Solvents from biorenewable sources: Ionic liquids based on fructose

Article abstract of DOI:10.1021/ol034778b

(Matrix presented) Fructose has been used as the starting material for the preparation of a new class of room-temperature ionic liquids (RTILs). These liquids exhibit tunable solvent properties much like conventional imidazole-based RTILs. They have been

Full text of DOI:10.1021/ol034778b

ORGANIC
LETTERS
2003
Vol. 5, No. 14
2513-2515
Solvents from Biorenewable Sources:
Ionic Liquids Based on Fructose
Scott T. Handy,* Maurice Okello, and Gary Dickenson
Department of Chemistry, State UniVersity of New York at Binghamton,
Binghamton, New York 13902-6000
shandy@binghamton.edu
Received May 7, 2003
ABSTRACT
Fructose has been used as the starting material for the preparation of a new class of room-temperature ionic liquids (RTILs). These liquids
exhibit tunable solvent properties much like conventional imidazole-based RTILs. They have been applied as recyclable solvents for the Heck
reaction of aryl iodides.
In the ongoing quest to improve the environmental compat-
ibility of chemical synthesis, alternatives to conventional
organic solvents have attracted increasing attention.1 Among
the numerous alternatives that have been or are being ex-
plored (including water, supercritical fluids, and fluorous sol-
vents), a particularly promising option is room-temperature
ionic liquids (RTILs). A number of RTILs based on large
organic cations (particularly dialkylimidazolium) in combi-
nation with weakly coordinating anions such as tetrafluoro-
borate, hexafluorophosphate, and bis(trifluoromethanesulfo-
nyl)imide (triflimide) have been reported.2 These compounds
are all reasonably air and moisture stable and are capable of
dissolving a wide range of organic, organometallic, and inor-
ganic compounds. Further, as salts, they exhibit no detectable
vapor pressure under ambient conditions and are nonflam-
mable. Finally, while the toxicity has yet to be determined
for many of these materials, the lack of volatility greatly
reduces any chance of exposure other than by direct physical
contact through spills or intentional ingestion.
Another potential problem with most conventional solvents
and alternative solvents is the fact that they are based on
petroleum. As this resource continues to be consumed at a
prodigious pace, alternative materials based on biorenewable
resources are of considerable interest.3 Within the area of
RTILs, only a few of these new solvent alternatives have
been based on biorenewable natural products.4 In this
communication, we report our initial studies of a family of
RTILs based on fructose.
Using chemistry developed many years ago by Trotter and
Darby, fructose can be readily converted into hydroxymeth-
yleneimidazole 1 (Scheme 1).5 From imidazole 1, double
alkylation of this core generates the corresponding imida-
(3) For examples, see: Descotes, G., Ed.; Carbohydrates as Organic
Raw Materials II; VCH: Weinheim, 1993. Stevens, E. S. Green Plastics:
An Introduction to the New Science of Biodegradable Plastics; Princeton
University Press: Princeton, 2002.
(4) For examples, see: Abbott, A. P.; Capper, G.; Davies, D. L.; Munro,
H. L.; Rasheed, R. K.; Tambyrajah, V. Chem. Commun. 2001, 2010. Abbott,
A. P.; Capper, G.; Davies, D. L.; Rasheed, R. K.; Tambyrajah, V. Green
Chem. 2002, 4, 24. Abbott, A. P.; Capper, G.; Davies, D. L.; Rasheed, R.
K.; Tambyrajah, V. Chem. Commun. 2003, 70. Wasserscheid, P.; Boesmann,
A.; Bolm, C. Chem. Commun. 2002, 200.
(5) Trotter, J.; Darby, W. In Organic Syntheses; Wiley: New York, 1955;
Collect. Vol. III, p 460. We have optimized the workup and isolation
conditions of this original report by replacing the use of hydrogen sulfide
to decompose the copper salts with thioacetimide and isolating the final
product 1 by chromatography instead of precipitation as a perchlorate salt.
Preparations of 1 on a 0.5 mmol scale have been routinely carried out using
these modifications. Details can be found in Supporting Information.
(1) For a recent volume reviewing many of these alternatives, see:
Knochel, P., Ed. Topics in Current Chemistry; Springer: New York, 1999;
Vol. 206.
(2) For recent reviews, see: Welton, T. Chem. ReV. 1999, 99, 2071-
2083. Rogers, R. D., Seddon, K. R., Eds.; Ionic Liquids: Industrial
Applications to Green Chemistry; American Chemical Society: Washington,
DC, 2002. 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.
10.1021/ol034778b CCC: $25.00 © 2003 American Chemical Society
Published on Web 06/14/2003

was partially miscible. The dicyanimide salt, on the other
hand, was completely miscible with water and methanol, as
well as acetonitrile, acetone, and methylene chloride. With
ethyl acetate, a small degree of immiscibility was noted,
while this salt was completely immiscible with ether, toluene,
and hexane.
Scheme 1. Synthesis of Fructose-Derived Ionic Liquids
One important question was if these protic RTILs could
be effective solvents for the same types of reactions that have
been performed in conventional imidazolium RTILs. Initial
investigations have focused on the Heck reaction (Scheme
2).8 The reaction of methyl acrylate with iodobenzene was
Scheme 2. Heck Reaction of Methyl Acrylate and
Iodobenzene
zolium cation. For the first alkylation, the ultrasound-
promoted conditions of Diez-Barra were most effective on
a small scale.6 Increasing the reaction scale above 1 g of 1,
however, resulted in poor reproducibility. For larger-scale
alkylations, conventional solution-phase alkylations in etha-
nol with potassium tert-butoxide as the base afforded
reproducible results.7 As a result, 10-20 g scale reactions
afforded 65-72% yields of the monoalkylation product 2.
It is worth noting that this first alkylation is not entirely
regiospecific but instead affords an inseparable 9:1 mixture
of the two regioisomers. On the basis of the observation of
a NOE enhancement between the butyl group and the
methylene group in the minor isomer, the major isomer is
that of alkylation at the nitrogen more distant from the
hydroxymethylene group. At this point, the second alkylation
with methyl iodide in methylene chloride proceeded un-
eventfully to afford the final iodide salt 3 in nearly quantita-
tive yield.
examined using a simple catalyst (palladium acetate).9 The
reaction was quite rapid, being complete in under 1 h at 100
°C, and methyl cinnamate was isolated as the only product
in >95% yield. The solvent and catalyst could be recycled
several times. Indeed, the catalyst/solvent could be recycled
up to five times by simply extracting the cinnamate product
with cyclohexane. By the fifth reaction, however, the medium
had become relatively viscous, doubtless due to the buildup
of triethylammonium salts. Fortunately, washing with water
and drying in vacuo at 80 °C for 5 h was sufficient to return
the ionic liquid to its original state. This restored solvent/
catalyst could then be reused at least another three times.
Additional substrates were explored under these conditions
(Scheme 3). In terms of aryl iodides, both electron-rich and
Standard anion metathesis was employed to afford a range
of new RTILs. The tosylate, trifluoroacetate, and acetate salts
were all extremely viscous, while the tetrafluoroborate,
triflimide, and dicyanimide salts were less viscous (qualita-
tively similar to the corresponding simple N-butyl, N-
methylimidazolium [bmim] salts).
Another physical property that was investigated was the
solubility/miscibility properties of the different salts. Atten-
tion was focused on the least viscous RTILs, those with
triflimide and dicyanimide anions. These two salts displayed
complementary properties. The triflimide salt was completely
immiscible in water and poorly miscible in methanol. At the
same time, it was completely miscible with acetonitrile,
acetone, methylene chloride, ethyl acetate, and even ether.
Only upon reaching simple hydrocarbons such as hexane and
toluene was immiscibility again noted, and even here, toluene
(8) For reports of the Heck reaction in RTILs, see: Calo, V.; Nacci, A.;
Lopez, L.; Napola, A. Tetrahedron Lett. 2001, 42, 4701. Battistuzzi, G.;
Cacchi, S.; Fabrizi, G. Synlett 2002, 439. Calo, V.; Nacci, A.; Monopoli,
A.; Lopez, L.; Cosmo, A. Tetrahedron 2001, 57, 6071. Boelm, V. P. W.;
Herrmann, W. A. Chem. Eur. J. 2000, 6, 1017. Calo, V.; Nacci, A.; Lopez,
L.; Mannarini, N. Tetrahedron Lett. 2000, 41, 8973. Herrmann, W. A.;
Boehm, V. P. W. J. Organomet. Chem. 1999, 572, 141. Kaufmann, D. E.;
Nouroozian, M.; Henze, H. Synlett 1996, 1091. Xu, L.; Chen, W.; Ross, J.;
Xiao, J. Org. Lett. 2001, 3, 295. Okubo, K.; Shirai, M.; Yokoyama, C.
Tetrahedron Lett. 2002, 43, 7115. See also ref 10.
(9) Representative Reaction Conditions. A solution of 0.045 g (0.20
mmol) of Pd(OAc)2 in 4 mL of 3-butyl-(4)5-hydroxymethyl-1-methyl-3H-
imidazol-1-ium triflimide salt was heated while stirring to 60 °C, whereupon
a dark reddish solution formed. The mixture was cooled to room temper-
ature, and a combination of 2.04 g (10.0 mmol, 1.119 mL) of iodobenzene,
1.1257 mL (12.5 mmol) of methyl acrylate, and 1.51 g (15.0 mmol, 2.010
mL) of triethylamine were added. The reaction mixture was heated to 100
°C under argon for 30 min. The cooled solution was extracted with 10 ×
10 mL portions of cyclohexane. The combined organic extracts were
concentrated in vacuo to afford 1.6086 g (99.2%) of methyl cinnamate (pure
(6) Diez-Barra, E.; de la Hoz, A.; Sanchez-Migallon, A.; Tejeda, J. Synth.
Commun. 1993, 23, 1783.
(7) Interestingly, despite the fact that the use of potassium tert-butoxide
in ethanol should generate sodium ethoxide in situ, attempts to use a solution
of sodium ethoxide in ethanol generated by the addition of sodium metal
to ethanol have not led to satisfactory results.
1
by HNMR in CDCl3).
2514
Org. Lett., Vol. 5, No. 14, 2003

Scheme 3. Other Heck Reactions
electron-deficient compounds worked well. In both cases,
the catalyst/RTIL layer could be recycled three or four times
with no detectable loss in activity.
Figure 1. Reaction rates and pseudo first-order rate constants for
the Heck reaction of methyl acrylate and iodobenzene
With aryl bromides, the case was somewhat different.
Using bromobenzene as the substrate, no reaction was
observed at 100 °C. Even attempts at 140 °C afforded only
traces of the desired product after 12 h. Finally, the same
reaction was attempted using a combination of palladium
acetate and tri-(ortho-tolyl)phosphine. Seddon and co-
workers have noticed that, in conventional imidazole-derived
RTILs, the addition of phosphine ligands improves the rate
of reaction with aryl bromides.10 In the fructose-derived
RTILs, however, no reaction was observed, even after 24 h
at 140 °C.
Given the failure with aryl bromides, there is a significant
question regarding the effect of the hydroxyl group present
in the fructose-derived RTILs on the rate of the Heck
coupling. Initially, we suspected that it might enhance the
reactivity of aryl iodides since these couplings could be
readily performed at temperatures much lower than those
reported for [bmim][BF4] (60 versus 100 °C). Subsequently,
it was observed that even the [bmim] salts were effective at
60 °C. With respect to the rate of the Heck reaction, there
are two main differences between the fructose-derived RTIL
and [bmim][BF4]: the anion and the hydroxyl group. To
dissect the relative effects of these two modifications, a study
of the kinetics of the reaction of methyl acrylate and
iodobenzene in [bmim][BF4], [bmim][NTf2], and RTIL 3 (X
) NTf2) was undertaken. These reactions were conducted
under pseudo first-order conditions using a 10-fold excess
of methyl acrylate with 1-methylnaphthalene as the internal
standard. They were monitored by removing aliquots at the
specified times and analyzing the ratio using GC/MS. The
results are seen in Figure 1.
From these results, it is apparent that in the normal [bmim]
series, the Heck reaction in the triflimide salt is slightly faster
than in the tetrafluoroborate salt, presumably due to poorer
coordination of the triflimide anion with any vacant coor-
dination sites on the palladium metal than is observed with
the tetrafluoroborate anion. In keeping with the hypothesis
that coordination of the solvent with the palladium metal
will slow the reaction rate, RTIL 3 was observed to have an
even slower reaction rate than either of the [bmim] salts,
although the magnitude of this difference is not great (less
than a factor of 2).
On the basis of even the current efforts, it seems likely
that a number of other transition metal-catalyzed reactions
could be performed equally well in the fructose-derived
RTILs instead of the conventional [bmim] series. At the same
time, the protic nature of this new solvent affords the
opportunity to explore other features not accessible using
conventional RTILs. Efforts along these lines are currently
underway and will be reported in due course.
Acknowledgment. The authors thank the State University
of New York at Binghamton, the Research Foundation, and
the New York State Energy Research and Development
Authority (NYSERDA) for financial support of this work.
Supporting Information Available: Experimental pro-
cedures and spectral data for all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(10) Carmichael, A. J.; Earle, M. J.; Holbrey, J. D.; Seddon, K. Org.
Lett. 1999, 1, 997.
OL034778B
Org. Lett., Vol. 5, No. 14, 2003
2515