The 2-position of imidazolium ionic liquids: Substitution and exchange

Article abstract of DOI:10.1021/jo0480850

(Chemical Equation Presented) The 2-position of imidazolium cations is known to be relatively acidic, leading to the useful Arduengo-type carbenes. At the same time, the acidity of this site can lead to undesired side reactions when using imidazolium-based ionic liquids as solvents. In this note, we describe the surprisingly facile deuterium exchange at this position and also the synthesis and exchange under modestly basic conditions (triethylamine) of a series of 2-methyl-substituted compounds.

Full text of DOI:10.1021/jo0480850

reactions.4 Most typically the deprotonation of the imid-
azolium cation is conducted with an alkoxide base.
Despite all of these positive applications, this facile
deprotonation can also be detrimental (Scheme 1). This
was clearly noted by Aggarwal in his studies of the
Baylis-Hillman reaction in imidazolium RTILs.5 The
basic catalysts employed in these reactions (such as
DABCO) resulted in deprotonation at C2 on the butyl-
methylimidazolium (BMIM) salt. The resulting carbene
was then responsible for a number of side reactions with
the aldehyde component of the reaction mixture, consid-
erably reducing the yield of the desired Baylis-Hillman
product.
The 2-Position of Imidazolium Ionic
Liquids: Substitution and Exchange
Scott T. Handy* and Maurice Okello
Department of Chemistry, State University of New York at
Binghamton, Binghamton, New York 13902
shandy@binghamton.edu
Received October 28, 2004
We have made a similar observation in our efforts
exploring RTILs as homogeneous supports.6 Attempts to
acylate a fructose-derived RTIL under basic conditions
at ambient temperature led to considerable darkening
of the reaction. The isolated, acylated RTIL was a dark
brown liquid that was actually a complex mixture of
compounds, including some that appeared to be C2-
modified RTILs. In the final cleavage of the supported
product from the ionic liquid, it was also noted that the
pale yellow reaction mixture became dark brown upon
treatment with either aqueous sodium hydroxide or
sodium methoxide in methanol. In either case, the
recovered RTIL was again a dark brown liquid that
contained a number of imidazolium byproducts. In both
cases (acylation and cleavage), the darkening could be
avoided by employing less basic conditions.
The 2-position of imidazolium cations is known to be
relatively acidic, leading to the useful Arduengo-type car-
benes. At the same time, the acidity of this site can lead to
undesired side reactions when using imidazolium-based ionic
liquids as solvents. In this note, we describe the surprisingly
facile deuterium exchange at this position and also the
synthesis and exchange under modestly basic conditions
(triethylamine) of a series of 2-methyl-substituted com-
pounds.
These observations raised two questions: (1) Could a
2-substituted fructose-derived ionic liquid be readily
prepared that would avoid this deprotonation problem
and (2) how acidic is this 2-position? Both of these issues
have been addressed and are the central topic of this
note.
2-Substituted imidazolium RTILs are certainly known.
Indeed, a few of the 2-methyl compounds are com-
mercially available.7 Most typically, such compounds are
used to demonstrate that Arduengo carbenes are likely
intermediates in various transition metal catalyzed reac-
tions in RTILs. At the same time, there is a report that
employs a 2-methylimidazolium RTIL as a solvent for
some organozinc reagents.8 More recently, Chu and co-
workers have also investigated a 2-substituted imid-
azolium ionic liquid as a solvent for the Baylis-Hillman
reaction and found that it avoids the problems encoun-
tered by Aggarwal with the simple BMIM salts.9 As a
result, it certainly appears that these compounds are
Interest in room temperature ionic liquids (RTILs)
continues to grow.1 These materials, once an object of
scientific curiosity, are making their mark as solvents
for a variety of applications, including organic and
inorganic synthesis, battery applications, chromato-
graphic stationary phases, and separation science. One
of the attributes of this class of materials that has been
widely publicized is the potential to generate a wide
range of types of RTILs by the combination of various
organic cations with numerous organic and inorganic
anions. At the same time, the vast majority of the
research in this area continues to focus on materials
derived from imidazole. In part this is due to the
simplicity in preparing these compounds and the gener-
ally lower melting points and viscosities of the RTILs
derived from imidazole as compared to other classes (such
as pyridinium and tetraalkylammonium RTILs). Under
most circumstances, the imidazole-derived RTILs are
considered to be “inert” solvents, but this is not always
the case.2 It is well-established that the 2-position can
be deprotonated to form a stabilized carbene.3 Indeed,
this Arduengo-type carbene has been implicated in a
number of the advantageous applications of RTILs,
particularly in the area of transition metal catalyzed
(4) In addition to the many applications of these carbene species
found in the reviews in ref 1, specific study and characterization of
these complexes can be found in the following two articles. Xu, L.;
Chen, W.; Xiao, J. Organometallics 2000, 19, 1123-1127. Mathews,
C. J.; Smith, P. J.; Welton, T.; White, A. J. P.; Williams, D. J.
Organometallics 2001, 20, 3848-3850.
(5) Aggarwal, V. K.; Emme, I.; Mereu, A. Chem. Commun. 2002,
1612-1613.
(6) Handy, S. T.; Okello, M. Tetrahedron Lett. 2003, 44, 8399-8402.
(7) Companies such as Aldrich, Strem, and EMD all sell 2-methyl-
imidazolium ionic liquids.
(1) Welton, T. Chem. Rev. 1999, 99, 2071-2083. Freemantle, M.
C&E News 2000, 37-50. Wasserscheid, P.; Keim, W. Angew. Chem.,
Int. Ed. 2000, 39, 3772-3789. Dupont, J.; de Souza, R. F.; Suarez, P.
A. Z. Chem. Rev. 2002, 102, 3667-3692.
(2) Dupont, J.; Spencer, J. Angew. Chem., Int. Ed. 2004, 43, 5296-
5297.
(3) Arduengo, A. J., III; Harlow, R. L.; Kline, M. J. Am. Chem. Soc.
1991, 113, 361-363
(8) Sirieix, J.; Ossberger, M.; Betzemeier, B.; Knochel, P. Synlett
2000, 1613-1615.
(9) Hsu, J.-C.; Yen, Y.-H.; Chu, Y.-H. Tetrahedron Lett. 2004, 45,
4673-4676.
10.1021/jo0480850 CCC: $30.25 © 2005 American Chemical Society
Published on Web 01/21/2005
J. Org. Chem. 2005, 70, 1915-1918
1915

SCHEME 1. Aggarwal’s Baylis-Hillman Reaction
(Scheme 3). In previous work on the 2-unsubstituted
series, the use of potassium tert-butoxide in ethanol had
proved to be the optimal method on large scale.6 Applica-
tion of these same conditions to 2 resulted in the
formation of very modest amounts of 4 (35%) along with
significant amounts of the product of N and O alkylation
(38%). After considerable experimentation, it was dis-
covered that the use of potassium hydroxide in DMF at
55 °C afforded imidazole 4 in reasonable (60%) yield with
only a small amount of double alkylation (<5%). With 4
in hand, the remaining imidazole nitrogen was then
alkylated with methyl iodide to afford the iodide salt 3.
This salt turned out to be a highly viscous, orangish
liquid at room temperature. Following anion metathesis
with either silver dicyanimide or lithium triflimide, the
triflimide and dicyanimide salts were obtained as clear
liquids which were qualitatively slightly more viscous
than the corresponding 2-unsubstituted fructose-derived
compounds.
in a RTIL
SCHEME 2. C2-Substituted Fructose-Derived
Imidazole Preparation
With a synthetic route to the 2-methyl ionic liquids in
hand, the way was paved to address the question of how
acidic this position is.11 To that end, a series of deuterium-
exchange experiments were initiated (Table 1). To es-
tablish a point of comparison, the 2-unsubstituted RTILs
were examined first. Although it was expected that the
2-position in these compounds would be relatively acidic,
it was surprising to note that the dicyanimide salt of
2-unsubstituted fructose-derived RTIL underwent ex-
change quite rapidly in the absence of any added base
(entry 1). Indeed, a similar result was obtained for the
corresponding simple BMIM dicyanimide salt (entry 2).
One possible explanation for these observations is the fact
that aqueous solutions of sodium dicyanimide are re-
ported to be slightly basic (pH of roughly 8).12 As a result,
the anion may be serving as the base in these salts. To
test this hypothesis, a similar exchange experiment was
conducted on BMIM tetrafluoroborate (entry 3). In this
case, no exchange was observed for up to 72 h without
the addition of some base.
The 2-methyl RTILs were expected to be much less
susceptible to exchange. As a result, it was not surprising
that these salts failed to undergo any detectable exchange
in the absence of added base. What was more surprising
was the observation that even a very mild base such as
triethylamine was sufficient to induce a slow but mea-
surable exchange at the 2-methyl group. Thus, both the
fructose-derived compound 6 and the 3-butyl-1,2-dimeth-
ylimidazolium salts13 (with either the dicyanimide or
chloride anion) would eventually undergo complete ex-
change of the three protons at this position (entries 5, 7,
and 9). The presence of a stronger base such as potassium
hydroxide greatly increased this rate, but still resulted
in little difference in terms of rate of exchange on the
three compounds (entries 4, 6, and 8).
considerably less acidic and more base tolerant than the
simple 2-unsubstituted compounds.
For the preparation of a 2-methylfructose-derived ionic
liquid, the most promising route was based on the
historical route to 2-methylimidazole, which involves the
condensation of glyoxal with the acetaldehyde ammonia
trimer 1.10 Armed with this information, the degradation/
condensation reaction of fructose with trimer 1 in place
of formaldehyde and ammonium hydroxide was at-
tempted (Scheme 2). Much to our delight, this did afford
the desired C2 methyl product 2, although in modest
yield (30%). After reexamining the literature surrounding
the use of trimer 1, one distinct possible reason for this
low yield was that trimer 1 was dissociating under the
reaction conditions and evolving ammonia.10 To compen-
sate for this loss of ammonia, the same reaction was
attempted again, but this time with the addition of
ammonium hydroxide. Indeed, the yield of 2 did improve
to 77%, or nearly the same as we reported earlier for the
2-unsubstituted compound.6
On the basis of this success, one further variation was
attempted, this time with ammonium hydroxide and
acetaldehyde instead of trimer 1. Although this did also
afford the desired imidazole 2 in 34% yield, the reaction
afforded numerous products and the isolation of 2
required extensive chromatography. It is likely that
further optimization of these reaction conditions can
improve upon this result, including pregeneration of an
equilibrating mixture of acetaldehyde/ammonia adducts
including trimer 1.
Further evidence that the 2-unsubstituted and 2-meth-
yl-substituted positions are reasonably acidic came from
a series of attempted alkylation reactions on RTIL 7
(11) There has been a single report measuring the pKa of the
2-position on an imidazolium salt. Kim, Y.-J.; Streitwieser, A. J. Am.
Chem. Soc. 2002, 124, 5757-5761.
With imidazole 2 in hand, the next step leading to
imidazolium salt 3 involved alkylation of the nitrogens
(12) Sodium dicyanimide, Merck Safety Data Sheet, catalog no.
490015.
(13) The 2-methyl BMIM salts were prepared in the same fashion
as the simple BMIM salts, except starting with 1,2-dimethylimidazole.
(10) Radziszewski, B. Chem. Ber. 1886, 15, 2706-2708.
1916 J. Org. Chem., Vol. 70, No. 5, 2005

SCHEME 3. Imidazolium Cation Preparation
TABLE 1. Deuterium Exchange Results for
Imidazolium RTILs
material, indicating that the alkoxide is being formed
first and then subsequent deprotonations occur at the
2-position (entry 2). Thus, the acidity of the 2-position is
less than that of the hydroxyl group.
The result of these studies indicates that the 2- position
of imidazolium ILs can be rendered less acidic by replac-
ing the hydrogen with a methyl group. Even this methyl
group is not completely inert, however, and can be subject
to deprotonation under surprisingly mild conditions. Still,
this added stability is sufficient to avoid some of the
competing reactions that are so common in the simple
imidazolium series.9 Efforts are underway to employ
these 2-methyl-substituted ILs derived from fructose as
more base-stable supports and to explore the preparation
of even more base-stable derivatives. These efforts will
be reported in due course.
a
entry
R1
R2
R3
X
base rate (min-1
)
1
2
3
4
5
6
7
8
9
H
HOCH2
D
D
D
N(CN)2 none
N(CN)2 none
BF4
3.0 × 10-3
H
H
H
H
4.1 × 10-2
none
no exchange
CH3 HOCH2 CD3 N(CN)2 KOH 2.3 × 10-3
CH3 HOCH2 CD3 N(CN)2 Et3N
0.04 × 10-3
CH3
CH3
CH3
CH3
H
H
H
H
CD3 N(CN)2 KOH 1.0 × 10-3
CD3 N(CN)2 Et3N
0.04 × 10-3
CD3 Cl
CD3 Cl
KOH 2.1 × 10-3
Et3N
0.04 × 10-3
a Determined by 1H NMR integration. Average of 3 exchange
Experimental Section
reactions.
2-Methyl-1H-imidazol-4(5)-ylmethanol [2]. To 11.1 g (0.0500
mol) of copper carbonate basic was added 75 mL of water,
followed by 11.4 mL (0.600 mol) of 28-30% aqueous ammonium
hydroxide. During this addition, the green suspension became
a homogeneous deep blue solution. To this solution was added
11.9 g (0.0600 mol) of acetaldehyde/ammonia trimer 1 and
4.5 g (0.024 mol) of D-fructose. The resulting mixture was swirled
and then placed in a boiling water bath. After 30 min of heating
and occasional swirling, a moderate current of air was bubbled
through the solution with continued heating for 2 h. The reaction
mixture was then chilled in an ice bath for 3 h and the olive-
brown precipitate was filtered off. The filter cake was washed
with 50 mL of chilled water or until no more blue color was
observed. The residue was then suspended in 200 mL of water
and 4.5 g (0.060 mol) of thioacetamide was added. The resultant
brown-black suspension was stirred at 50 °C for 2 h. The
precipitated CuS was removed via filtration and the reddish/
brown filtrate was concentrated in vacuo. The resulting residue
was purified on silica gel with use of aqueous ammonium
hydroxide/methanol/ethyl acetate (5:35:60) to afford 2.06 g (77%)
of 2 as an off-white solid (mp 112-113 °C). IR (Nujol) 3213, 2972,
TABLE 2. Alkylation Reactions on Imidazolium Salt 7
equiv
of NaH
equiv
of MeI
time
(h)
ratio of
entry
productsa (7:8:9)
1
2
3
0
0.5
6
3
3
10
12
1
12
all 7
1.85:1.0:0.0
all 9
a Determined by 1H NMR.
2928, 1664, 1540, 1419, 1118, 1037, 993, 757 cm-1 1H NMR
;
(360 MHz, D2O) δ 6.89 (s, 1H), 4.44 (s, 2H), 2.27 (s 3H); 13C NMR
(90 MHz, D2O with a dioxane reference) δ 146.4, 137.0, 117.49,
56.7, 13.2. Mass spec (EI) calcd for C5H8N2O 112.0634, found
112.0633.
(Table 2). Treatment of 7 with an excess of sodium
hydride and methyl iodide resulted in the formation of
2-ethyl methyl ether 9 as the sole product (entry 3). A
similar result was obtained starting with RTIL 4. This
differs slightly from the observation of Begtrup on 1,3-
dimethylimidazolium iodide, in which treatment under
the same conditions resulted in the formation of 2-iso-
propyl-1,3-dimethylimidazolium iodide.14 The effective
acidity of the 2-position is also apparent from alkylation
studies on RTIL 7. Treatment with a limiting amount of
sodium hydride and excess methyl iodide results in the
formation of only methyl ether 8 and recovered starting
1-Butyl-2-methyl-1H-imidazol-4(5)-ylmethanol [3]. To a
stirred solution of 0.1072 g (0.9559 mmol) of 2 in 2.4 mL of
anhydrous DMF was added 0.2676 g (4.7793 mmol) of finely
powdered potassium hydroxide, followed by 114 µL (1.0515
mmol) of butyl bromide. The mixture was heated to 55 °C for
7 h. Once TLC indicated complete consumption of starting
material, the reaction was quenched by the addition of powdered
sodium bicarbonate. The mixture was filtered through Celite,
the filter cake rinsed with ethanol, and the filtrate concentrated
in vacuo. The crude product was purified on silica gel with use
of ammonium hydroxide/methanol/ethyl acetate (1:10:89) to
afford 0.096 g (60%) of 3 as a pale yellow oil. IR (neat) 3305,
2959, 1659, 1632, 1546, 1518, 1460, 1425, 1380, 1335, 1232,
(14) Begtrub, M. Bull. Soc. Chim. Belg. 1988, 97, 573-577.
1165, 1116, 1040, 1010, 824, 736 cm-1
;
1H NMR (360 MHz,
J. Org. Chem, Vol. 70, No. 5, 2005 1917

CDCl3) δ 7.1 (s, 1H), 4.7 (s, 2H), 4.0 (t, 2H) 2.7 (s 3H), 1.8
(m, 2H), 1.4 (m, 2H), 1.0 (t, 3H); 13C NMR (90 MHz, CDCl3)
δ 142.8, 132.9, 118.3, 53.7, 47.5, 31.6, 19.5, 13.5, 11.2. Mass spec
(EI) calcd for C9H16N2O 168.1258, found 168.1259.
(s, 2H), 3.99 (t, 2H), 3.72 (s 3H), 2.56 (s, 3H), 1.70 (m, 2H), 1.30
(m, 2H), 0.88 (t, 3H); 13C NMR (90 MHz, acetone-d6) δ 144.4,
133.4, 119.7 (q, JC,F ) 315 Hz), 119.2, 118.0, 144.4, 53.2, 38.2,
32.2, 31.4, 19.4, 13.3, 9.6. Anal.: C 39.12 (39.09), H 5.20 (5.28),
N 11.4 (11.44).
3-Butyl-4(5)-hydroxymethyl-1,2-dimethyl-3H-imidazo-
lium Iodide [4]. To a solution of 0.9284 g (5.4844 mmol) of 3 in
3.7 mL of methylene chloride under argon was added 0.3759
mL (6.0345 mmol) of iodomethane. This solution was heated to
reflux and stirred overnight. After cooling, the solvent was
removed in vacuo. The residual material was extracted with
ether (5 × 5 mL) and the ether layer decanted. The remaining
ionic liquid layer was dissolved in 5 mL of methylene chloride
and filtered through Whatman #1 filter paper. Removal of the
solvent in vacuo afforded 1.6051 g (94%) of iodide 4 as a pale
orange liquid. IR (neat) 3318, 3088, 2958, 2932, 2872, 1664, 1621,
3-Butyl-4(5)-hydroxymethyl-1,2-dimethyl-3H-imidazo-
lium Dicyanimide [6]. To a solution of 0.2010 g (0.6477 mmol)
of 3-butyl-4(5)-hydroxymethyl-1,2-dimethyl-3H-imidazol-1-ium
iodide 4 in 1.5 mL of methanol was added 0.1127 g (0.6477 mmol)
of silver dicyanimide. The resultant mixture was left to stir
overnight at room temperature. The reaction was worked up by
filtration of the reaction mixture through a Hirsch funnel under
vacuum. The filtrate was concentrated in vacuo, to afford 0.1700
g of the crude salt. Further purification was achieved by
dissolving the crude in about 4 mL of dichloromethane, drying
with MgSO4, gravity filtration, and concentration in vacuo to
afford 0.1424 g (88%) of 6 as a pale yellow liquid. IR (neat) νmax
3522, 3366.33, 3135, 2965, 2939, 2878, 2245, 2201, 2142.26,
1624, 1539, 1465, 1420, 1351, 1193, 1136, 1057, 789, 762, 740,
1536, 1463, 1418, 1382, 1156, 1128, 1064, 1028, 808, 732 cm-1
;
1H NMR (360 MHz, acetone-d6) δ 7.76 (s, 1H), 4.66 (s, 2H), 4.32
(t, 2H), 3.92 (s 3H), 2.85 (s, 3H), 1.82 (m, 2H), 1.41 (m, 2H), 0.92
(t, 3H); 13C NMR (90 MHz, acetone-d6) δ 145.6, 134.3,120.2, 52.9,
48.7, 33.9, 32.5, 20.0, 13.9, 11.4. Mass spec (EI) calcd for
C10H19N2O 183.1492, found 183.1490.
654, 617 cm-1 1H NMR (360 MHz, acetone-d6) δ 7.54 (s, 1H),
.
4.72 (s, 1H), 4.25 (t, 2H), 3.89 (s, 3H), 2.78 (s, 3H), 1.84 (m, 2H),
1.40 (m, 2H), 0.94 (t, 3H). 13C NMR (90 MHz, acetone-d6) δ 146.1,
134.7, 119.9, 53.8, 48.7, 32.8, 32.5, 20.1, 13.8, 10.0. Anal.: C
57.81 (57.77), H 7.68 (7.70), N 28.09 (28.01).
3-Butyl-4(5)-hydroxymethyl-1,2-dimethyl-3H-imidazo-
lium Triflimide [5]. Lithium triflimide (0.19 g, 0.66 mmol) of
was dissolved in 0.33 mL of water. In a separate flask, 0.2054 g
(0.6619 mmol) of iodide 4 was dissolved in 0.66 mL of water.
Both solutions were heated to 70 °C and then combined. The
combined solution was allowed to cool to room temperature while
stirring was continued for 3 h. The top aqueous layer was
decanted and extracted with methylene chloride (2 × 5 mL).
These washes were combined with the original remaining ionic
liquid layer and washed with water (2 × 5 mL). The organic
layer was filtered through a short plug of basic alumina and
the solvent removed in vacuo. The residue was dried on a
vacuum line overnight to afford 0.2193 g (72%) of 5 as a pale
yellow liquid. IR (neat) 3523, 3137, 2967, 2939, 2879, 1624, 1590,
1539, 1467, 1420, 1351, 1191, 1136, 1056, 789, 762, 740, 654,
Acknowledgment. The authors thank the New
York State Energy Resource Development Agency (NY-
SERDA) and the Research Foundation for support of
this project.
Supporting Information Available: Methods for deute-
rium exchange studies and spectra for compounds 2, 3, 4, 5,
and 6. This material is available free of charge via the Internet
at http://pubs.acs.org.
616 cm-1
;
1H NMR (360 MHz, acetone-d6) δ 7.19 (s, 1H), 4.58
JO0480850
1918 J. Org. Chem., Vol. 70, No. 5, 2005