based on the free
energy differences between the ions and the protonated forms at
31
infinite dilution in water at 25 °C, the pK values of the term-
b
inal nitrogen of the CN-group were estimated to be 13.9 for
N(CN)2−, 16.7 for C(CN)3− and 20.7 for B(CN) . The pKb
−
4
values of the central atoms of N(CN)2 and C(CN)3− were calcu-
lated to be less basic than the terminal nitrogen atoms in both
cases. The pKb values are 17.2 for N(CN)2− and 18.5 for
C(CN)3−, respectively.
−
Discussion
Preliminary hydrolysis tests according to OECD guideline 311 at
3
23 K for 5 days permit a first estimate of the hydrolytic stability
of a compound. The results of these tests indicate that the anions
B(CN)4−, (CF SO ) N , (C F ) PF and H(C F )SO were
−
−
−
3
2 2
2
5 3
3
2
4
3
hydrolytically stable under all the test conditions and no
influence of the counter cation (inorganic or IM12) was
observed. Thus, the half-lives of these compounds at 298 K
(
25 °C) can be assumed to be >1 year. Under environmentally
Fig. 6 Comparison of
a
non-hydrolysed solution of NaN(CN)
straight line, EC50 = 50 ± 9%), a 50%-hydrolysed solution of NaN-
CN) (dashed line, EC50 = 72 ± 10%), a 100%-hydrolysed solution of
(dotted line, EC50 = 68 ± 8%) and a non-hydrolysed solution
(dashed-dotted line, EC50 = 12 ± 2%). The highest concen-
tration for each solution is set at 100%. 50%- and 100%-hydrolysed sol-
utions of KC(CN) were inactive (data not shown).
2
relevant conditions (pH 7 and 9) this was also demonstrated for
the other two anions tested – N(CN)2− and C(CN)3−. Hence, the
hydrolytic degradation of all the anions investigated here as an
alternative pathway to their biodegradation is thought to be un-
likely. Studies in buffer systems of pH 1 and 13 were performed
to cover the two extremes that might occur during technical
(
(
2
NaN(CN)
of KC(CN)
2
3
3
−
−
−
applications. Although B(CN)4 , (CF SO ) N , (C F ) PF
3
3
2 2
2
5 3
−
compared to controls (EC50 = 12.0%), no half-maximal effect
could be observed, even for undiluted solutions of the 50% and
and H(C F )SO can be used even under harsh conditions, the
2
4
3
rapid hydrolysis of N(CN)2 and C(CN)3− even at low tempera-
−
1
00% hydrolysed samples.
tures reduces their technical applicability.
The kinetic rate constants of the hydrolysis of these two
anions were determined; the constants displayed different ten-
dencies. Whereas for N(CN)2− k was greater at pH 1 than at pH
13, the opposite was observed for C(CN)3− (k(pH 1) < k(pH 13)).
This may be explained by the different reaction mechanisms
assumed for the hydrolysis in acidic and alkaline solutions. Fol-
lowing the standard textbook reaction mechanisms, hydrolysis is
initiated by the protonation of the terminal nitrogen at low pH,
Theoretical calculations
According to the mechanism of degradation the hydrolytic
process for N(CN)2− and C(CN)3 starts with the protonation of
the terminal nitrogen. Attempting to explain the finding that the
different cyano-based anions exhibit different hydrolytic stab-
−
ilities, we calculated the pK values of the CN-groups. Applying
b
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 2474–2483 | 2481

View Online
whereas in a basic environment direct attack of the hydroxide
ion on the carbon of the nitrile group is the first step. Especially
the second variant might be too simple for the anions investi-
gated in this study. The nucleophilic attack of a negatively
charged anion by a hydroxide ion is expected to be quite un-
favorable and one would expect a more complicated hydrogen
bond assisted mechanism instead. Nevertheless, the discussion
of the partial charges of the cyano nitrogen atoms can be used
for some first considerations: a partial negative charge at the
nitrogen favours acidic hydrolysis, and a partial positive charge
at the carbon supports alkaline hydrolysis. When considering the
resonance structures or by comparing qualitatively the electro-
negativity and the resulting negative partial charges at the N atoms
to be detected by ESI-MS. Further, assuming that the second
nitrile group is hydrolysed in an identical way, urea, carbamic
acid and ultimately ammonia will be formed. These degradation
products could also be identified during hydrolysis of
41
−
HN(CN)2. For the C(CN)3 anion (Fig. 5) the first hydrolysis
−
step would lead to 2,2-dicyanoacetamide (m/z = 108), which
would be further hydrolysed either at a second nitrile group,
−
resulting in 2-cyanopropandiamide (m/z = 126), or at the amide
group to form 2,2-dicyanoacetic acid. The acid would then
decompose to carbon dioxide and malononitrile (m/z− = 65).
Further hydrolysis at the second and third nitrile groups with the
same mechanism could result in 2-cyanoacetamide, 2-cyano-
acetic acid, acetonitrile, acetamide, acetic acid and finally
methane. For both anions the presence of the first hydrolysis
product (1-cyanoisourea and 2,2-dicyanoacetamide respectively)
was observed even after a few days, indicating that subsequent
hydrolysis reactions are slow.
in the anions N(CN)2−, C(CN)3 and B(CN)4 , it is decreasing
in the given order. Under acidic conditions the reaction rate
therefore decreases as well in the same order owing to the
−
−
lesser protonation capability of B(CN) than of C(CN)3− and
−
4
N(CN)2−. This is in accordance with the observations from our
It is known that some of the identified hydrolysis products are
classified as toxic to man and the environment. In order to
compare this potential with the native compounds, a rapid and
well-established cellular test system was used to screen the cyto-
toxic potential of all the substances investigated in this hydro-
lysis study. The results indicate that most of the identified
hydrolysis products have a lower acute cytotoxicity. Studies in
solutions of pH 1 were also performed in order to simulate the
relevant conditions found in the animal or human stomach. This
can be helpful to further predict the toxicological behaviour of
experiments and the calculated pKb values. N(CN)2−, even
though it still has a high pK , is the strongest base of the three
b
anions investigated and therefore exhibits a higher protonation
potential than the other ions. Additionally, the calculated pKb
values support the proposed reaction mechanism, initiated by the
protonation of the terminal nitrogen. As the values are higher for
the central atoms, a protonation of the terminal groups is
expected to be favoured. For base-induced hydrolysis the partial
charge at the nitrile carbon may be regarded as explaining the
−
−
observed differences in stability. In B(CN)4 the partial charge
the compounds. In fact, hydrolysis of the N(CN)2 anion at
should be slightly negative (electronegativity B = 2.0, C = 2.5)
37 °C is very fast (t0.5 ≈ 20 min) and will most probably occur
in the case of (accidental) oral ingestion. However, the results
obtained in the cytotoxicity test for non-hydrolysed and 50% or
100% hydrolysed solutions showed no significant difference in
their acute cytotoxicological behaviour, even though some of
the identified hydrolysis products are classified as toxic. Since
the concentrations of the single hydrolysis compounds in the
mixture are not known, this can most likely be explained by the
excessively low concentrations. Additional antagonistic mixture
effects are possible. But in order to test this hypothesis, further
studies for quantifying the hydrolysis products will be necessary.
and therefore reduce the possibility of a nucleophilic attack. In
−
−
comparison to N(CN)2 and C(CN)3 this could be an expla-
−
nation for the observed stability of B(CN)4 even at pH 13.
However, this simple model cannot be used to explain the higher
−
−
reaction rate constant for C(CN)3 than for N(CN) , since the
2
−
positive charge should be greater in N(CN)2 . Another approach
that might help to explain the higher rate constant for the
−
hydrolysis of C(CN)3 at pH 13 is the larger number of nitrile
groups in C(CN)3− than in N(CN)2 and therefore the greater
possibility of an attack. However, more detailed theoretical cal-
culations, taking into account the electronic as well as the steric
characteristics of the anions, will be necessary in order to under-
stand this trend. The generally lower frequency factors observed
for hydrolysis at pH 13 may be due to the orientation of the col-
liding entities: a collision between the oxygen of the hydroxide
anion and the nitrile carbon leading to a reaction is less likely
than a collision between a proton and the nitrile nitrogen.
−
Conclusion
−
The hydrolysis of technologically relevant IL anions (N(CN)
,
2
−
−
−
−
−
C(CN) , B(CN) , (CF SO ) N , (C F ) PF and H(C F )SO )
3
4
3
2 2
2
5 3
3
2
4
3
combined with alkaline or IM12 cations was studied. The
results of preliminary tests (323 K for 5 days) showed that
none of the investigated anions was hydrolysed under environ-
mentally relevant conditions. Therefore, the hydrolytic degra-
dation pathway does not represent an appropriate alternative to
biodegradation for removing these anions from aqueous environ-
mental media. It has to be remembered, however, that these tests
were performed under laboratory conditions. Under real environ-
mental conditions, the stability of the investigated anions in
aqueous solution could well be affected by such factors as
changes in dissolved oxygen, different salinities/electrical con-
ductivity, and the presence of reactive oxygen species.
−
By means of the CN cuvette test and MS measurements we
were able to identify the hydrolysis degradation products and
postulate a degradation pathway (Fig. 4 and 5). The results of the
cuvette test indicate that there is no reaction of water with the
−
central atom and hence no formation of CN during hydrolysis.
The first step of our proposed hydrolysis degradation mechanism
−
for N(CN)2 (Fig. 4) would lead to the formation of 1-cyano-
isourea (m/z− = 84). If both cyano groups of the anion were
attacked, biuret would be produced. According to the general
hydrolysis mechanism of nitrile groups, 1-cyanoisourea can be
further hydrolysed to form N-cyano-carbamic acid. This com-
pound is unstable, however, and will decompose further to CO2
N(CN)2− and C(CN)3− were shown to be unstable under
strongly acidic or alkaline conditions. This may limit their
−
and cyanamide. But again, cyanamide is too small (m/z = 41)
2482 | Green Chem., 2012, 14, 2474–2483
This journal is © The Royal Society of Chemistry 2012

View Online
technical applicability under such conditions. Kinetic studies,
including generated Arrhenius equations, helped to estimate the
reaction velocity and half-lives of the compounds at every temp-
14 A. Yokozeki, M. B. Shiflett, C. P. Junk, L. M. Grieco and T. Foo, J. Phys.
Chem. B, 2008, 112(51), 16654–16663.
15 M. A. Harmer, C. Junk, V. Rostovtsev, L. G. Carcani, J. Vickery and
Z. Schnepp, Green Chem., 2007, 9(1), 30–37.
16 D. R. MacFarlane, J. Golding, S. Forsyth, M. Forsyth and G. B. Deacon,
−
erature tested. The hydrolysed solutions of N(CN)2 and
C(CN)3− showed a similar or increased cytotoxicity towards
IPC-81 compared to the non-hydrolysed samples, although some
identified hydrolysis products are known to be toxic to rodents
after oral application. However, further studies, including the
quantification of hydrolysis products, will be helpful for acquir-
ing a more detailed hazard assessment. Summing up, the results
known to date indicate that the stability of these anions is a
matter of great concern in hazard assessments.
Chem. Commun., 2001 (16), 1430–1431.
1
1
7 Y. Yoshida, K. Muroi, A. Otsuka, G. Saito, M. Takahashi and T. Yoko,
Inorg. Chem., 2004, 43(4), 1458–1462.
8 T. Koller, M. H. Rausch, P. S. Schulz, M. Berger, P. Wasserscheid,
I. G. Economou, A. Leipertz and A. P. Froba, J. Chem. Eng. Data, 2012,
57(3), 828–835.
1
9 M. Marszalek, Z. F. Fei, D. R. Zhu, R. Scopelliti, P. J. Dyson,
S. M. Zakeeruddin and M. Gratzel, Inorg. Chem., 2011, 50(22),
11561–11567.
20 C. L. Hussey, T. B. Scheffler, J. S. Wilkes and A. A. Fannin, J. Electro-
chem. Soc., 1986, 133(7), 1389–1391.
21 M. Ponikvar, B. Zemva and J. F. Liebman, J. Fluorine Chem., 2003,
1
23(2), 217–220.
22 A. V. Plakhotnyk, L. Ernst and R. Schmutzler, J. Fluorine Chem., 2005,
26(1), 27–31.
Acknowledgements
1
The authors gratefully thank the whole UFT ionic liquids
research group and Dr Tobias Borrmann for the helpful discus-
sions. They thank IoLiTec GmbH for supplying the chemicals
KH(C F )SO and IM12 H(C F )SO . Furthermore, we are
especially grateful to Merck KGaA for providing chemicals and
for their support within our strategic partnership. The authors
would like to acknowledge the financial support of the European
project MINILUBES (FP7 Marie Curie ITN network 216011-2)
by the European Commission.
23 M. G. Freire, C. M. S. S. Neves, I. M. Marrucho, J. A. P. Coutinho and
A. M. Fernandes, J. Phys. Chem. A, 2010, 114(11), 3744–3749.
2
4 R. P. Swatloski, J. D. Holbrey and R. D. Rogers, Green Chem., 2003,
(4), 361–363.
25 G. A. Baker and S. N. Baker, Aust. J. Chem., 2005, 58(3), 174–177.
5
2
4
3
2
4
3
26 J. Neumann, C. W. Cho, S. Steudte, J. Köser, M. Uerdingen, J. Thöming
and S. Stolte, Green Chem., 2012, 14(2), 410–418.
27 OECD, OECD Guidelines for the Testing of Chemicals, Section 1, 2004,
1–16.
28 S. Stolte, S. Steudte, A. Markowska, J. Arning, J. Neumann and
P. Stepnowski, Anal. Methods, 2011, 3(4), 919–926.
2
9 N. Lacaze, G. Gombaudsaintonge and M. Lanotte, Leuk. Res., 1983,
(2), 145–154.
7
3
0 J. Ranke, K. Mölter, F. Stock, U. Bottin-Weber, J. Poczobutt,
J. Hoffmann, B. Ondruschka, J. Filser and B. Jastorff, Ecotoxicol.
Environ. Saf., 2004, 58(3), 396–404.
Notes and references
1
F. Endres, M. Bukowski, R. Hempelmann and H. Natter, Angew. Chem.,
Int. Ed., 2003, 42(29), 3428–3430.
B. Li, C. H. Fan, Y. Chen, J. W. Lou and L. G. Yan, Electrochim. Acta,
31 A. Klamt, COSMO-RS: From Quantum Chemistry to Fluid Phase Ther-
modynamics and Drug Design, Elsevier, Amsterdam, 2005.
32 A. Klamt and F. Eckert, COSMOtherm [Computer Program], COSMO-
logic GmbH & Co. KG, Leverkusen, Germany, 2011.
33 A. Klamt and G. Schüürmann, J. Chem. Soc., Perkin Trans. 2, 1993 (5),
799–805.
34 A. D. Becke, Phys. Rev. A: At., Mol., Opt. Phys., 1988, 38(6), 3098–3100.
35 S. H. Vosko, L. Wilk and M. Nusair, Can. J. Phys., 1980, 58(8),
1200–1211.
2
2
011, 56(16), 5478–5482.
3
4
5
6
M. Kömpf, Linde Technology 2006, 2006 (January 2006), 24–26.
R. Adler and H. Mayer, DE102005026916A1, 2006, p. 3.
M. Gratzel, J. Photochem. Photobiol., C, 2003, 4(2), 145–153.
S. J. Sun, J. Song, R. X. Feng and Z. Q. Shan, Electrochim. Acta, 2012,
6
9, 51–55.
7
8
I. Minami, Molecules, 2009, 14(6), 2286–2305.
F. Zhou, Y. M. Liang and W. M. Liu, Chem. Soc. Rev., 2009, 38(9),
2
36 J. P. Perdew, Phys. Rev. B, 1986, 33(12), 8822–8824.
37 A. Schafer, C. Huber and R. Ahlrichs, J. Chem. Phys., 1994, 100(8),
5829–5835.
38 K. Eichkorn, F. Weigend, O. Treutler and R. Ahlrichs, Theor. Chem. Acc.,
1997, 97(1–4), 119–124.
39 TURBOMOLE [Computer Program], University of Karlsruhe,
Forschungszentrum Karlsruhe GmbH, TURBOMOLE GmbH, 2007.
40 S. Stolte, J. Arning, U. Bottin-Weber, M. Matzke, F. Stock, K. Thiele,
M. Uerdingen, U. Welz-Biermann, B. Jastorff and J. Ranke, Green
Chem., 2006, 8(7), 621–629.
590–2599.
L. Vidal, M. L. Riekkola and A. Canals, Anal. Chim. Acta, 2012, 715,
9–41.
0 C. F. Poole and S. K. Poole, J. Sep. Sci., 2011, 34(8), 888–900.
1 W. T. Bi, J. Zhou and K. H. Row, Sep. Sci. Technol., 2012, 47(2),
60–369.
9
1
1
1
1
1
3
2 N. V. Ignat’ev, U. Welz-Biermann, A. Kucheryna, G. Bissky and
H. Willner, J. Fluorine Chem., 2005, 126(8), 1150–1159.
3 H. Tokuda, K. Hayamizu, K. Ishii, M. bu Bin Hasan Susan and
M. Watanabe, J. Phys. Chem. B, 2004, 108(42), 16593–16600.
41 S. V. Hill, A. Williams and J. L. Longridge, J. Chem. Soc., Perkin Trans.
2, 1984 (6), 1009–1013.
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 2474–2483 | 2483