Decolorization of ionic liquids for spectroscopy

Article abstract of DOI:10.1021/ac061481t

It has been widely recognized that although ionic liquids should be colorless, they are frequently not. Colored samples appear to be pure by most analytical techniques (e.g., NMR spectroscopy, mass spectrometry, HPLC, and ion chromatography), and there have been many attempts to identify the source of color in our own laboratories and others-after 20 years the best that can be said is that the impurities are at a very low level (probably parts per billion) with very high molar extinction coefficients. In this paper, we do not identify these impurities but describe a practical method for removing them for spectrochemical applications. We clearly note that the method is not "green", but we anticipate that it will only be applied to the small volumes of ionic liquids required for fundamental spectroscopic studies in academia but not in industrial processes.

Full text of DOI:10.1021/ac061481t

Anal. Chem. 2007, 79, 758-764
Technical Notes
Decolorization of Ionic Liquids for Spectroscopy
Martyn J. Earle,† Charles M. Gordon,‡,§ Natalia V. Plechkova,† Kenneth R. Seddon,*,† and
Thomas Welton|
QUILL, The Queen’s University of Belfast, Stranmillis Road, Belfast, BT9 5AG, Northern Ireland, UK, Department of Pure
and Applied Chemistry, University of Strathclyde, Thomas Graham Building, 295 Cathedral Street, Glasgow G1 1XL, UK,
and Department of Chemistry, Imperial College London, London SW7 2AZ, UK
It has been widely recognized that although ionic liquids
should be colorless, they are frequently not. Colored
samples appear to be pure by most analytical techniques
(e.g., NMR spectroscopy, mass spectrometry, HPLC, and
ion chromatography), and there have been many attempts
to identify the source of color in our own laboratories and
otherssafter 20 years the best that can be said is that the
impurities are at a very low level (probably parts per
Figure 1. Typical examples of ionic liquids: (a) 1-butyl-3-meth-
billion) with very high molar extinction coefficients. In this
ylimidazolium bis{(trifluoromethyl)sulfonyl}amide (bistriflamide)
([C4mim][NTf2]); (b) 1-octyl-3-methylimidazolium chloride ([C8mim]Cl);
paper, we do not identify these impurities but describe a
practical method for removing them for spectrochemical
applications. We clearly note that the method is not
(c) 1-hexyl-3-methylimidazolium chloride ([C6mim]Cl); (d) 1-propyl-
2,3,5-trimethylpyrazolium methanesulfonate ([C3tmpyr][OMs]).
“green”, but we anticipate that it will only be applied to
the small volumes of ionic liquids required for fundamen-
tal spectroscopic studies in academia but not in industrial
hexafluorophosphate ([PF6]-) anion, prepared from halide ionic
liquids. Hexafluorophosphate ionic liquids are immiscible with
water (hydrophobic ionic liquids) in general and are frequently
processes.
purified by aqueous extraction.2,3 As the halide is removed (halide
Since the earliest days of ionic liquids, they have been
concentration can be determined by ion chromatography4,5) water
analytically pure to NMR spectroscopy and mass spectrometric
is introduced concomitantly into the ionic liquid, which then itself
microanalysis, but have frequently been yellow, brown or in the
has to be removed. If these ionic liquids containing water from
the washings are heated under vacuum at 65 °C, they undergo
worst cases black in color (Figure 1). (The ionic liquids marketed
as Basionics (BASF) are visibly colored, many being black and
partial hydrolytic decomposition to liberate HF and form complex
oxo- and hydroxophosphates. Additionally, if ionic liquids are
hydrophilic, every time a water wash is carried out, a considerable
amount of ionic liquid is lost in the aqueous layer.
With regard to the chromophoric impurities in ionic liquids,
there have been attempts to decolorize (or remove the chro-
mophores) for the last two decades. It is still not known which
chromophores they contain, except that they must be present at
the ppb level. However, some sources say that the chromophores
are the products of side reactions occurring at high temperatures
(higher than 75 °C de bene esse) in the alkylation reactions when
ionic liquid halides are made.6
opaque (http://www.basionics.com, BASF (2006), Merck KGaA,
Ionic Liquids.)
Overheating when making ionic liquids can cause the color,
but this is not the only source of color (vide infra): color can be
generated in isothermal reactions at room temperature.
Ionic liquids may contain water, halides, and chromophores
(the word “chromophore” is applied here for any sort of impurity
that imparts color to ionic liquids). Both high water and halide
content generally make ionic liquids unsuccessful candidates for
catalytic processes.1 It is possible to minimize the halide content
in ionic liquids by washing them with water and then subsequent
drying in vacuo at 65 °C for 24 h. However, there is a problem
for certain types of ionic liquids, such as those that contain the
(2) Boˆnhote, P.; Dias, A. P.; Papageorgiou, N.; Kalyanasundaram, K.; Gr¨atzel,
M. Inorg. Chem. 1996, 35, 1168-1178.
(3) Huddleston, J. G.; Willauer, H. D.; Swatloski, R. P.; Visser, A. E.; Rogers,
* To whom correspondence should be addressed. E-mail: k.seddon@qub.ac.uk.
† The Queen’s University of Belfast.
‡ University of Strathclyde.
R. D. Chem. Commun. 1998, 1765-1766.
§ Current address: Pfizer Global Research and Development, Ramsgate Rd.,
Sandwich, Kent CT 13 9NJ, UK; e-mail: charles.gordon@pfizer.com.
(4) Billard, I.; Moutiers, G.; Labet, A.; El Azzi, A.; Gaillard, C.; Mariet, C.;
Lutzenkirchen, K. Inorg. Chem. 2003, 42, 1726-1733.
(5) Anderson, J. L.; Ding, J.; Welton, T.; Armstrong, D. W. J. Am. Chem. Soc.
2002, 124, 14247-14254.
|
Imperial College London.
(1) Chauvin, Y.; Mussmann, L.; Olivier, H. Angew. Chem., Int. Ed. Engl. 1996,
34, 2698-2700.
(6) Farmer, V.; Welton, T. Green Chem. 2002, 4, 97-102.
758 Analytical Chemistry, Vol. 79, No. 2, January 15, 2007
10.1021/ac061481t CCC: $37.00 © 2007 American Chemical Society
Published on Web 12/08/2006

According to Gordon and co-workers,7 the possibilities to get
“spectroscopic grade” ionic liquids can be represented in four
ways, notably (i) purification of starting materials, (ii) control of
conditions for quaternisation reactions, (iii) anion exchange, and
(iv) cleaning of the ionic liquid. Some of these procedures have
been taken from the paper by Welton et al.8
and cooling rates for both were 10 °C min-1. The melting point
and glass transitions were defined as peak maxima. Calibration
of the DSC machines was carried out with iridium samples.
Ultraviolet-visible (UV-vis) spectra were recorded on a Perkin-
Elmer Lambda 800 instrument. The samples were contained in
1-mm quartz cuvettes. Distilled dichloromethane was used as a
reference, and ionic liquid was used as a sample. The water
content of ionic liquids was measured with an Aquapal III Karl
Fischer titrator. The compartment solutions were obtained from
Riedel deHae¨n (Hydranal-Coulomat AG and Hydranal-Coulomat
CG). All NMR spectra were recorded at room temperature on a
Bruker Avance spectrometer DPX 300.
When sodium sulfate was used to eliminate water from ionic
liquids, both metal and sulfur analyses were performed (after
insoluble sodium sulfate was separated from the ionic liquid by
filtration): no traces of either sodium or sulfur were detected.
All ionic liquids studied in this work were routinely monitored
by NMR spectroscopy, mass spectrometry, and ion chromatog-
raphy: there was no example where these methods could
distinguish between a colored and a decolorized ionic liquid.
Moreover, Karl Fischer analyses were performed before and after
the decolorization, and in no case did the water content increase
after treatment.
Quaternization. 1-Butyl-3-methylimidazolium Chloride. 1-Meth-
ylimidazole (123 g, 1.50 mol) and 1-chlorobutane (157 g, 1.70 mol)
were placed in a 2-L round-bottomed flask. A slight excess of the
haloalkane was used to guarantee the consumption of 1-meth-
ylimidazole. Ethanenitrile (100 cm3) was added to reduce the
viscosity of the mixture, which was left to stir under reflux at
70 °C for 96 h. The halide salt separated as a second phase from
the ethanenitrile. The excess of ethanenitrile was removed by
decantation.
The halide salt was then recrystallized from ethyl ethanoate.
The volume of ethyl ethanoate used for the recrystallization was
approximately half that of the halide salt. The ethyl ethanoate was
decanted, followed by the addition of fresh ethyl ethanoate, and
this step was repeated twice. After the third cycle, the remaining
ethyl ethanoate (bp 77 °C) and 1-chlorobutane (bp 77 °C) were
removed under reduced pressure at 70 °C and the chloride salt
was finally dried in vacuo at 70 °C. The product was obtained as
a very pale yellow solid (251 g, 1.44 mol, 96.0% yield; mp (DSC),
69 °C; lit. mp (DSC), 69 °C).10
The first two methods focus on the purification of the starting
materials and control of the quaternization process temperature.
The last two methods deal with creating the right conditions for
the synthesis of the final product and its cleaning. Most ap-
proaches in the literature center upon making the ionic liquids
pure in the first place.9 In this paper, the problem of how to remove
colored impurities once generated is tackled, so the proposed
method deals with the cleaning of an ionic liquid. The halide ionic
liquid precursors used in this paper were as follows: 1-butyl-3-
ethylimidazolium chloride ([C4mim]Cl), 1-butyl-2,3-dimethylimi-
dazolium chloride ([C4dmim]Cl), 1-decyl-3-methylimidazolium
bromide ([C10mim]Br), 1-butylpyridinium chloride ([C4py]Cl),
1-decyl-3-methylimidazolium chloride ([C10mim]Cl). The final ionic
liquids include the following: 1-butyl-3-methylimidazolium bistri-
flamide ([C4mim][NTf2]), 1-butyl-3-methylimidazolium tetrafluo-
roborate ([C4mim][BF4]), 1-butyl-3-methylimidazolium hexafluo-
rophosphate ([C4mim][PF6]), 1-butylpyridinium bistriflamide
([C4py][NTf2]), 1-butyl-2,3-dimethylimidazolium bistriflamide ([C4-
dmim][NTf2]), 1-decyl-3-methylimidazolium triflate ([C10mim]-
[OTf]), 1-butyl-3-methylimidazolium triflate ([C4mim][OTf]), and
trihexyl(tetradecyl)phosphonium diisobutylphosphate ([P6 6 6 14]-
[iBu2PO2]).
EXPERIMENTAL SECTION
The ionic liquids used in this paper were prepared at QUILL,
except for trihexyl(tetradecyl)phosphonium diisobutylphosphate,
which was donated by Cytec.
Li[NTf2] was purchased from 3M. 1-Methylimidazole, chloro-
alkanes, bromooctane, pyridine, Na[BF4], Na[PF6], Na[OTf],
dichloromethane, ethyl ethanoate, and 1,2-dimethylimidazole were
purchased from Aldrich. 1-Methylimidazole, 1,2-dimethylimidazole,
and pyridine were distilled before the quaternization step. For the
decolorization of ionic liquids, decolorizing charcoal was pur-
chased from Acros Organics, silica gel for flash chromatography
from Fluorochem, and Celite from Aldrich. Deuteriated solvents
were obtained from Aldrich.
1-Decyl-3-methylimidazolium Chloride. 1-Methylimidazole (123
g, 1.50 mol) and 1-chlorodecane (300 g, 1.70 mol) were placed in
a 2-L round-bottomed flask. A slight excess of the haloalkane was
used to guarantee the consumption of 1-methylimidazole. Ethane-
nitrile (100 cm3) was added to reduce the viscosity of the mixture,
which was left to stir under reflux at 70 °C for 168 h. The halide
salt separated as a second phase from the ethanenitrile. The
excess of ethanenitrile was removed by decantation. The excess
of 1-chlorodecane was removed during recrystallization.
The halide salt was then recrystallized from ethyl ethanoate.
The volume of ethyl ethanoate used for the recrystallization was
approximately half that of the ionic liquid phase. The ethyl
ethanoate was decanted, followed by the addition of fresh ethyl
ethanoate, and this step was repeated twice. After the third cycle,
the remaining ethyl ethanoate (bp 77 °C) was removed under
reduced pressure at 70 °C and the chloride salt was finally dried
Because many of the ionic liquids are either hygroscopic or
deliquescent, the reactions to produce them or use them have to
be carried out under a dry dinitrogen atmosphere. Every ionic
liquid made in this paper was stored in a sealed, dark glass bottle
having been dried in vacuo (5-10 mbar) at 70 °C for 24 h prior
to use.
Differential scanning calorimetry (DSC) traces were recorded
with two different systems from Perkin-Elmer. For measurements
below room temperature down to -100 °C, a Pyris 1 instrument
with dinitrogen cooling was used. For temperature ranges from
room temperature to 300 °C, a DSC 7 was employed. The heating
(7) Gordon, C. M.; McLean, A. J.; Muldoon, M. J.; Dunkin, I. R. In Ionic Liquids
as Green Solvents: Progress and Prospects; Rogers, R. D., Seddon, K. R.,
Eds.; ACS Symposium Series 856; American Chemical Society: Washington,
DC, 2003.
(8) Cammarata, L.; Kazarian, S. G.; Salter, P. A.; Welton, T. Phys. Chem. Chem.
Phys. 2001, 3, 5192-5200.
(9) Nockemann, P.; Binnemans, K.; Driesen, K. Chem. Phys. Lett. 2005, 415,
131-136.
(10) Wilkes, J. S.; Levisky, J. A.; Wilson, R. A.; Hussey, C. L. Inorg. Chem. 1982,
21, 1263-1264.
Analytical Chemistry, Vol. 79, No. 2, January 15, 2007 759

in vacuo at 70 °C. The product was obtained as a yellow solid
(375 g, 1.45 mol, 97% yield).
layer containing the ionic liquid was washed with water (6 × 20
cm3) and then dried with anhydrous sodium sulfate (20 g), which
was removed by filtration. The filtrate was heated in vacuo at 40
°C to remove the excess solvent. The product was obtained as a
pale yellow liquid (147 g, 340 mmol, 97% yield).
1-Butylpyridinium Bistriflamide. [C4py]Cl (60.15 g, 350.4 mmol)
and Li[NTf2] (108.3 g, 377.2 mmol) were dissolved in a mixture
of dichloromethane (200 cm3) and water (20 cm3) and stirred for
6 h. The lower dichloromethane layer containing the ionic liquid
was washed with water (6 × 20 cm3) and then dried with
anhydrous sodium sulfate (20 g), which was removed by filtration.
The filtrate was heated in vacuo at 40 °C to remove the excess
solvent. The product was obtained as a yellow liquid (136 g, 326
mmol, 93% yield).
1-Butyl-2,3-dimethylimidazolium Chloride. 1,2-Dimethylimida-
zole (13.0 g, 135 mmol) and 1-chlorobutane (13.165 g, 142.17
mmol) were placed in a 2-L round-bottomed flask. A slight excess
of the haloalkane was used to guarantee the consumption of 1,2-
dimethylimidazole. Ethanenitrile (100 cm3) was added to reduce
the viscosity of the mixture, which was left to stir under reflux at
70 °C for 288 h. The halide salt separated as a second phase from
the ethanenitrile. The excess of ethanenitrile was removed by
decantation.
The halide salt was then recrystallized from ethyl ethanoate.
The volume of ethyl ethanoate used for the recrystallization was
approximately half that of the ionic liquid phase. The ethyl
ethanoate was decanted, followed by the addition of fresh ethyl
ethanoate, and this step was repeated twice. After the third cycle,
the remaining ethyl ethanoate (bp 77 °C) and chlorobutane (bp
77 °C) were removed under reduced pressure at 70 °C and the
chloride salt was finally dried in vacuo at 70 °C. The product was
obtained as a yellow solid (21.6 g, 125 mmol, 92% yield).
1-Butylpyridinium Chloride. Pyridine (126 g, 1.50 mol) and
1-chlorobutane (157 g, 1.70 mol) were placed in a 2-L round-
bottomed flask. A slight excess of the haloalkane was used to
guarantee the consumption of pyridine. Ethanenitrile (100 cm3)
was added to reduce the viscosity, and the mixture was left to
stir under reflux at 70 °C for 288 h. The halide salt separated as
a second phase from the ethanenitrile. The excess of ethanenitrile
was removed by decantation.
1-Butyl-3-methylimidazolium Tetrafluoroborate.12 [C4mim]Cl (125
g, 716 mmol) was dissolved in 200 cm3 of water, an aqueous
solution (100 cm3) containing Na[BF4] (82.56 g, 751.8 mmol) was
added, and the resulting mixture was left to stir overnight. The
resulting homogeneous mixture ([C4mim][BF4] is hydrophilic at
ambient temperature), was extracted 5 times with dichlo-
romethane (400 cm3 each time), and then dried with anhydrous
sodium sulfate (30 g), which was removed by filtration. The
dichloromethane (bp 40 °C) was then removed by heating at 40
°C under reduced pressure and finally by heating at 60 °C in vacuo
for 24 h. The product was obtained as a yellow liquid (150 g, 665
mmol, 93% yield).
1-Butyl-3-methylimidazolium Hexafluorophosphate. Since the
hydrolysis of hexafluorophosphate is known to be acid catalyzed,13
the synthesis of [C4mim][PF6] was performed by metathesis from
Na[PF6] rather than the acid.
Hexafluorophosphoric acid (250 g, 60 wt % aqueous solution,
1.03 mol) was slowly added to water (250 cm3) in an ice-water
bath. This solution was then added dropwise from a polyethylene
dropping funnel to an equimolar aqueous solution of NaOH (41
g dissolved in 160 cm3 of water), placed in an ice-water bath, to
generate aqueous Na[PF6]. All steps described above are exo-
thermic, and they should be performed with care. Periodically
the pH of the solution was checked with universal pH paper to
ensure that the pH of the final solution was ∼7.
The halide salt was then recrystallized from ethyl ethanoate.
The volume of ethyl ethanoate used for the recrystallization was
approximately half that of the ionic liquid phase. The ethyl
ethanoate was decanted, followed by the addition of fresh ethyl
ethanoate, and this step was repeated twice. After the third cycle,
the remaining ethyl ethanoate (bp 77 °C) and chlorobutane (bp
77 °C) were removed under reduced pressure at 70 °C and the
chloride salt was finally dried in vacuo at 70 °C. The product was
obtained as a yellow solid (224 g, 1.31 mol, 87% yield).
1-Octyl-3-methylimidazolium Bromide.11 1-Methylimidazole (275
g, 3.30 mol) was poured into a 2-L three-necked round-bottomed
flask and placed in an oil bath at 70 °C. Distilled bromooctane
(714 g, 3.70 mol) was added dropwise at a rate to maintain the
reaction temperature below 90 °C. The mixture was left to stir
for 24 h under reflux at 70 °C. The reaction flask was covered
with foil to avoid photolytic side reactions. Ethanenitrile (50 cm3)
was then added to reduce the viscosity, and the halide salt was
then recrystallized by addition of ethyl ethanoate. The volume of
ethyl ethanoate used for the recrystallization was approximately
half that of the ionic liquid phase. The ethyl ethanoate was
decanted, followed by the addition of fresh ethyl ethanoate, and
this step was repeated twice. After the third decanting, the
remaining ethyl ethanoate (bp 77 °C) was removed under reduced
pressure at 70 °C and the bromide salt was finally dried in vacuo
at 73 °C. The product was obtained as a yellow liquid (817 g, 2.97
mol, 90% yield).
[C4mim]Cl (179 g, 948 mmol) was added to the Na[PF6] (167
g, 996 mmol) solution in water, and the resulting mixture was
left to stir overnight. The upper, aqueous phase was removed by
decantation and the product, which appeared as a yellow oily liquid
at the bottom of the round-bottom flask, was dissolved in
dichloromethane (50 cm3) and extracted 7 times with water cooled
to 4 °C (7 × 30 cm3). The dichloromethane was then removed by
heating at 40 °C under reduced pressure and finally by heating
at 60 °C in vacuo. The product was obtained as a yellow liquid
(251 g, 882 mmol, 93% yield).
1-Decyl-3-methylimidazolium Triflate. [C10mim]Cl (90.69 g, 350.4
mmol) and Na[OTf] (64.90 g, 377.2 mmol) were dissolved in water
(200 cm3) and stirred for 6 h. Then the aqueous solution was
extracted with dichloromethane (6 × 100 cm3). The dichlo-
romethane layer was dried with sodium sulfate (20 g), filtered,
and the solvent evaporated under reduced pressure and finally
by heating at 60 °C in vacuo. The product was obtained as a yellow
Metathesis. 1-Butyl-2,3-dimethylimidazolium Bistriflamide. [C4-
dmim]Cl (66.12 g, 350.4 mmol) and Li[NTf2] (108.3 g, 377.2 mmol)
were dissolved in a mixture of dichloromethane (200 cm3) and
water (20 cm3) and stirred for 6 h. The lower dichloromethane
(12) Lancaster, N. L.; Welton, T.; Young, G. B. J. Chem. Soc., Perkin Trans. 2
2001, 2267-2270.
(11) Chan, B. K. M.; Chang, N. H.; Grimmett, M. R. Aust. J. Chem. 1977, 30,
2005-2013.
(13) Swatloski, R. P.; Holbrey, J. D.; Rogers, R. D. Chem. Commun. 2003, 5,
361-363.
760 Analytical Chemistry, Vol. 79, No. 2, January 15, 2007

Figure 2. Experimental apparatus for decolorizing of ionic liquids.
liquid (121 g, 326 mmol, 93% yield).
finally, dichloromethane (200 cm3) was passed in order to elute
the remainder of the ionic liquid. In order to purge ionic liquid
through the column, dinitrogen was passed over the column.
Once eluted, the eluate was dried under reduced pressure and
finally by heating at 60 °C in vacuo.
1-Butyl-3-methylimidazolium Triflate. [C4mim]Cl (61.20 g, 350.4
mmol) and Na[OTf] (64.90 g, 377.2 mmol) were dissolved in water
(200 cm3) and stirred for 6 h. Then the aqueous solution was
extracted with dichloromethane (6 × 100 cm3). The dichlo-
romethane layer was dried with sodium sulfate (20 g), filtered,
and the solvent evaporated under reduced pressure and finally
by heating at 60 °C in vacuo. The product was obtained as a yellow
liquid (90.91 g, 315.4 mmol, 90% yield).
RESULTS
Synthesis and Characterization. Generally the synthesis of
most ionic liquids consists of a quaternization step followed by a
metathesis stepsanion exchange.14 All the procedures used were
standard literature ones with slight modifications.
1-Butyl-3-methylimidazolium Bistriflamide.2 [C4mim]Cl (61.20
g, 350.4 mmol) and Li[NTf2] (108.3 g, 377.2 mmol) were dissolved
in a mixture of dichloromethane (200 cm3) and water (20 cm3),
and the resultant mixture was stirred for 6 h. The lower
dichloromethane layer containing the ionic liquid was washed with
water (6 × 20 cm3) and then dried with anhydrous sodium sulfate
(20 g), which was then removed by filtration. The filtrate was
heated in vacuo at 40 °C to remove the excess solvent. The
product was obtained as a pale yellow liquid (139 g, 331 mmol,
95% yield).
Decolorization of Halide Precursors and Ionic Liquids.
A special column was designed for the decolorization. The column
was packed with the following compounds: Celite on the bottom
to trap charcoal particles, flash chromatographic silica gel in the
middle for decolorizing/removal of polar and inorganic impurities,
and charcoal on the top for decolorizing (Figure 2). The dimen-
sions of the column are illustrated in the Figure 2b.
The ionic liquids were analytically pure. 1H NMR spectra were
recorded for all compounds.15,16 There were no detectable signs
of organic impurities, which was true whether or not the
compounds were colored. The water content, as measured by Karl
Fischer titration for all ionic liquids, was below 20 ppm.17
Quaternization Step. A slight molar excess (1.05-1.1 equiv)
of alkylating agent, usually a 1-haloalkane, must be used. The
removal of unreacted 1-methylimidazole from 1-alkyl-3-methylimi-
dazolium chloride, and 1,2-dimethylimidazole from an 1-butyl-2,3-
dimethylimidazolium chloride, is extremely difficult due to their
high boiling points correspondingly (198 and 202 °C). The reaction
is usually carried out in ethanenitrile. The solvent helps to reduce
the viscosity of the reaction mixture. It was found that it is best
to perform the quaternization step at ∼80-90 °C (conventional
heating) for several days, rather than at higher temperatures for
a shorter period, to avoid charring and alkyl chain scrambling
(eq 1).9,18
The column was similar to the one used in conventional
chromatography. The size of the column was 50 cm long, 42-mm
outer diameter, and 36-mm inner diameter.
An ionic liquid (200 cm3) was mixed with dichloromethane (600
cm3) (v/v 1:3). In the case of halide salts that were more viscous,
the volume ratio between a halide salt and dichloromethane was
1:6; the dichloromethane was added in order to decrease the
viscosity of the ionic liquid. If an ionic liquid was only slowly
miscible with dichloromethane, it could be heated to less than
the boiling point of dichloromethane (40 °C), to speed up the
dissolution.
Equation 1 shows the scrambling of the 1-R1-3-R2-imidazolium
cation.
(14) Carmichael, A. J.; Deetlefs, M.; Earle, M. J.; Fro¨hlich, U.; Seddon, K. In
Ionic Liquids as Green Solvents: Progress and Prospects; Rogers, R. D.,
Seddon, K. R., Eds.; ACS Symposium Series 856; American Chemical
Society: Washington, DC, 2003; pp 14-31.
(15) Avent, A. G.; Chaloner, P. A.; Day, M. P.; Seddon, K. R.; Welton, T. J. Chem.
Soc., Dalton Trans. 1994, 3405-3413.
(16) Stark, A. Ph.D., The Queen’s University of Belfast, 2001.
The column was pretreated with dichloromethane (200 cm3)
prior to its use with ionic liquids. Then the dichloromethane
solution of the ionic liquid was passed through the column, and
Analytical Chemistry, Vol. 79, No. 2, January 15, 2007 761

Figure 3. (a) [C4py][NTf2] before (left) and after (right) passing
through the column. (b) Electronic absorption spectra of [C4py][NTf2]
before (blue) and after (red) passing through the column.
Figure 5. (a) Trihexyl(tetradecyl)phosphonium diisobutylphosphate
before (left) and after (right) passing through the column. (b) Electronic
absorption spectra of trihexyl(tetradecyl)phosphonium diisobutylphos-
phate before (blue) and after (red) passing through the column.
Metathesis Step. Most of the ionic liquids studied were water
immiscible; they are mainly [NTf2]-, [PF6]-, and [BF4]- salts,
which simplifies the metathesis step because the removal of halide
ion can be carried out by washing with water.
The halide salt [Q]X and the respective metal salts MY (M )
Li+, Na+, K+, or [NH4]+; Y ) [NTf2]-, [PF6]-, [BF4]-, or [OTf]-)
dissolved in dichloromethane (∼300 cm3/mol of ionic liquid) were
repeatedly extracted with small amounts of water (20-50 cm3).
After removing the organic solvent, the ionic liquid was dried
ready for use. Drying is usually carried out under high vacuum
at elevated temperatures. However, when working with tetrafluo-
roborates or hexafluorophosphates, temperatures of 50-60 °C
should not be exceeded (to avoid excessive hydrolytic decomposi-
tion).
Decolorization Step. Although the colored impurities might
be aesthetically displeasing, there is no evidence that the chro-
mophoric impurities affect either the chemistry or the physical
properties of ionic liquids. However, they have a major effect upon
measurements involving light absorption or emission; thus, they
interfere with studies involving electronic absorption, lumines-
cence, or Raman spectroscopy.
Figure 4. (a) [C4dmim][NTf2] before (left) and after (right) passing
through the column. (b) Electronic absorption spectra of [C4dmim]-
[NTf2] before (blue) and after (red) passing through the column.
A number of methods in the literature for specific ionic liquids
have been reported. We are attempting to create a methodology
for generic ionic liquid decolorization. Thus, after many trial-and-
error experiments, the column reported here, which combines
learning from previous attempts of decolorization,7-9 seems the
most effective.
The removal of the haloalkanes and reaction solvents is
generally not a problem, especially for the relatively volatile
shorter chain haloalkanes. In the case of longer chain haloalkanes,
their boiling points are higher than shorter ones, and they are
removed in the washing (recrystallization) cycles.
Decolorization by the column was attempted for both halide
salts and ionic liquids. Pictures of several ionic liquids before and
after decolorization by the column were taken (Figures 3a-5a
and Figures Ia-IVa, Supporting Information), and their electronic
(17) Widegren, J. A.; Laesecke, A.; Magee, J. W. Chem. Commun. 2005, 1610-
1612.
(18) Abdul-Sada, A. K.; Ambler, P. W.; Hodgson, P. K. G.; Seddon, K. R.; Stewart,
N. J. World Pat. WO 95 21871, 1995.
762 Analytical Chemistry, Vol. 79, No. 2, January 15, 2007

Figure 7. (a) [C4mim][OTf] before decolorization; (b) after the first
decolorization using the column described above; (c) after the second
attempt of decolorization using the second column. (d) [C4mim][OTf]
before decolorization (blue); after the first attempt of decolorization
(red); after the second attempt of decolorization (green).
In particular, excellent results were obtained for ionic liquids
containing [NTf2]-, [BF4]-, and [PF6]- anions, irrespective of the
cation.
Figure 6. (a) [C4mim][NTf2] before (left) and after (right) decolori-
zation. (b) Several fractions of [C4mim][NTf2] collected.
Decolorization for trihexyl(tetradecyl)phosphonium diiso-
butylphosphate is not effective, due to the fact this ionic liquid
appears to interact with the charcoal and dissolves some colored
material. Another possibility is that the diisobutylphosphate anion
forms colored complex ions with traces of metals.
Some ionic liquids are not decolorized effectively by this
method. These tend to be ionic liquids with anions that undergo
strong hydrogen-bonding interactions, for example, halides or
phosphinates. In addition, ionic liquids based upon the triflate
anion are not so efficiently decolorized with this system (see
Figure III, Supporting Information).
Our results suggest that the best strategy is to decolorize the
final ionic liquid rather than the halide salt. Using this methodol-
ogy, it is possible to decolorize ∼200 cm3 of an ionic liquid in
4-5 h. While we have optimized the column for use with ionic
liquids including the [NTf2]-, [BF4]-, and [PF6]- anions, different
configurations of the column might be better for other anions.
A second equivalent column could be used if further decol-
orization is required. [C4mim][OTf] was chosen as an example,
as it could not be decolorized completely during the first attempt.
It can be seen in Figure 7 that the second attempt of decolorization
using the second parallel column improves the results of decol-
orization compared to the first attempt.
absorption spectra were recorded (Figures 3b-5b and Figures
Ib-IVb, Supporting Information). The spectra of purified ionic
liquids are generally featureless in the region of absorption 270-
600 nm but show an increase to shorter wavelengths, as the tail
of the cation charge-transfer band is approached.
During the first attempt to decolorize the [C4mim][NTf2] ionic
liquid (200 cm3 of the ionic liquid in 600 cm3 of dichloromethane),
several fractions of ∼100 cm3 were collected and UV-vis spectra
recorded on each of them. There was no significant difference
between spectra, so the decolorization process did not lose its
efficiency (Figure 6).
Relationship between the Chromophore and Fluores-
cence. Since the early days of ionic liquids, it has been observed
that there is often a background fluorescence stimulated by UV
irradiation, which can cause problems when attempting to
measure Raman spectra.19 One means of removing these has been
brief exposure to lithium “peas”, which become rapidly coated
with a black deposit, with concomitant elimination of the fluores-
cent impurities. We report here that these fluorescent impurities
are disconnected with the colored impurities, as the fluorescence
of the samples studied here is reduced but not eliminated by the
decolorization procedure.
We recommend this methodology for ionic liquids where
extremely low levels of chromophores are needed. We have
optimized the methodology for the specific ionic liquids described
here, but we believe the approach is versatile enough to be
adapted for other ionic liquids, providing they do not contain
strongly hydrogen-bonding anion, or cations that interact signifi-
cantly with carbon.
CONCLUSION
The results presented here demonstrate that, for most ionic
liquids, it was possible to remove the chromophoric impurities.
(19) Anderson, J. L.; Armstrong, D. W.; Wei, G.-T. Anal. Chem. 2006, 78, 2892-
2902.
Analytical Chemistry, Vol. 79, No. 2, January 15, 2007 763

ACKNOWLEDGMENT
SUPPORTING INFORMATION AVAILABLE
Additional information as noted in text. This material is
available free of charge via the Internet at http://pubs.acs.org.
The authors gratefully acknowledge QUILL and its Industrial
Advisory Board, as well as the EPSRC (Portfolio Partnership
Scheme, grant EP/D029538/1) for funding. We are also extremely
grateful to Dr. Al Robertson (Cytec) for supplying the phospho-
nium ionic liquids used.
Received for review August 9, 2006. Accepted November
4, 2006.
AC061481T
764 Analytical Chemistry, Vol. 79, No. 2, January 15, 2007