An intermediate for the clean synthesis of ionic liquids: Isolation and crystal structure of 1,3-dimethylimidazolium hydrogen carbonate monohydrate

Article abstract of DOI:10.1002/chem.200700055

1,3-Dimethylimidazolium-2-carboxylate and carbonic acid have been used to prepare a 1,3-dimethylimidazolium hydrogen carbonate salt by means of a Krapcho reaction. The ability to form hydrogen carbonate azolium salts allows for them to be used as precursors for fast, efficient, environmentally benign, and halide-free syntheses of many ionic liquids by a simple, acid-base reaction of virtually any acid (inorganic, organic, and organic noncarboxylic) with a pK_a less than that of HCO₃ ^-. ditionally, the kinetics of this reaction can be accelerated by employing catalytic amounts of DMSO (a traditional Krapcho solvent used in decarboxylation reactions) to catalyze the decarboxylation. The crystal structure of 1,3-dimethylimidazolium hydrogen carbonate monohydrate is the first example of an imidazolium-based hydrogen carbonate salt. There is a strong 2D hydrogen-bonded network with facially it-stacked imidazolium cations located in the cavities created by this framework.

Full text of DOI:10.1002/chem.200700055

FULL PAPER
DOI: 10.1002/chem.200700055
An Intermediate for the Clean Synthesis of Ionic Liquids:
Isolation and Crystal Structure of 1,3-Dimethylimidazolium
Hydrogen Carbonate Monohydrate
[
a]
Nicholas J. Bridges, C. Corey Hines, Marcin Smiglak, and Robin D. Rogers*
Abstract: 1,3-Dimethylimidazolium-2-
carboxylate and carbonic acid have
been used to prepare a 1,3-dimethyli-
midazolium hydrogen carbonate salt by
means of a Krapcho reaction. The abil-
ity to form hydrogen carbonate azoli-
um salts allows for them to be used as
precursors for fast, efficient, environ-
mentally benign, and halide-free syn-
theses of many ionic liquids by a
simple, acid–base reaction of virtually
any acid (inorganic, organic, and or-
DMSO (a traditional Krapcho solvent
used in decarboxylation reactions) to
catalyze the decarboxylation. The crys-
tal structure of 1,3-dimethylimidazoli-
um hydrogen carbonate monohydrate
is the first example of an imidazolium-
based hydrogen carbonate salt. There
is a strong 2D hydrogen-bonded net-
work with facially p-stacked imidazoli-
um cations located in the cavities creat-
ed by this framework.
ganic noncarboxylic) with a pK less
a
ꢀ
than that of HCO . Additionally, the
3
kinetics of this reaction can be acceler-
ated by employing catalytic amounts of
Keywords: halide-free syntheses ·
imidazolium salts · ionic liquids ·
Krapcho
reaction
·
X-ray diffraction
Introduction
gateway to halide-free syntheses for both hydrophobic and
hydrophilic ILs by decarboxylation of the carboxylate with
strong protic acids (for example, HPF , HBF , picric acid,
Currently the most common synthetic routes for ionic liq-
uids (ILs) involve metathesis reactions from halide inter-
mediates, by an inorganic halide precipitation or liquid/
liquid separation, to the desired products. Though this pro-
cedure is simple, complete removal of the halide is time-
consuming, costly, and leads to depressed yields, attributable
to inherent loss of the IL during the purification process. In
some applications, halide contamination is acceptable, but
for others, even ppm quantities of chloride can poison ex-
pensive catalysts, rendering them useless. Current ap-
proaches circumventing halide contaminants include halide-
free syntheses of imidazolium-based ILs by using dialkylsul-
or dimethylcarbonate (DMC) methylating agents,
and these have provided efficient routes.
6
4
[7]
and H SO ). It was noted that the strength of the protic
2
4
acid affects the decarboxylation of 1,3-diMIM-2-(COO); the
[
1]
lower the pK of the acid, the faster it reacts, forming 1,3-di-
a
+
methylimidazolium ([1,3-diMIM] )-based ILs. In the case of
HNO , the primary product formed in the reaction was the
3
protonated salt of the starting material, 2-carboxy-1,3-dime-
thylimidazolium nitrate ([2-(COOH)-1,3-diMIM]
which is stable at room temperature. The decarboxylation of
[2-(COOH)-1,3-diMIM][NO ] was achieved by either heat-
A
C
H
T
R
E
U
N
G
[NO ])
3
[2]
ACHTREUNG
3
ing the sample or chemically by the addition of DMSO,
which catalyzes decarboxylation through a Krapcho mecha-
nism. From these observations, it was determined that
strong acids will cause decarboxylation, leading to the prod-
uct without the need for heating or addition of catalytic
[
3,4]
fate
[
5,6]
Additionally, our group has recently investigated the uti-
lization of 1,3-dimethylimidazolium-2-carboxylate (1,3-
diMIM-2-(COO)), focusing on the carboxylate functionali-
ty appended to the C2 carbon atom of an azolium core as a
[7]
amounts of a traditional Krapcho reaction solvent.
[6]
In the past few years it has been recognized that to ach-
ieve the goal of ultra-pure ILs, the effort can no longer be
focused on improvement of currently used purification tech-
niques, rather, focus should be on the design of synthetic
routes for the formation of ILs in processes that assure con-
taminant-free products. Thus, highly desirable, clean, con-
taminant-free intermediates for a variety of ILs could come
from a class of anions which can easily be used by further
metathesis reactions, leaving no byproducts or original
[
a] N. J. Bridges, C. C. Hines, M. Smiglak, Prof. R. D. Rogers
Department of Chemistry and Center for Green Manufacturing
The University of Alabama
Box 870336, Tuscaloosa, AL 35467-0336 (USA)
Fax : (+1)205-348-0823
E-mail: RDRogers@bama.ua.edu
Chem. Eur. J. 2007, 13, 5207 – 5212
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5207

[
11,12]
anion, even in trace amounts. One approach is to have a re-
active precursor to the desired anion which could form the
desired product and inert byproduct(s). The carbonate
anion would be an ideal candidate as it cleanly decomposes
to gaseous CO and H O in the presence of stronger acids.
carbonate anion.
Additionally, DMC has also been
shown to yield tetramethylammonium carbonates with ILs
being employed as catalysts, but with very depressed
yields. Previous patent literature has proposed the use-
fulness of a [HCO ] -based IL for the halide-free synthesis
of many other ILs through titration with a Brønsted acid
and has recently been scaled-up according to the most
recent Sigma–Aldrich catalog.
Here, we demonstrate the use of a stable, yet reactive,
azolium-based zwitterion, 1,3-diMIM-2-(COO), and H CO3
[13]
[14]
ꢀ
2
2
3
ꢀ
The ability to synthesize a stable [HCO ] -based IL
3
would thus allow the synthesis of an intermediate for the
[15]
formation of virtually any IL by simple reaction of this pre-
ꢀ
cursor with any protic acid stronger than HCO (pKa1
3
[8]
(
H CO ) 6.352). Due to the limited stability of H CO
2
3
2
3
2
formed in the reaction and its fast decomposition, utilization
of this synthetic protocol would allow for the formation of
halide and metal-free ILs, with the only byproducts being
H O and gaseous CO . Additionally, the evacuation of gas-
in the synthesis of a novel 1,3-dimethylimidazolium hydro-
gen carbonate ([1,3-diMIM]
ceeds by means of a Krapcho reaction in which HCO3
A
H
R
U
G
ꢀ
serves as a weak nucleophile and water as the polar solvent
in the decarboxylation of 2-carboxy-1,3-dimethylimidazoli-
2
2
eous byproducts would greatly contribute to the process as
an additional driving force for the reaction.
um hydrogen carbonate ([2-(COOH)-1,3-diMIM]
A
C
H
T
R
E
U
N
G
[HCO ]),
3
ꢀ
It should also be pointed out that the reaction of [HCO3]
resulting in [1,3-diMIM][HCO ]. We have also found that
ACHTREUNG
3
with protic acids and the decomposition of H CO is not
the synthesis can be completed in one step by protonation/
decarboxylation in the presence of DMSO in deionized
2
3
limited to acids with pK <6.35 due to the associated specia-
a
tion of the carbonate anion. Under ambient conditions at a
pH of 7.25, 10% of all carbonate present in the system still
exists as H CO . Thus, in the presence of excess protons
water, with CO absorbed from the atmosphere or purged
through the system. The kinetics of the reaction can be in-
2
creased by increasing the concentration of H CO3 in the
2
3
2
from an acid source, an increase in the concentration of
H CO occurs, and thermodynamic decomposition of H CO
3
system, which can be achieved by increasing the pressure of
CO2.
2
3
2
[9]
is favored. This decomposition process might not be kineti-
cally fast with weaker protic acids, but heating or reduction
of pressure would increase the kinetics.
Results and Discussion
2ꢀ
Even though the synthesis of a [CO ] -based imidazolium
3
salt might seem to be even more attractive over the forma-
It was expected that spontaneous decarboxylation would not
occur when 1,3-diMIM-2-(COO) and H CO were reacted
even at an elevated temperature and pressure. H CO was
2 3
thought to be too weak of an acid to lead to decarboxyla-
tion, thus, the reaction was expected to proceed from 1,3-
ꢀ
2ꢀ
tion of [HCO ] -based salts, the formation of a CO
salt
3
ꢀ
3
2
3
[8]
[7]
(
pKa2 (HCO ) 10.329) would still not allow for metathesis
3
ꢀ
with weaker protic acids than those of a [HCO ] -based salt
3
2ꢀ
(
as CO3 must first undergo a protonation reaction to form
ꢀ
HCO3 anyway) and thus would not extend the potential
product scope. Additionally, the formation of the more basic
diMIM-2-(COO) to [2-(COOH)-1,3-diMIM][HCO ], which
A
H
R
U
G
3
then could undergo the solvent-supported decarboxylation
by using a traditional Krapcho solvent (polar aprotic sol-
2ꢀ
CO3 anion would only decrease the stability of the imida-
zolium cation, as a result of its mildly acidic C2 proton,
which could react leaving an unstable or reactive imidazyli-
dene. Recently, for example, this route for carbene forma-
tion has been exploited in the synthesis of 2-carboxylate
vent, for example, DMSO), yielding [1,3-diMIM][HCO ],
A
C
H
T
R
E
U
N
G
3
Scheme 1. This hypothesis was based on the assumption that
traditionally, Krapcho decarboxylation reactions require:
[16,17]
1) a strong nucleophile and 2) a polar aprotic solvent.
Surprisingly, over time, [2-(COOH)-1,3-diMIM][HCO ] de-
[10]
azolium zwitterions.
ACHTREUNG
3
Recent patent literature has proposed high temperature,
high pressure reaction conditions to initially form imidazoli-
um-based (but not limited to the azolium-based cation) for-
mate salts, which are oxidized with H O /Pd to the desired
carboxylated through a Krapcho reaction in the presence of
only HCO3 and water (Scheme 1, path B). Spontaneous de-
carboxylation has not been observed in previous Krapcho
reactions with such a weak nucleophile (for example,
ꢀ
2
2
Scheme 1. Synthetic routes to [1,3-diMIM]
A
H
R
U
G
3
] from 1,3-diMIM-2-(COO).
5208
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 5207 – 5212

Synthesis of Ionic Liquids
FULL PAPER
ꢀ
HCO ) and a polar protic solvent (for example, H O). Par-
3
2
allel to this finding, we also confirmed that the decarboxyla-
tion of [2-(COOH)-1,3-diMIM][HCO ] does occur under
traditional Krapcho conditions, with the use of DMSO
AHCTREUNG
3
(
Scheme 1, path C).
NMR spectroscopic characterization of all three imidazo-
lium species (with the carboxylate-terminus on C2, on the
starting material, 1,3-diMIM-2-(COO), the intermediate [2-
(
COOH)-1,3-diMIM]
A
H
R
U
G
uct [1,3-diMIM][HCO ]) is especially difficult due to very
ACHTREUNG
3
small changes in the electron density around the C2 carbon
atom in the imidazolium cores. To correctly analyze and dis-
tinguish between the compounds, an in depth analysis of all
1
13
three with both H and C NMR spectra are needed
Table 1).
(
Figure 1. Reaction scheme, path D, 1,3-diMIM-2-(COO) to [1,3-diMIM]-
1
13
Table 1. Comparison of H and C NMR shifts in [D
6
]DMSO.
A
H
R
U
G
3
[HCO ] as a function of water ratio.
Me C4/
C5
C2
C(O
2
)
C(O
2
)H HCO
3
1
1
H
C
3.96 7.52
36.5 123.5 141.0 157.1
–
–
–
–
–
–
3
in which additional [D ]DMSO and H O (polished to 18.1–
6
2
1
8.2 MWcm, then saturated with atmospheric CO2 before
[
1,3-diMIM-2-(COO)]
use) were added at variable ratios with the total volume
kept constant at 1mL. All samples where capped and al-
lowed to sit for 72 h at room temperature before being
1
1
H
C
3.917.58
36.2 124.5 138.4
–
–
–
ꢁ9.4
3
–
ꢁ154
160.6
1
transferred to NMR tubes for H NMR spectroscopic analy-
sis. The water:1,3-diMIM-2-(COO) ratio was determined by
peak integrations from the NMR spectra (Figure 1).
[
2-
A
C
H
T
R
E
U
N
G
(COOH)-1,3-
diMIM]
A
C
H
T
R
E
U
N
G
[HCO
3
]
The results can be explained as follows: 1) the more water
present, the higher the ratio of H CO to imidazolium salt;
1
1
H
C
3.84 7.59 8.04
36.4 123.2 137.6
–
–
–
–
ꢁ9.4
2
3
3
159.0
2
) DMSO causes decarboxylation of the protonated imida-
zolium salt; 3) in the presence of excess H O, the reaction is
2
driven toward formation of the final product, thus when the
water concentration is low, the starting material, 1,3-diMIM-
[
A
C
H
T
R
E
U
N
G
3
1,3-diMIM][HCO ]
2
-(COO), is favored. Given enough time, complete decar-
We have also found that when [1,3-diMIM]
solved in methanol, an unexpected, slow reaction occurs be-
tween the solvent and the product with the formation of 1,3-
A
H
R
U
G
boxylation of 1,3-diMIM-2-(COO) can occur, as long as
H O, CO , and DMSO are present in the solution. We spec-
ulate that to further increase the kinetics of this reaction,
3
2
2
dimethylimidazolium
methylcarbonate
([1,3-diMIM]-
and scale-up the process, pressurized CO would increase
2
A
C
H
T
R
E
U
N
G
[MeCO ]) and H O (confirmed by NMR spectroscopy). This
the solubility of carbonate species in water, thus increasing
the attainable concentration of H CO in solution.
3
2
reaction is driven forward through the formation of a
weaker acid, methylcarbonate, but nonetheless, in the pres-
ence of a protic acid with a pK less than that of MeCO ,
metathesis can still occur, leaving only methanol, H O, and
CO2.
2
3
We have also noticed that the process described in
Figure 1, path D is reversible if the solid sample is stored
“dry” with trace amounts of DMSO, as the system is in equi-
librium with the starting material, 1,3-diMIM-2-(COO). This
was visually observed for the solid [1,3-diMIM]-
ꢀ
a
3
2
Dissolution of 1,3-diMIM-2-(COO) in [D ]DMSO with
6
variable concentrations of H O, in which H CO is also pres-
A
H
R
U
G
2
2
3
ent, revealed an unknown set of peaks with variable ratios
to the starting material. Further investigation showed that
the ratio of the unknown peaks had a kinetic and water con-
centration dependency in the NMR spectra (Scheme 1,
path D). As the concentration of water increases, more 1,3-
back to the thermodynamically favored 1,3-diMIM-2-(COO)
with the release of one mole of water which further solvated
the sample. We speculate that, though kinetically slow, re-
verse carboxylation of the C2 carbon atom is thermodynam-
ically favored due to formation of a fully conjugated, overall
neutral, zwitterionic molecule (1,3-diMIM-2-(COO)) in
comparison to the charge-separated organic salt ([1,3-
diMIM-2-(COO) reacts to form [1,3-diMIM]
Krapcho decarboxylation reaction (Figure 1).
From stock solution of 1,3-diMIM-2-(COO) in
D ]DMSO, equal amounts were added to fresh glass vials
A
H
R
U
G
a
diMIM][HCO ]). This reverse reaction was monitored with
A
H
R
U
G
3
1
13
[
H and C NMR spectroscopy.
6
Chem. Eur. J. 2007, 13, 5207 – 5212
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5209

R. D. Rogers et al.
Structural Characterization
Due to the small differences in the NMR spectra of the
three imidazolium salts, definitive proof of the synthesis by
NMR spectroscopy is problematic. The crystallographic
characterization of the highly deliquescent, colorless, plate-
like crystals of [1,3-diMIM][HCO ]·H O isolated from the
A
H
R
U
G
3 2
mother liquor provides the first structurally characterized
imidazolium-based hydrogen carbonate salt. The asymmetric
unit is provided in Figure 2 and consists of one cation, H O,
2
ꢀ
ꢀ
and HCO . The HCO anions form hydrogen-bonded
3
3
dimers around crystallographic centers of inversion.
Figure 2. ORTEP illustration (50% probability thermal ellipsoids) of the
[
A
C
H
T
R
E
U
N
G
3 2
1,3-diMIM][HCO ]·H O hydrogen-bonded dimer.
Figure 3. Supramolecular features of [1,3-diMIM]
drogen bonding about each HCO dimer includes four water molecules
and eight cations, b) close contacts around each cation, c) contacts
around each water molecule, and d) the two-dimensional hydrogen-bond-
2 2 3
ing network made up of 1D H O···O3···H O chains linked by HCO
A
H
R
U
G
3
]·H
2
O: a) hy-
3ꢀ
Hydrogen bonding about the anion is shown in Figure 3a,
in which a total of eight significant interactions are ob-
served. The hydrogen bound to O1has a slightly elongated
bond (1.19(4) Š) and is disordered, which is not unusual for
ꢀ
dimers in a brickwork lattice. Pairs of cations are p-stacked within the
framework.
[18–21]
carbonate dimers.
The remaining close contacts about
the anion include: O4-H2O4···O3=1.97(2) Š, 171(2)8; C5-
H5A···O2=2.38(2) Š, 151(1)8; O1-H1O1···O2=1.42(4) Š,
ꢀ
ꢀ
ꢀ
3
171(3)8; O4-H1O4···O3=2.02(2) Š, 166(2)8. Contacts
these H O···HCO and HCO ···HCO interactions are
2 3 3
[21]
around each cation (Figure 3b) include interactions with
common,
but the combination of linear chains cross-
linked by HCO3 dimers in this array has not previously
been observed.
ꢀ
ꢀ
four different HCO3 dimers and one water molecule: C5-
H5A···O2=2.38(2) Š, 151(1)8; C2-H2A···O1=2.40(2) Š,
1
2
53(1)8; C4-H4A···O3=2.41(2) Š, 148(2)8; C7-H7B···O3=
.47(3) Š, 167(2)8; and C6-H6A···O4=2.60(3) Š, 156(2)8.
It is, however, the hydrogen bonding between the anionic
Conclusions
dimer and the water molecules which give rise to the inter-
esting supramolecular features observed for this compound.
Although this 2D anionic network has a unique topology,
The chemistry of the carbonate anion in H O is closely tied
to CO , pH, and pressure. Through the proper manipulation
2
2
the H O–dimer interactions are similar to 1D chains ob-
served in a previous structure, for which sets of HCO3
dimers, oriented in the same direction as the chain propaga-
of these variables, the synthesis and subsequent isolation of
the first azolium-based hydrogen carbonate salt was accom-
plished. Initially, this was achieved through the use of pres-
sure and CO (by increased H CO concentrations in situ)
2
ꢀ
tion, are connected via the terminal oxygen by two H O
2
2
2
3
molecules, resulting in rows of chains separated by large
to protonate 1,3-diMIM-2-(COO), followed by decarboxyla-
tion by means of a Krapcho reaction in an aqueous media.
Next, we found that the kinetically slow decarboxylation re-
[21]
hetero-organic molecules.
In the framework we report
ꢀ
here, linear chains of H O···O3 are connected by HCO
2
3
dimers (Figure 3c) to form 2D sheets. Water interacts with
alternating rows of dimers to make up the framework yield-
ing a brickwork pattern. Oppositely oriented pairs of p-
stacked (3.53(3) Š) cations are locked within the voids of
the brickwork-like hydrogen-bonded lattice (Figure 3d).
action of [2-(COOH)-1,3-diMIM]
aqueous environment, and that it can be accelerated by the
addition of DMSO to yield [1,3-diMIM][HCO ]. Further-
more, we have shown that the same product can be obtained
when the reaction is carried out in an aqueous solution with
A
H
R
U
G
AHCTREUNG
3
The incorporation of H O in these networks results in a
addition of DMSO, naturally absorbed CO from the atmos-
2
2
bending of the sheets at the H O locations. Individually,
phere, and without added pressure.
2
5210
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 5207 – 5212

Synthesis of Ionic Liquids
FULL PAPER
2
H), 161.90 ppm (HCO );
3ꢀ
Carbonate-based imidazolium salts not only promise a
wide range of applications themselves, but also offer clean
intermediates for halide-free syntheses of a large number of
imidazolium-based ILs. The speciation of H CO is pH de-
124.43 (CH), 132.07 (C), 155.84 (C(O)
1
H NMR (500 MHz, [D
(
4
]MeOH, 258C): d=4.09 (s, 6H; CH
s, 2H; CH); C NMR (126 MHz) 37.99 (CH ), 124.97 (CH), 142.60 (C),
); H NMR (500 MHz, [D ]DMSO,
3
), 7.48 ppm
1
3
3
ꢀ
1
1
53.43 (C(O)
2
H), 158.00 ppm (HCO
3
6
2
3
258C): d=3.91(s, 6H; CH ), 7.58 (s, 2H; CH), ꢁ9.40 ppm (b, 1H,
3
ꢀ
13
pendent (at pH 7.25, 10% of the carbonate is present as
H CO ) and in the presence of a proton source from a stron-
HCO
3
); C NMR (126 MHz): d=36.18 (CH
3
), 124.54 (CH), 138.72 (C),
ꢀ
ꢁ
154 (C(O)
2
H), 160.60 ppm (HCO
3
).
2
3
ꢀ
1
,3-Dimethylimidazolium hydrogen carbonate ([1,3-diMIM]
A
C
H
T
R
E
U
N
G
[HCO
), 7.32 ppm (s, 2H;
), 123.00 (CH), 133.12 (C),
161.02 ppm (HCO3 ); H NMR (500 MHz, [D ]MeOH, 258C): d=3.86 (s,
3
]):
ger acid more of the HCO3 will be protonated forming
H CO , thus promoting further decomposition of H CO to
1
H NMR (500 MHz, D
2
O, 25 8C): d=3.80 (s, 6H; CH
3
2
3
2
3
13
CH); C NMR (126 MHz): d=38.57 (CH
3
CO and H O. Thus, clean metathesis reactions providing a
ꢀ
1
2
2
4
1
3
large number of new ILs are possible. Though these reac-
tions would be kinetically slow, under mild heating and re-
6H; CH ), 7.51(s, 2H; CH), 9 0. 1ppm (s, H1 ; CH);
C NMR
3
(126 MHz): d=36.47 (CH
(HCO3 ); H NMR (500 MHz, [D ]DMSO, 258C): d=3.84 (s, 6H; CH ),
3
), 124.18 (CH), 137.19 (C), 159.87 ppm
ꢀ
1
ꢀ
6
3
duced pressure the titration of HCO3 with acid will still
ꢀ
7
.58 (s, 2H; CH), 8.02 ppm (s, 1H; CH), ꢁ9.40 (b, 1H, HCO
3
);
proceed to completion.
13
C NMR (126 MHz): d=36.40 (CH
3
), 123.21 (CH), 137.58 (C),
ꢀ
1
59.00 ppm (HCO
3
).
1
,3-Dimethylimidazolium
methylcarbonate
([1,3-diMIM]
A
C
H
T
R
E
U
N
G
[MeCO
3
]):
1
H NMR (500 MHz, [D
3H; CH ), 7.75 (s, 2H; CH), 9.08 ppm (s, 1H; CH);
4
]MeOH, 258C): d=4.09 (s, 6H; CH
3
), 4.24 (s,
1
3
Experimental Section
C NMR
3
ꢀ
(
126 MHz): d=36.49 (CH
3
), 39.52 (CH
3
CO
3
), 124.85 (CH), 138.44 (C),
ꢀ
1
60.96 ppm (CH
3
CO
3
).
DMC and 1-methylimidazole were purchased from Aldrich (Milwaukee,
WI) and 1-methylimidazole was distilled prior to use. Synthesis of 1,3-
diMIM-2-(COO) followed a previously published method, and purity was
confirmed by means of H and C NMR spectroscopy. Into a 10 mL
thick-walled glass pressure tube, 1,3-diMIM-2-(COO) (0.5 g, 4.8 mmol)
was added to room temperature carbonated water (7 mL, about 3% wt/
A
H
R
U
G
3 2
Crystallographic data for [1,3-diMIM][HCO ]·H O: Crystal size=0.25”
0.20”0.10 mm; monoclinic; space group=P2 /n; a=7.2671(18), b=
9.224(2), c=12.750(3) Š; b=98.350(5)8; V=845.6(4) Š ; 1calcd =
1
1
13
[4]
3
ꢀ
3
1.384 Mgm ; 2qmax =56.628; MoKa radiation (graphite monochromated)
l=0.71073 Š; w scans; T=173(2) K; reflections: measured=6034, inde-
pendent=2011, 6026 reflections included in refinement, GOF=1.094
treated for absorption by using SADABS (m (mm )=0.116, min/max
transmission=0.868791); structure solution by using SHELXS, structure
refinement with SHELXTL; 157 parameters; hydrogen atoms were lo-
cated from a difference Fourier map and refined isotropically; R , wR
[I>2s(I)]=0.0401, 0.1022, R , wR (all data)=0.0470, 0.1060; refined
against jF j ; largest residual peak=0.213 eŠ . CCDC-632450 contains
the supplementary crystallographic data for this paper. These data can be
obtained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
wt H
water, and heated at 1008C for 7 days (Scheme 1, path A). The inter-
mediate [2-(COOH)-1,3-diMIM][HCO ] was dried by purging with dry
2 3
CO ), purchased from Publix Super Market as Syfo brand seltzer
[22]
ꢀ1
A
H
R
U
G
3
[
23]
air until the bulk water had been removed, followed by 24 h of drying in
vacuo at room temperature, yielding an amorphous white solid. The
sample was stored in a closed vial for 3 months at room temperature,
during which time the sample absorbed atmospheric water and large col-
orless plate-like crystals grew by means of decarboxylation. The crystals
1
2
1
2
2
ꢀ3
were isolated and identified as [1,3-diMIM]
liquor contained solvated [2-(COOH)-1,3-diMIM]
A
H
R
U
G
3
]·H
2
O. The mother
3
] (confirmed by
A
H
R
U
NMR spectroscopy) and became a thick liquid due to adsorbed atmos-
pheric humidity (Scheme 1, path B).
In later studies, we found that [2-(COOH)-1,3-diMIM]
A
H
R
U
G
3
[HCO ] can under-
go decarboxylation in much shorter times with the addition of a Krapcho
solvent (for example, DMSO) as previously reported. The rate of decar-
boxylation is a function of the concentration of DMSO, but even with
Acknowledgements
[
6]
This research was supported by the Alabama Space Grant Consortium,
NASA Training Grant NNG05GE80H) and the Air Force Office of Sci-
entific Research (Grant F49620-03-1-0357).
catalytic amounts of DMSO (10 mL in 5 mL of H
2
O with ꢁ0.5 g of [2-
(
(
COOH)-1,3-diMIM][HCO ]), complete decarboxylation occurs in less
A
C
H
T
R
E
U
N
G
3
1
than 72 h (as monitored by H NMR spectroscopy; Scheme 1, path C).
For larger-scale syntheses, path D (Scheme 1) allows for multigram quan-
2
tities without the use of pressure or heat. In 50 mL of deionized H O
[
[
[
1] K. R. Seddon, A. Stark, M.-J. Torres, Pure Appl. Chem. 2000, 72,
(
Nanopure, Barnstead, Dubuque, Iowa, 18.1–18.2 MWcm), 3.9906 g of
2
275–2287.
1
,3-diMIM-2-(COO) was dissolved and DMSO (10 mL) was added. The
2] V. Gallo, P. Mastrorilli, C. F. Nobile, G. Romanazzi, G. P. Suranna, J.
Chem. Soc. Dalton Trans. 2002, 4339–4342.
sample was allowed to sit for 3 weeks in an open beaker to allow absorp-
tion of CO into the solution (forming H CO ) before the water was
2
2
3
3] New ionic liquids based on alkylsulfate and alkyl oligoether sulfate
anions: synthesis and application: P. Wasserscheid, R. van Hal, A.
Bçsmann, J. Esser, A. Jess in Ionic Liquids as Green Solvents: Prog-
ress and Prospects (Eds.: R. D. Rogers, K. R. Seddon), American
Chemical Society Symposium Series 856, American Chemical Soci-
ety, Washington DC, 2003, pp. 57–69.
evaporated. The sample was initially dried by purging air across the
sample until a thick liquid was left. After transferring the liquid to a
smaller vial, it was dried in vacuo at room temperature, yielding NMR
pure [1,3-diMIM]
,3-Dimethylimidazolium-2-carboxylate (1,3-diMIM-2-(COO)): H NMR
500 MHz, D O, 25 8C): d=4.07 (s, 6H; CH ), 7.48 ppm (s, 2H; CH),
C NMR (126 MHz): d=37.04 (CH ), 123.30 (CH), 139.98 (C),
); H NMR (500 MHz, [D ]MeOH, 258C): d=3.96 (s,
), 7.44 ppm (s, 2H; CH); C NMR (126 MHz): d=37.55 (CH ),
); H NMR (500 MHz,
), 7.52 ppm (s, 2H; CH);
), 123.52 (CH), 141.01 (C),
A
C
H
T
R
E
U
N
G
3 2
[HCO ]·nH O approaching quantitative yield.
1
1
(
2
3
[
4] J. D. Holbrey, W. M. Reichert, R. P. Swatloski, G. A. Broker, W. R.
Pitner, K. R. Seddon, R. D. Rogers, Green Chem. 2002, 4, 407–412.
5] Unprecedented synthesis of 1,3-dialkylimidazolium-2-carboxyalte;
applications in the synthesis of halogen-free ionic liquids and reac-
tivity as carbon dioxide transfer agent to activate C H bonds: M.
Aresta, I. Tkatchenko, I. Tommasi in Ionic Liquids as Green Sol-
vents: Progress and Prospects (Eds.: R. D. Rogers, K. R. Seddon),
ACS Symposium Series 856, American Chemical Society, Washing-
ton DC, 2003, pp. 93–99.
1
3
3
1
1
6
1
58.31 ppm (C(O)
H; CH
23.52 (CH), 141.44 (C), 157.05 ppm (C(O)
]DMSO, 258C): d=3.97 (s, 6H; CH
C NMR (126 MHz): d=36.47 (CH
57.06 ppm (C(O) ).
-Carboxy-1,3-dimethylimidazolium hydrogen carbonate ([2-(COOH)-
2
4
[
1
3
3
3
1
2
ꢀ
[
D
6
3
1
3
3
1
2
2
1
1
,3-diMIM]
A
C
H
T
R
E
U
N
G
[HCO
3
]): H NMR (500 MHz, D
2
O, 25 8C): d=4.02 (s, 6H;
),
[6] J. D. Holbrey, W. M. Reichert, I. Tkatchenko, E. Bouajila, O.
1
3
CH
3
), 7.46 ppm (s, 2H; CH); C NMR (126 MHz): d=38.14 (CH
3
Walter, I. Tommasi, R. D. Rogers, Chem. Commun. 2003, 28–29.
Chem. Eur. J. 2007, 13, 5207 – 5212
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5211

R. D. Rogers et al.
[
7] M. Smiglak, J. D. Holbrey, S. T. Griffin, W. M. Reichert, R. P. Swat-
loski, A. R. Katritzky, H. Yang, D. Zhang, K. Kirichenko, R. D.
Rogers, Green Chem. 2007, 9, 90–98.
[16] A. P. Krapcho, Synthesis 1982, 805–822.
[17] A. P. Krapcho, Synthesis 1982, 893–914.
[18] Q. Li, T. C. W. Mak, J. Inclusion Phenom. Mol. Recognit. Chem.
1995, 20, 73–88.
[19] Q. Li, W. H. Yip, T. C. W. Mak, J. Inclusion Phenom. Mol. Recognit.
Chem. 1995, 23, 233–244.
[20] D. Braga, E. DꢁOria, F. Grepioni, F. Mota, J. J. Novoa, C. Rovira,
Chem. Eur. J. 2002, 8, 1173–1180.
[
[
8] D. C. Harris, Quantitative Chemical Analysis, 6th ed., W. H. Free-
man, New York, 2003, p. AP14.
9] Kohlensäure H CO und ihre Lonen: L. Gmelin in Gmelin Hand-
2 3
th
buch der Anorganischen Chemie: PartC3 , 8 ed, Verlag Chemie,
Weinhein, 1973, pp. 117–125.
[
[
10] I. Tommasi, F. Sarrentino, Tetrahedron Lett. 2006, 47, 6453–6456.
11] G. Degen, K. Ebel (BASF AG), WO 2006/027070, 2006 [Chem.
Abstr. 2006, 144, 292756].
12] G. Degen, K. Ebel (BASF AG), WO 2006/027069, 2006 [Chem.
Abstr. 2006, 144, 292757].
[21] F. Benetollo, R. Bertani, P. Ganis, G. Pace, L. Pandolfo, L. Zanotto,
J. Organomet. Chem. 2002, 642, 64–70.
[22] G. M. Sheldrick, SADABS, Program for Semiempirical Absorption
Correction of Area Detector Data, University of Gçttingen, Germa-
ny, 1997.
[
[
[
13] Z. Zheng, T. Wu, X. Zhou, Chem. Commun. 2006, 1864–1865.
14] R. Kalb, W. Wesner, R. Hermann, M. Kotschan, M. Schelch, W.
Staber, WO 2005/021484, 2005 [Chem. Abstr. 2005, 142, 299841].
15] http://www.sigmaaldrich.com/aldrich/brochure/al_chemfile_v6_n9.pdf
[23] G. M. Sheldrick, SHELXTL, Version 6.14, Siemens Analytical X-
Ray Instruments, Inc., 1997.
[
Received: January 12, 2007
(Last accessed on the 8th of January 2007).
Published online: April 16, 2007
5212
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 5207 – 5212