Benign coupling of reactions and separations with reversible ionic liquids

Article abstract of DOI:10.1016/j.tet.2009.11.014

Reversible ionic liquids are a novel class of solvents that combine an effective medium where reactions occur with a 'built-in' separation ability for facile recovery of the products and catalysts, making the solvent available for recycle. We report the utility of these solvents in a number of reactions (Claisen-Schmidt condensation, Heck C-C coupling, and CO₂ capture) and discuss the effectiveness of the separation. We also provide insight into the challenges and limitations of using these unique solvent systems to couple reactions and separations.

Full text of DOI:10.1016/j.tet.2009.11.014

Tetrahedron 66 (2010) 1082–1090
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Benign coupling of reactions and separations with reversible ionic liquids
Ryan Hart a, Pamela Pollet b c, Dominique J. Hahne d, Ejae John b, Veronica Llopis-Mestre b,
,
,
b
,
,b,c,
*
Vittoria Blasucci a, Hillary Huttenhower b, Walter Leitner d, Charles A. Eckert a c, Charles L. Liotta a
a Department of Chemical & Biomolecular Engineering, Georgia Institute of Technology, Atlanta GA, 30332, USA
b Department of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta GA, 30332, USA
c Specialty Separations Center, Georgia Institute of Technology, Atlanta GA, 30332, USA
d Institut fu¨r Technische und Makromolekulare Chemie, RWTH-Aachen, 52074 Aachen, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 24 July 2009
Accepted 24 October 2009
Available online 10 November 2009
Reversible ionic liquids are a novel class of solvents that combine an effective medium where reactions
occur with a ‘built-in’ separation ability for facile recovery of the products and catalysts, making the
solvent available for recycle. We report the utility of these solvents in a number of reactions (Claisen–
Schmidt condensation, Heck C–C coupling, and CO2 capture) and discuss the effectiveness of the sepa-
ration. We also provide insight into the challenges and limitations of using these unique solvent systems
to couple reactions and separations.
Published by Elsevier Ltd.
1. Introduction
problems a chemist or engineer could face. However, the one issue
most researchers do not address when dealing with reactions in ILs
In academic research laboratories and commercial scale pro-
cessing alike, reactions and separations are often times thought of
as two separate unit operations. Commonly, chemists first develop
an optimal set of reaction conditions for a given reaction, and then
hand off the product mixture to an engineer to design the sepa-
ration. The end result is that separations in industrial chemical
production comprise a large fraction of the total production cost.
Additionally, many multi-step synthetic procedures involve
switching between polar and nonpolar solvents to create a homo-
geneous reaction phase, thereby increasing the number of separa-
tion steps required. To eliminate the intermediate separation steps
and final product purification stage, one would ideally want a sol-
vent that switches polarity and solubility properties as desired. To
ensure viability in commercial scale processing and to reduce fur-
ther the environmental impact that the chemical industry has, one
would also want the ideal solvent system to be recyclable. The
desire to combine a solvent that will facilitate a reaction and then
offer a separation, combined with the ability to be reused for sus-
tainable processing, led to the development of reversible ionic
liquids (RevILs) as a specific class of ‘switchable solvents.’1–4
Room temperature ionic liquids (ILs) have received much at-
tention from the scientific community in recent years. There is al-
most a limitless combination of anions and cations that can be used,
and some say that ILs are the answer to all solution chemistry
is the separation step. ILs have proven to be very effective as re-
action media, but isolating products or catalysts from the solvent
remains difficult, except for very volatile products that can be easily
stripped from the IL. This problem becomes even more severe if
purification of the ILs for recovery or recycling is considered. It
quickly becomes evident that when the separation step is consid-
ered in the overall processing scheme, the appeal of using ILs in
organic synthesis is quickly diminished.
In recent years we have examined many techniques to couple
reactions and separations, including: supercritical fluids,5–15 near-
critical fluids,11,16–20 phase-transfer catalysts,21–23 gas-expanded liq-
uids,10,11,24–27 the Organic Aqueous Tunable Solvents process
(OATS),28–30 and piperylene sulfone (a recyclable DMSO sub-
stitute).31–33 The most recent and exciting addition to our arsenal of
novel solvent systems are reversible ionic liquids. Just like conven-
tional ILs, reversible ionic liquids can be designed for a specific ap-
plication, where we can adjust the molecular architecture of the
solvent system to achieve the desired solvent properties. Going be-
yond the synthesis and characterization of RevILs, we have examined
the use of these unique solvent systems for a variety of industrially
relevant applications. In addition to a discussion of the positive as-
pects of RevILs in reactions and separations, the challenges and lim-
itations of these solvent systems will also be addressed.
2. Background on reversible ionic liquids
A large number of chemical processes involving reactions and
separations are performed in solution. The solvent provides
* Corresponding author. Fax: þ404 385 3210.
E-mail address: Charles.liotta@chemistry.gatech.edu (C.L. Liotta).
0040-4020/$ – see front matter Published by Elsevier Ltd.
doi:10.1016/j.tet.2009.11.014

R. Hart et al. / Tetrahedron 66 (2010) 1082–1090
1083
a means of bringing reactants together to facilitate reaction. It also
provides a means of temperature control for both endothermic and
exothermic reactions. The difference in solute solubilities in various
solvents is the basis for crystallization and extraction as unit op-
erations in separation processes. The chemical industry has a con-
tinuous demand for better solvent alternatives for chemical
processesdsolvents in which reactions can occur in high yields at
reasonable rates under relatively mild reaction conditions, where
product isolation is facile, and where recycle of the solvent system
is possible.
Tunable solvents are defined as solvents that change their
physico–chemical properties continuously upon variation of ex-
ternal stimuli such as pressure, temperature or composition. An
illustrative example is provided by the biphasic system containing
room temperature ILs and supercritical carbon dioxide (scCO2),
where the physico–chemical properties such as gas solubility, vis-
cosity, and melting points depend strongly on the amount of dis-
solved CO2 present in the IL phase.34 Such changes can drastically
influence the chemical reactivity in these systems, as demonstrated
by catalytic hydrogenation reactions in IL/scCO2.35 In contrast,
switchable solvents change properties as a step-wise function via
changes in their molecular structure upon application of external
stimuli. In other words, they can be switched ‘on’ and ‘off’ by re-
versible chemical transformations. The Liotta–Eckert–Jessop
groups reported the first example of switchable solvent, which can
be switched back and forth from a molecular liquid to an ionic
liquid upon reaction with CO2.2,3 A schematic representation is
shown for the 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) and alco-
hol system (Fig. 1).
The two-component systems presented here are composed of an
equimolar mixture of N,N,N0,N0-tetramethyl-N00-butylguanidine
(TMBG) or DBU and an alcohol such as methanol. Although not
discussed in this article, it should be noted that Weiss and co-
workers have investigated other two-component systems involving
guanidines or amidines with amines.42–44 Just like the DBU/alcohol
system, the TMBG/alcohol system reacts with CO2 (under atmo-
spheric conditions) to form the corresponding ionic liquid, N,N,N0,N0-
tetramethyl-N00-butyl-guanidinium alkylcarbonate (Fig. 2). The
reaction of alcohols with CO2 in the presence of amidines or guani-
dines to form ionic liquids is an equilibrium reaction governed by
many factors such as structure, temperature, and pressure. Studies to
understand the effect of each parameter are currently ongoing.
Figure 2. Reversible switch from a molecular liquid mixture of TMBG and alcohol (R–
OH) to the ionic liquid [TMBGH]þ[RCO3]ꢀ upon addition of CO2. R¼C1 to C12
.
The major distinction between the one-component and the
two-component systems is that the one-component system does
not require the presence of an equimolar amount of alcohol, sim-
plifying the reaction system and subsequent processing schemes.
Reaction of CO2 with the amine precursors, trialkoxy- or trialkyl-
silylpropylamines, forms the corresponding ionic liquids (Fig. 3). It
should be emphasized that the reaction of amines with CO2 to form
carbamic acids and/or carbamates has long been known and used
for CO2 separation.45–47 In addition, the formation of carbamates
decreases the nucleophilicity of the nitrogen, making the reversible
carbamate formation a successful protecting strategy for the amine
group.13,25,48
Figure 1. Reversible switch from a molecular liquid mixture of DBU and alcohol (R–
OH) to the ionic liquid [DBUH]þ[RCO3]ꢀ upon addition of CO2. R¼C1 to C12
.
Traditional ionic liquids, such as [BMIM][BF4], [BMIM][BTA], or
[BMIM][RSO4] (to name just a few popular ones) have been widely
explored during the past decade for synthetic transformations
with marked success.36,37 To this day however, efficiently
achieving both reaction and separation in traditional ionic liquids
remains a major challenge. The inherent ionic nature of molten
salts and their relatively high viscosity often precludes distillation
of the IL itself or precipitation of the product. Commonly, the
product is extracted from the IL with a nonpolar solvent such as
hexane.38 The extraction strategy is product-dependent, generates
significant amounts of organic waste, and disposal or regeneration
of the contaminated IL phase can be costly. Extraction of the
product with scCO2 has been shown to be a possible separation
technique, and in many cases represents an environmentally be-
nign alternative.8,39–41 In all cases, separations from ILs are typi-
cally restricted to removal of very nonpolar and/or highly volatile
components. With the advent of switchable systems, this situa-
tion is fundamentally changed. For the case of RevILs, the solvent
can take two different formsdmolecular and ionicdand each
form exhibits considerably different properties. The resulting
property change upon switching from one solvent form to another
is so drastic that it opens up unique opportunities to address si-
multaneously reactions and separations. This article discusses two
classes of RevILs, a two-component system and a single compo-
nent system.
Figure 3. One-component system: reversible switch from a molecular liquid (tri-
alkoxy- or trialkyl- silylpropylamine) to its corresponding ionic liquid upon addition of
CO2. R¼OC1, OC2, C2 or C3.
The synthesis and characterization of both classes of reversible
ionic liquids is reported herein, with an emphasis on the stepwise
property change upon switching from the neutral to the ionic state
and back. We discuss the use of RevILs for integration of reaction
and separation in the Claisen–Schmidt condensation reaction, Heck
reaction, and CO2 capture. Additionally, we also discuss some lim-
itations and challenges that have been observed.
3. Two-components RevILs
3.1. Physical properties and synthesis
The two-component reversible ionic liquids are prepared by
bubbling CO2 through equimolar solutions of TMBG or DBU and

1084
R. Hart et al. / Tetrahedron 66 (2010) 1082–1090
various alcohols (from methanol to dodecanol) (Figs. 1 and 2). The
reversible formation of the ionic products was characterized by 1H
and 13C NMR, elemental analysis, FTIR, melting points, differential
scanning calorimetry (DSC), and conductivity measurements.3 Re-
versal of the ionic liquid back to the molecular liquid can be achieved
by bubbling N2 (or other inert gases) through the ionic liquid or in the
case of TMBG based systems, by heating at temperatures above 60 ꢁC.
Conductivity measurements clearly demonstrate the reversibility
and repeatability of the formation of the ionic versus molecular
species for three consecutive cycles (Fig. 4). At the beginning of the
experiment, the equimolar mixture of TMBG/methanol in chloro-
and alcohol give a wide selection of solvent switches. Using shorter
alcohols gives a greater difference between the polarities of the
ionic and neutral forms of the solvent. Furthermore, switchable
solvents based on TMBG have a larger switch in polarity compared
to those based upon DBU; the ionic forms of the two systems are
nearly comparable while the neutral form of TMBG is significantly
less polar than that of DBU.
3.2. Claisen–Schmidt condensation reaction of butanone
and benzaldehyde
form does not conduct electricity (0
mS/cm). As CO2 is bubbled into
The Claisen–Schmidt condensation of butanone and benzalde-
hyde yields three products: the internal enone (3-methyl-4-phenyl-
but-3-en-2-one), the terminal enone (1-phenyl-pent-1-en-3-one)
and water (Fig. 6).20,50 Under basic conditions, the terminal enone
product is the predominant product formed.51
the solution, conductivity increases sharply to 250
m
S/cm. This is
consistent with the formation of the ionic species N,N,N0,N0-tetra-
methyl-N00-butyl-guanidinium methylcarbonate. Upon heating the
solution the conductivity decreases to w0 mS/cm, confirming the
complete and repeatable reversal of the ionic liquid to the molecular
liquid.
Figure 6. Claisen–Schmidt condensation of 2-butanone and benzaldehyde.
The reaction of butanone and benzaldehyde was carried out in
TMBG, which played the dual role of base catalyst and solvent.52,53
The isolation of the enone products was performed by adding oc-
tane and methanol to the reaction mixture, followed by the addi-
tion of CO2, which triggered the formation of the ionic liquid. Under
these conditions an octane phase separated from the newly de-
veloped ionic liquid phase. The enone products were pre-
dominantly in the octane phase and were easily separated by
decantation. After 24 h at room temperature or 3 h at 80 ꢁC, yields
of 48% and 44% in enone products were obtained, respectively
(Table 1). The formation of the terminal and internal enone prod-
ucts was also studied as a function of time at 80 ꢁC. As the reaction
time increased, the yields first increased from 13% at 1 h to 44% at
3 h and then began to decrease after 4 h (Table 1). The decrease was
attributed to competing condensation processes between the
enone products and the benzaldehyde, resulting in lower overall
yields of the desired products. As a consequence, short reaction
times to partial conversions were necessary in order to develop
a process in which isolated yields were maximized and solvent
recycle was possible (Fig. 7).
Figure 4. Conductivity of TMBG/methanol in chloroform solvent as a function of CO2
addition and elimination.
The relative polarity of the molecular liquid precursor and the
corresponding ionic liquid was determined via UV–vis absorption
measurement of the Nile Red dye (a common solvatochromic
probe).3 Equimolar mixtures of DBU/alcohol or TMBG/alcohol be-
come significantly more polar when exposed to CO2, as shown by
the shift of the lmax to longer wavelengths (Fig. 5). For example, the
TMBG/methanol mixture exhibits a lmax of 538.0 nm while the lmax
of the corresponding ionic liquid (N,N,N0,N0-tetramethyl-N00-butyl-
guanidinium methylcarbonate) is 554.0 nm, corresponding to
a shift of 16.0 nm. Such a shift in lmax represents a polarity switch
akin to going from chloroform to acetic acid. The Nile Red experi-
ments suggest that the polarity of both molecular and ionic forms
depend on the length of the alkyl chain on the alcohol. However,
the reported lmax values fall well within the range found for non-
switchable ionic liquids.49 The many possible combinations of base
Table 1
Reaction conditions and yields for the condensation of butanone and benzaldehyde
in the presence of TMBG
Temperature (ꢁC)
Time (h)
Yield (%)
Room Temp
24
1
2
48
13
24
44
80
80
80
3
The first attempt to recycle TMBG upon reversal of the ionic
liquid failed. Water, a product from the condensation reaction,
reacted with TMBG and CO2 to form the N,N,N0,N0-tetramethyl-N00-
butylguanidium carbonate. The formation of the carbonate salt
resulted in a dramatic increase in the viscosity of the ionic liquid.
This, coupled with the severe conditions for the reversal of the
Figure 5. The polarity of the neutral and ionic forms of the DBU/alcohol and TMBG/
alcohol solvents as a function of the alcohol chain length.

R. Hart et al. / Tetrahedron 66 (2010) 1082–1090
1085
the reversible ‘switch’ from ionic to molecular solvent should en-
able the separation of product and by-product sequentially, leading
to a recycling of the solvent system and the catalyst (Fig. 9).
Figure 7. Process that couples reaction and separation for the Claisen–Schmidt con-
densation of butanone and benzaldehyde.
carbonate salt, precluded recycling. However, the reversal of the
ionic liquid and recycle of TMBG were successful when the reaction
mixture was dried over magnesium sulfate (and filtered) prior to
reaction with CO2. By introducing the drying step into the process,
the TMBG was successfully recycled three times. The isolated yields
in enone products were 34, 32 and 34% for each cycle with a con-
sistent product distribution of 95% terminal enone product. Again,
it should be emphasized that partial conversion was necessary in
order to avoid the higher condensation products, thus providing
recycle for the Claisen–Schmidt condensation in RevILs.
Figure 9. General reaction/separation scheme for the Heck reaction in the reversible
ionic liquid solvent system.
As a benchmark reaction, the palladium catalyzed Heck reaction
of bromobenzene and styrene was investigated (Fig. 10). E-stilbene
was produced as the major product. Other isomers (1,1-diphenyl-
ethylene and Z-Stilbene) were obtained in yields of less than 5%.
3.3. Heck reactiondreaction of bromobenzene and styrene
Palladium catalyzed C–C coupling reactions between an aryl
halide and a substituted alkene, also known as Heck reactions, are
useful chemical transformations in many synthetic processes. Heck
reactions inherently cause the formation of HX (X¼halide; Fig. 8),
which needs to be neutralized by a base to recover the catalyst.
Therefore, stoichiometric amounts of salt are formed as a by-
product.
Figure 10. Heck reaction of bromobenzene and styrene in the reversible ionic liquid
[DBUH]þ[HexOCO2]ꢀ
.
Several forms of palladium were compared in the reaction
process. Palladium acetate and palladium chloride in the presence
and absence of triphenylphosphine ligand were compared. In ad-
dition, the preformed palladium chloride–triphenylphosphine
complex was also investigated (Table 2). The following general
procedure was employed. The required amount of palladium cat-
alyst was introduced into the reaction vessel in a toluene solution.
The solvent was then removed under reduced pressure, and the
preformed reversible ionic liquid was added. The starting materials,
bromobenzene (1 equiv) and the styrene (1.2 equiv), were in-
troduced into the reaction vessel and the vessel pressurized with
CO2. The reaction mixture was kept at temperature for three days.
After cooling and depressurization, the reaction products were
extracted with heptane and analyzed via GC with tridecane as in-
ternal standard.
Figure 8. General reaction scheme for Heck-type coupling reactions.
In attempts to facilitate product isolation and recycle of the
catalyst, ionic liquids have been widely explored to immobilize Pd
complex catalysts for C–C bond formation.37,54 In some cases, ben-
eficial effects of the ionic environment on the efficiency and selec-
tivity of the catalyst have been observed. It was also demonstrated
that the organic products can be extracted and catalyst solutions can
be reused. Some extractions have included the use of scCO2 as an
extraction medium.55,56 In most cases, however, the accumulation of
stoichiometric amounts of salts in the IL was not addressed. The
presence of salts makes any continuous processing problematic and
will ultimately limit batch-wise recycling. [For a notable exception
using a triphasic system of H2O/IL/organic, see Perosa, et al.57] The
unique properties of RevILs open new opportunities to make Heck
processes more efficient as well as more environmentally-friendly
by combining an efficient reaction with the ability to selectively
separate the desired products, salt by-product, and catalyst.
Table 2
Effect of catalyst type and concentration on the Heck reaction of bromobenzene and
styrene at 115 ꢁC and 30 bar CO2
Entry
Catalyst
Concentration (mol %)
Ligand
Yield (% E-stilbene)
1
Pd(OAc)2
Pd(OAc)2
PdCl2
0.5
0.5
0.5
0.5
0.5
1
d
0
2
2a
3
TPP
d
26
31
87
83
97
4a
5
7
8
PdCl2
TPP
d
Heck reactions were investigated in the two-component re-
versible ionic liquid mixture of DBU/hexanol. The overall process
was designed to couple the reaction, which was carried out under
ionic conditions, and a two-stage separation: first isolating the
nonpolar product from the ionic liquid, and second by precipitating
the salt by-product from the nonpolar solvent mixture. In principle,
PdCl2(TPP)2
PdCl2(TPP)2
PdCl2(TPP)2
d
2
d
a
- Catalyst precusor and ligand were dissolved in acetonitrile and combined in
the reaction vessel. After stirring for 20 min the solvent was removed under reduced
pressure.

1086
R. Hart et al. / Tetrahedron 66 (2010) 1082–1090
The isolated complex PdCl2(TPP)2 (TPP¼triphenylphosphine,
PPh3) was found to give the best performance, with up to a 97%
yield of E-stilbene using a catalyst loading of 2 mol % (Table 2). The
effect of temperature and CO2 pressure on the yields of product was
also investigated (Table 3). Carrying out the coupling under 30 bar
of CO2 at 115 ꢁC was found to afford the higher yield in E-stilbene.
No additional base was required in these experiments since free
DBU, in equilibrium with the ionic species, was found to act as
a scavenger for HBr.
Figure 11. The one-component reversible ionic liquid precursor molecules.
Table 3
Effect of CO2 pressure and temperature on the Heck reaction of bromobenzene and
styrene
the DSC results indicated that some of the alcohol is lost with the
release of CO2 (especially the lower MW alcohols). However, for the
one-component system the loss of CO2 occurs at a temperature
substantially different (>100 ꢁC) from the evaporation of the mo-
lecular liquid. The one-component system also affords ease of
processing and ionic liquid preparation, as the consideration of
reactant stoichiometry can be ignored (when compared to the two-
component systems).
Entry
Temperature (ꢁC)
Pressure (bar)
Yield (%)
1
100
115
115
115
140
30
30
0
50
30
9
87
22
39
65
2
3a
4
5
a
- 2 mol % PdCl2(TPP)2.
Because the presence of water can result in the cleavage of the
Si–O bond in the trialkoxysilylpropylamines, we developed ionic
liquids based on trialkylsilylpropylamines. Stability tests of the one-
component RevILs revealed that while under wet conditions the
trialkoxy- RevILs did in fact degrade quickly, the trialkyl- RevILs
showed no loss of structural integrity over the period of several
months.58 This is an important consideration for the application of
the one-component RevIL solvents to applications where water is
either present, required or produced.
The Nile Red experiments show for the one-component systems
that the conversion from the molecular solvent form to the ionic
form results in a drastic change in polarity (Fig. 12).58 The magni-
tude in change of polarity for all four RevILs is about 10 nm, which
is less than for some of the two-component solvents where the
polarity change was found to be a function of the base and alcohol
used. Just like the two-component systems, however, the polarity
of both the molecular liquid form and ionic liquid form decrease as
the alkyl chain length on the Si- group is increased.
Extraction of the product(s) from the reaction mixture was
carried out under CO2 atmosphere to ensure that
a
[DBUH]þ[HexOCO2]ꢀ remained in the ionic form. However, GC
analysis indicated that some free DBU and hexanol were also
extracted into the heptane phase. After phase separation, the ionic
phase was heated and purged with argon to revert back to the DBU/
hexanol molecular liquid. A colorless precipitate separated from the
solution; it was filtered off and identified as the HBr salt of DBU by
NMR spectroscopy. The supernatant solution remained yellow,
suggesting the presence of dissolved palladium catalyst.
The application of RevILs to the Heck reaction offers a new and
potentially effective solution for the problem of salt accumulation.
The reaction was carried out in the ionic liquid form of the solvent,
from which the nonpolar product could be isolated by extraction.
Reversing the solvent to its molecular liquid state caused the salt
by-product to precipitate. After separation of the salt, the solvent
(containing the catalyst) was switched back to its ionic form and
was used for another Heck reaction showing significant activityy.
We are currently developing an optimized recycle procedure for the
Heck reaction using RevILs.
4. One-component RevILs
4.1. Synthesis and properties
Just as in the two-component systems, the one-component re-
versible ionic liquids are prepared by bubbling CO2 through a pri-
mary amine molecular liquid precursor to cause the conversion to
the ionic liquid. Here we discuss four one-component RevIL solvent
systems (Fig.11). The reversible formation of the ionic products was
characterized by 1H and 13C NMR, elemental analysis, FTIR, and
DSC.4,58 Reversibility back to the molecular liquid with these sys-
tems is easily achieved by heating or stripping with an inert gas.
From a processing standpoint, the one-component system is su-
perior to the two-component system due to the elimination of the
need for stoichiometric alcohol. With the two-component system,
Figure 12. The polarity of the neutral and ionic forms of the one-component reversible
ionic liquid solvents.
4.2. CO2 Capture application
One of the most daunting tasks facing the chemical industry to-
day is the need to develop an energy-efficient process to capture CO2
from existing point source emitters. For the application of flue gas
treatment from coal-fired power plants, alkanolamines have re-
ceived much interest as a reactant to remove selectively the CO2
from dilute feed streams. Because the product from the reaction of
alkanolamines with CO2 is a solid, an additional solvent (commonly
water) is needed for continuous processing. The regeneration of
reactant is achieved through thermal processing, but much energy is
y
Heck Reaction Recycle: The DBU/hexanol solution containing the Pd-species
was treated with CO2 to switch back to the ionic form, and recharged with addi-
tional preformed ionic liquid to compensate for losses from handling. Fresh sub-
strates were added, the mixture was pressurized with 30 bar CO2 and stirred at
115 ꢁC for three days. After reaction, the mixture was treated as before for se-
quential product/salt separation and catalyst recycle. With 87% yield of E-Stilbene
in the first run and 55% in the second, the recovered catalyst solution displayed
lower activity than fresh catalyst.

R. Hart et al. / Tetrahedron 66 (2010) 1082–1090
1087
wasted through heating of the solvent. The one-component RevILs
overcome this energy waste by combining the reactant with the
solvent to eliminate the need of an added solvent. Here we are using
the reaction-induced switch of the RevILs to give separation.
As previously reported, the RevILs behave as dual-capture sol-
vents for CO2 capture: (1) the highly selective and efficient separa-
tion through the chemical reaction of CO2 with the molecular liquid
to form the ionic liquid and (2) the high capacities offered by the
physical absorption of the CO2 into the ionic liquid (Fig. 13).58 The
capacity fromreactionwas calculated using a stoichiometric ratiofor
CO2:amine of 1:2. Due to the relative bulkiness of the silylated
amines, the capacities for reaction are all fairly low and relatively
similar. But a significant enhancement in capacity is observed when
the physical absorption is combined with the capacity through re-
action. The physical absorption capacities indicate that the bulkier
side groups attached to the molecule result in higher capacity. We
believe that this is due to the decreased packing efficiency of the
molecules leaving more space for the CO2 in the ionic liquid matrix.
Figure 14. Cyanosilylation of cyclohexanone in TMBG.
heptane phase, thus facilitating its separation from the reaction
mixture.
The variation of reaction time (1.5 to 16 h) as well as of the molar
ratio of TMBG to cyclohexanone (0.02/1 to 1/1) had little or no ef-
fect on the reaction yields, which were essentially quantitative.
However, the analysis of the ionic liquid phase (via NMR) indicates
the formation of the ionic by-product N,N,N0,N0-tetramethyl-N00-
butyl-N00-trimethylsilylguanidinium cyanide (Fig. 15).59 The for-
mation of this ionic by-product is detrimental as it is irreversible
and its presence drastically increases the viscosity of the ionic
liquid phase, precluding the quantitative reversal of the ionic liquid
for recycle. The formation of the silyl salt was minimized, but not
eliminated, by adding heptane at the reaction stage as opposed to
the separation stage. The product yields were not affected; they
remained quantitative (96 to 98%). Under these conditions, the
recycling of TMBG upon reversal of the ionic liquid phase was in-
vestigated at two molar ratios of TMBG to cyclohexanone toTMSCN,
0.4/1/1.2 and 1/1/2, respectively (Table 4). In the first cycle, the 1-
trimethylsilyloxy-1-cyclohexanecarbonitrile product was isolated
with 86 and 96% yields (Table 4, entries 1–1 & 2–1, respectively).
The ionic liquid phase was then reversed upon heating at 50 ꢁC for
4 h. The resulting TMBG was reused for a second cycle, yielding
the desired product in 54 and 56% (Table 4, entry 1–2 & 2–2,
respectively). The loss of yields upon recycle has been attributed
to the contamination of the guanidine precursor with the by-
product (N,N,N0,N0-tetramethyl-N00-butyl-N00-trimethylsilylguanidi-
nium cyanide).
Figure 13. CO2 capacities (reported here as mol of CO2 per kg amine) for both capture
through reaction (calculated) and physical absorption (measured at 35 ꢁC and 62 bar),
showing the combined effect.
The one-component RevILs, just as the two-component RevILs
and conventional ILs, have an almost unlimited possibility of
structural variability that would give rise to unique solvents. The
drastic change in properties observed as they are switched from
molecular form to ionic indicates potential to be used as solvents
for synthesis to combine reaction and separation. All one-compo-
nent candidates show promise as CO2 capture agents that use re-
action to give separation, and we are continuing to investigate how
the branching on the side chain can result in more favorable
properties for such applications.
Figure 15. Formation of the by-product N,N,N0,N0-tetramethyl-N00-butyl-N00-tri-
methylsilyl guanidinium cyanide.
Table 4
Cyanosilylation of cyclohexanone coupled with the recycling of the ionic liquid
phase
5. Reversible ionic liquids: challenges and limitations
Entry
Mole Ratio TMBG/Cyclohexanone/TMSCN
Time (h)
Yield (%)
Realistically, reversible ionic liquids are not the solution to all
reaction-separation problems. In this section, their limitations and
challenges are discussed for three reactions and one processing
issue: 1) the cyanosilylation of cyclohexanone, 2) the Michael ad-
dition of 2-cyclohexenone, 3) the Michael addition of aniline to 1,3-
diphenyl-propenone and 4) viscosity of ionic liquid form.
1–1
1–2
2–1
2–2
0.4/1/1.2
0.4/1/1.2
1/1/1.2
17
16
16
16
96
54
86
56
1/1/1.2
5.2. Michael addition of dimethylmalonate
to 2-cyclohexenone
5.1. Cyanosilylation of cyclohexanone
The cyanosilylation of cyclohexanone with trimethylsilyl cyanide
(TMSCN) was carried out in TMBG at room temperature (Fig. 14). The
isolation of the product was performed by adding heptane and
methanol to the reaction mixture. The addition of CO2 triggered the
formation of the ionic liquid, which resulted in a two-phase system,
the ionic liquid and the heptane phases. The product 1-trimethylsi-
lyloxy-1-cyclohexanecarbonitrile partitioned predominantly into the
The base-catalyzed Michael addition of dimethyl malonate to 2-
cyclohexenone was carried out in TMBG at 80 ꢁC. In all experi-
ments, the ratio of TMBG to malonate to 2-cyclohexenone was 1/1/
1. Upon addition of methanol, hexane and CO2, the formation of the
ionic liquid yielded two phases, the ionic liquid and the hexane
phase. The analysis by GC–MS of the hexane phase showed the
complete disappearance of the starting materials. The 1H NMR

1088
R. Hart et al. / Tetrahedron 66 (2010) 1082–1090
analysis showed that the product existed in the ionic form and was
present exclusively in the ionic liquid phase. The reason for this is
simply that in a basic medium like TMBG the a-carbon of the b-
And although the data suggest there is potential for altering the
structure of the RevIL to decrease the viscosity of both molecular
and ionic solvent forms, the magnitude in change is still significant
enough to pose complications in continuous flow processing of the
RevILs and must be considered when designing a process.
diester product is deprotonated resulting exclusively in the ionic
product (Fig. 16). An acidic work-up with 10% aqueous HCl was
performed yielding the neutral form of the product, which now
partitioned predominantly into the hexane phase. The neutral
product was subsequently isolated in 86 to 100% yields. Although
TMBG can be recovered and eventually recycled, the necessity of an
acidic work-up complicates the overall process providing little
advantage for coupling the reaction and separation.
6. Conclusions
While conventional room temperature ionic liquids are often
cited as environmentally-friendly alternative solvents, current re-
search continues to indicate that they will face severe limitations in
applicability due to separation difficulties. In this paper, we present
two reactions (Claisen–Schmidt condensation and Heck coupling)
that were successfully carried out in RevILs with the added capa-
bilities of easy separation of the product and recycling of the sol-
vent system. In addition, we emphasized the limitations and
challenges of RevILs for the cyanosilylation of cyclohexanone and
two examples of Michael additions. The ability to couple reaction
and separation in a common solvent system allows for the de-
velopment of sustainable processing in the chemical industry, that
is, simpler, more efficient, and more environmentally-benign than
solvents commonly used today. By treating a typical reaction/sep-
aration process inter-dependently rather than independently, we
are changing the approach that research laboratories and com-
mercial plants take in the design of processes. With proper care and
technique, we show that reversible ionic liquids are capable of
being recycled, completely eliminating the need to purify or dis-
pose of waste. Additionally, they facilitate catalyst recovery, yield-
ing efficient separations of expensive catalysts for reuse. We
continue to investigate other possible systems where switchable
solvent systems have much promise in our attempt to push the
envelope of sustainable solvent systems.
Figure 16. Base-catalyzed Michael addition of dimethylmalonate and 2-cyclohexenone.
5.3. Michael addition of aniline to 1,3-diphenyl-propenone
The Michael addition of aniline to 1,3-diphenyl-propenone
(chalcone) was carried out by first dissolving the reaction partners
in a methanol:heptane mixture (1:4) followed by the addition of
TMBG (in equimolar ratio relative to methanol) (Fig. 17). The re-
action was run at 80 ꢁC for 18 h. The isolation of the 1,3-diphenyl-3-
phenylamino-propan-1-one product was performed by bubbling
CO2 throughout the mixture (forming the ionic liquid) resulting in
a two-phase ionic liquid and heptane system. The 1,3-diphenyl-3-
phenylamino-propan-1-one partitioned predominantly into the
heptane phase, from which it was isolated in 86% yield. After re-
versal of the ionic liquid at 60 ꢁC for 4 h, the TMBG was recycled. In
the second cycle, the isolated yield decreased to 45%. We conjecture
that the accumulation of organic materials in the ionic liquid phase
changes the product partitioning in recycle attempts when com-
pared to the first cycle.
7. Experimental
7.1. General
GC–MS analyses to quantify the concentrations of reactants and
products were conducted on an HP 6890 GC–MS. Standard cali-
bration curves of all starting materials and products were prepared.
If the reaction was recycled, the desired products were isolated,
characterized by NMR and compared to literature values.
7.2. Claisen–Schmidt condensation of butanone
and benzaldehyde
Figure 17. TMBG-catalyzed addition of aniline to 1,3-diphenyl-propenone.
The butanone (1.7 mL, 0.019 mol) and the benzaldehyde
(2.03 mL, 0.019 mol) were added to TMBG (4.1 g, 0.024 mol). The
reaction was then stirred at room temperature or at 80 ꢁC for the
desired amount of time. After cooling, MgSO4 was added to the re-
action mixture and stirred for w30 min before being filtered.
Methanol (1 mL) and heptane (1 mL) were then added and CO2 was
bubbled through the mixture for w1 h. The heptane phase was an-
alyzed by GC–MS. The products isolated from the hexane phase and
from the ionic liquid phase were analyzed by 1H NMR.1-phenyl-pent-
5.4. Viscosity of reversible ionic liquids
Another important processing factor examined for the RevILs is
viscosity (Fig. 18),58 where the viscosity change in going from
molecular form to ionic form is about three orders of magnitude.
1
1-en-3-one60: H NMR: (CDCl3): 1.16 (3H, t), 2.7 (2H, q), 6.7 (1H, d),
7.37–7.4 (3H, m), 7.5–7.6 (3H, m). 4-phenyl-3-methyl-3-buten-2-one:
1H (DMSO-d6): 1.87 (d, 3H), 2.37 (s, 3H), 7.3–7.45 (m, 4H), 7.59 (s,1H).
7.3. General procedure for the Heck Reaction
Styrene was distilled under reduced pressure, filtered through
neutral alumina oxide and stored under argon at ꢀ20 ꢁC. DBU was
dried over CaH2, distilled under reduced pressure and stored under
argon. Acetonitrile was dried by molecular sieves and kept under
argon.
Figure 18. The viscosity of the neutral and ionic forms of the one-component
reversible ionic liquid solvents.

R. Hart et al. / Tetrahedron 66 (2010) 1082–1090
1089
The required amount of catalyst was introduced as a toluene
solution into a 10 ml windowed autoclave. The solvent was re-
moved under reduced pressure. DBU (2.00 mL, 13.2 mmol) and
hexanol (1.68 mL, 13.2 mmol) were combined in a Schlenk flask
under nitrogen atmosphere. CO2 was sparged trough the solution
until the exothermic reaction cooled down. The produced ionic
liquid was added to the catalyst, followed by bromobenzene
(0.21 mL, 2.0 mmol) and styrene (0.27 mL, 2.4 mmol). The reaction
was pressurized with the required CO2 pressure and stirred for
three days at the respective reaction temperature. After cooling
down and depressurizing, the reaction mixture was transferred to
a Schlenk flask, which was kept under CO2 at atmospheric pressure.
The reaction mixture was extracted with heptane. The heptane
phase was evaporated under reduced pressure and analyzed by HP
5890 Series II GC, using tridecane as an internal standard. The re-
versible ionic liquid was sparged with argon and heated to convert
to the molecular phase, precipitating the salt. The supernatant
solvent was removed, converted to the ionic liquid and fresh ionic
liquid was added. The ionic liquid, containing the catalyst was
charged with fresh substrates in the autoclave and treated as
mentioned before. The precipitate was washed with pentane and
dried under reduced pressure to be analyzed by NMR. DBU$HBr salt:
1H (CDCl3): d¼1.54–1.77 (6H, m), 1.88–2.00 (2H, m), 2.81 (2H, br s),
3.30–3.47 (6H, m), 4.82 (1H, br s). 13C (CDCl3) d¼20.3, 24.5, 27.4,
29.3, 33.5, 39.4, 48.8, 54.3, 165.3.
through the mixture for w1 h. The heptane phase was analyzed by
GC–MS. The ionic liquid phase and the isolated product charac-
terized were analyzed by 1H NMR. 3-(N-phenylamino)-1,3-diphenyl-
1-acetone. 1H (CDCl3): 3.46 (d,1H), 3.49 (d,1H), 5.02 (m,1H), 6.57 (d,
2H), 6.67 (m, 1H), 7.05 (m, 2H), 7.25 (d, 1H), 7.3–7.34 (m, 2H), 7.78–
7.59 (m, 5H), 7.92 (d, 2H).
Acknowledgements
This project was funded by the US Department of Energy grant
NT0005287 and by the National Science Foundation grant
07,22,905. We would like to thank Dr. Leslie Gelbaum, the NMR
Center Manager at Georgia Tech, for his assistance with the analysis
and interpretation of the NMR spectra.
References and notes
1. Heldebrant, D. J.; Jessop, P. G.; Thomas, C. A.; Eckert, C. A.; Liotta, C. L. J. Org.
Chem. 2005, 70, 5335–5338.
2. Jessop, P. G.; Heldebrant, D. J.; Li, X. W.; Eckert, C. A.; Liotta, C. L. Nature 2005,
436, 1102.
3. Phan, L.; Chiu, D.; Heldebrant, D. J.; Huttenhower, H.; John, E.; Li, X. W.; Pollet,
P.; Wang, R. Y.; Eckert, C. A.; Liotta, C. L.; Jessop, P. G. Ind. Eng. Chem. Res. 2008,
47, 539–545.
4. Blasucci, V.; Dilek, C.; Huttenhower, H.; John, E.; Llopis-Mestre, V.; Pollet, P.;
Eckert, C. A.; Liotta, C. L. Chem. Commun. 2009, 116–118.
5. Dillow, A. K.; Brown, J. S.; Liotta, C. L.; Eckert, C. A. J. Phys. Chem. A 1998, 102,
7609–7617.
6. Brown, J. S.; Lesutis, H. P.; Lamb, D. R.; Bush, D.; Chandler, K.; West, B. L.; Liotta,
C. L.; Eckert, C. A.; Schiraldi, D.; Hurley, J. S. Ind. Eng. Chem. Res. 1999, 38, 3622–
3627.
7.4. Cyanosilylation of cyclohexanone
7. Thompson, R. L.; Glaser, R.; Bush, D.; Liotta, C. L.; Eckert, C. A. Ind. Eng. Chem. Res.
1999, 38, 4220–4225.
8. Brown, R. A.; Pollet, P.; McKoon, E.; Eckert, C. A.; Liotta, C. L.; Jessop, P. G. J. Am.
Chem. Soc. 2001, 123, 1254–1255.
9. Nolen, S. A.; Lu, J.; Brown, J. S.; Pollet, P.; Eason, B. C.; Griffith, K. N.; Glaser, R.;
Bush, D.; Lamb, D. R.; Liotta, C. L.; Eckert, C. A.; Thiele, G. F.; Bartels, K. A. Ind.
Eng. Chem. Res. 2002, 41, 316–323.
10. Jessop, P.G.; Stanley, R.R.; Brown, R.A.; Eckert, C.A.; Liotta, C.L.; Ngo, T.T.; Pollet,
P. Neoteric solvents for asymmetric hydrogenation: supercritical fluids, ionic
liquids, and expanded ionic liquids, 2003.
Trimethylsilyl cyanide (0.8 mL, 0.006 mol) was added to TMBG
(0.003–0.006 mol) at room temperature with stirring. Cyclohexa-
none (0.5 mL, 0.005 mol) was then added slowly at room temper-
ature. After reaction completion, methanol (1 equivalent relative to
TMBG) and heptane (0.8 mL) were added. CO2 was bubbled through
the mixture for 1 h to form the ionic liquid. The heptane phase
was analyzed by GC–MS and the ionic liquid phase and the isolated
product by 1H
&
13C NMR. 1-trimethylsilyloxy-1-cyclohexane-
11. Eckert, C. A.; Liotta, C. L.; Bush, D.; Brown, J. S.; Hallett, J. P. J. Phys. Chem. B 2004,
108, 18108–18118.
carbonitrile. 1H (DMSO-d6): 0.2 (s, 9H), 1.48–1.7 (m, 8H), 1.9 (m, 2H).
13C (DMSO-d6): 1.4, 22.4, 23.8, 38.6, 70.5, 121. TMBG/TMS-CN salt: 1H
(DMSO-d6): 0.1 (s, 9H), 0.82 (t, 3H), 0.98 (m, 2H), 1.25 (m, 2H), 2.8 (s,
12H). 13C (DMSO-d6): 1.1, 13.9, 18.8, 19.7, 31.7, 39.1, 123.4, 161.3.
12. Koch, D.; Leitner, W. J. Am. Chem. Soc. 1998, 120, 13398–13404.
13. Furstner, A.; Ackermann, L.; Beck, K.; Hori, H.; Koch, D.; Langemann, K.; Liebl,
M.; Six, C.; Leitner, W. J. Am. Chem. Soc. 2001, 123, 9000–9006.
14. Solinas, M.; Jiang, J. Y.; Stelzer, O.; Leitner, W. Angew. Chem., Int. Ed. 2005, 44,
2291–2295.
15. Maayan, G.; Ganchegui, B.; Leitner, W.; Neumann, R. Chem. Commun. 2006,
2230–2232.
16. Chandler, K.; Deng, F. H.; Dillow, A. K.; Liotta, C. L.; Eckert, C. A. Ind. Eng. Chem.
Res. 1997, 36, 5175–5179.
7.5. Michael addition of dimethyl malonate
to 2-cyclohexenone
17. Chandler, K.; Liotta, C. L.; Eckert, C. A.; Schiraldi, D. Aiche J. 1998, 44, 2080–2087.
18. Lesutis, H. P.; Glaser, R.; Liotta, C. L.; Eckert, C. A. Chem. Commun. 1999, 2063–
2064.
19. Patrick, H. R.; Griffith, K.; Liotta, C. L.; Eckert, C. A.; Glaser, R. Ind. Eng. Chem. Res.
2001, 40, 6063–6067.
20. Nolen, S.; L.L., C.; Eckert, C. A.; Glaser, R. Green. Chem. 2003, 5, 663–669.
21. Chandler, K.; Culp, C. W.; Lamb, D. R.; Liotta, C. L.; Eckert, C. A. Ind. Eng. Chem.
Res. 1998, 37, 3252–3259.
22. Wheeler, C.; Lamb, D. R.; Jayachandran, J. P.; Hallett, J. P.; Liotta, C. L.; Eckert, C.
A. Ind. Eng. Chem. Res. 2002, 41, 1763–1767.
23. Jayachandran, J. P.; Wheeler, C.; Eason, B. C.; Liotta, C. L.; Eckert, C. A. J. Supercrit.
Fluids 2003, 27, 179–186.
24. Chamblee, T. S.; Weikel, R. R.; Nolen, S. A.; Liotta, C. L.; Eckert, C. A. Green Chem.
2004, 6, 382–386.
25. Xie, X. F.; Liotta, C. L.; Eckert, C. A. Ind. Eng. Chem. Res. 2004, 43, 7907–7911.
26. Weikel, R. R.; Hallett, J. P.; Liotta, C. L.; Eckert, C. A. Ind. Eng. Chem. Res. 2007, 46,
5252–5257.
27. Hallett, J. P.; Ford, J. W.; Jones, R. S.; Pollet, P.; Thomas, C. A.; Liotta, C. L.; Eckert,
C. A. Ind. Eng. Chem. Res. 2008, 47, 2585–2589.
Cyclohexenone (0.17 mL, 1.75 mmol) was added to TMBG (0.3 g,
1.75 mmol). Dimethyl malonate (0.2 mL, 1.75 mmol) was then
added at room temperature. The reaction was heated to 80 ꢁC for
16 h. After cooling, methanol was added (0.07 mL, 1.75 mmol) and
CO2 was bubbled through the mixture for approximately 1 h. The
heptane phase was analyzed by GC–MS and the ionic liquid phase
by 1H NMR. When HCl work up was performed, the reaction mix-
ture was treated with aqueous HCl (10%) for w1 h. Ether was then
added and the biphasic mixture was stirred for w1 h. The ether
phase was then analyzed by GC–MS. After removal of the ether, the
isolated product was characterized by 1H NMR. 3[bis(methoxy-
carbonyl)methyl]cyclohexanone. 1H (CDCl3): 1.4–1.55 (m, 1H), 1.65–
1.8 (m, 2H), 1.9–2.0 (m, 1H), 2–2.15 (m, 1H), 2.3–2.55 (m, 5H), 3.3 (d,
1H), 3.75 (s, 6H).
28. Lu, J.; Lazzaroni, J.; Hallett, J. P.; Bommarius, A. S.; Liotta, C. L.; Eckert, C. A. Ind.
Eng. Chem. Res. 2004, 43, 1586–1590.
29. Broering, J. M.; Hill, E. M.; Hallett, J. P.; Liotta, C. L.; Eckert, C. A.; Bommarius, A.
S. Angew. Chem., Int. Ed. 2006, 45, 4670–4673.
7.6. Michael addition of aniline to 1,3-diphenyl-propenone
Aniline (0.05 mL, 0.5 mmol) and 1,3-diphenyl-propenone
(0.11 g, 0.5 mmol) are added to the mixture methanol (0.5 mL) and
heptane (2 mL). TMBG was then added (1.8 g, 0.01 mol) and the
reaction mixture heated at 80 ꢁC for 16 h. After cooling, methanol
(0.5 mL) and heptane (0.5 mL) were added and CO2 was bubbled
30. Hill, E. M.; Broering, J. M.; Hallett, J. P.; Bommarius, A. S.; Liotta, C. L.; Eckert, C.
A. Green Chem. 2007, 9, 888–893.
31. Vinci, D.; Donaldson, M.; Hallett, J. P.; John, E. A.; Pollet, P.; Thomas, C. A.; Grilly,
J. D.; Jessop, P. G.; Liotta, C. L.; Eckert, C. A. Chem. Commun. 2007, 1427–1429.
32. Jiang, N.; Vinci, D.; Liotta, C. L.; Eckert, C. A.; Ragauskas, A. J. Ind. Eng. Chem. Res.
2008, 47, 627–631.

1090
R. Hart et al. / Tetrahedron 66 (2010) 1082–1090
33. Donaldson, M. E.; Mestre, V. L.; Vinci, D.; Liotta, C. L.; Eckert, C. A. Ind. Eng.
Chem. Res. 2009, 48, 2542–2547.
34. Topics in Organometallic Chemistry; Langanke, J., Leitner, W., Eds.; Springer:
Berlin/Heidelberg, 2008; Vol. 23, pp 91–108.
35. Solinas, M.; Pfaltz, A.; Cozzi, P. G.; Leitner, W. J. Am. Chem. Soc. 2004,126, 16142–16147.
36. Ionic Liquids III: Fundamentals, Progress, Challenges, and Opportunities; American
Chemical Society: Washington, DC, 2005.
48. Wittmann, K.; Wisniewski, W.; Mynott, R.; Leitner, W.; Kranemann, C. L.;
Rische, T.; Eilbracht, P.; Kluwer, S.; Ernsting, J. M.; Elsevier, C. L. Chem.dEur. J.
2001, 7, 4584–4589.
49. Carmichael, A. J.; Seddon, K. R. J. Phys. Org. Chem. 2000, 13, 591–595.
50. Sundberg, R. J.; Francis, A. C. Advanced Organic Chemistry, Part B: Reactions and
Synthesis, 3rd ed.; Plenum: New York, NY, 1990.
51. M. Stiles, D. W.; Hudson, G. JACS 1959, 81, 628–632.
37. Ionic Liquids in Synthesis 2nd, completely rev. and enl. ed.; Wiley–VCH: Wein-
heim, 2008.
52. Costa, M.; Chiusoli, G. P.; Taffurelli, D.; Dalmonego, G. J. Chem. Soc. Perkin Trans.
1 1998, 1541–1546.
38. Welton, T. Chem. Rev. 1999, 99, 2071–2083.
39. Blanchard, L. A.; Hancu, D.; Beckman, E. J.; Brennecke, J. F. Nature 1999, 399, 28–29.
40. Leitner, W. Acc. Chem. Res. 2002, 35, 746–756.
53. Toshio Isobe, K. F.; Ishikawa, T. J. Org. Chem. 2000, 65, 7770–7773.
54. Mo, J.; Xiao, J. L. Angew. Chem., Int. Ed. 2006, 45, 4152–4157.
55. Carmichael, A. J.; Earle, M. J.; Holbrey, J. D.; McCormac, P. B.; Seddon, K. R. Org.
Lett. 1999, 1, 997–1000.
41. Cole-Hamilton, D. J. Science 2003, 299, 1702–1706.
42. Yamada, T.; Lukac, P. J.; George, M.; Weiss, R. G. Chem. Mater. 2007, 19, 967–969.
43. Yamada, T.; Lukac, P. J.; Yu, T.; Weiss, R. G. Chem. Mater. 2007, 19, 4761–4768.
44. Yu, T.; Yamada, T.; Gaviola, G. C.; Weiss, R. G. Chem. Mater. 2008, 20, 5337–5344.
45. Dibenedetto, A.; Aresta, M.; Fragale, C.; Narracci, M. Green Chem. 2002, 4, 439–443.
46. Aresta, M.; Ballivet-Tkatchenko, D.; Dell’Amico, D. B.; Bonnet, M. C.; Boschi, D.;
Calderazzo, F.; Faure, R. E.; Labella, L.; Marchetti, F. Chem. Commun. 2000, 1099–
1100.
56. Yoon, B.; Yen, C. H.; Mekki, S.; Wherland, S.; Wai, C. M. Ind. Eng. Chem. Res. 2006,
45, 4433–4435.
57. Perosa, A.; Tundo, P.; Selva, M.; Canton, P. Chem. Commun. 2006, 4480–4482.
58. Blasucci, V.; Hart, R.; Llopis-Mestre, V.; Hahne, D.J.; Burlager, M.; Huttenhower,
H.; B.J.R., T.; Liotta, C.L.; Eckert, C.A. Fuel, in press.
59. Wang, L.; Jiang, X. H. J.; Liu, X.; Feng, X. Tetrahedron Lett. 2006, 47, 1581–1584.
60. Phillip Peac, D. J. C.; Jennefer, A. K.; Inderjit, M.; Ian, H.; Lynne, C.; Tim, W.;
Martin, W. Tetrahedron 2006, 62, 1864–1876.
47. Hirst, L. L.; Pinkel, I. I. Indu. Eng. Chem. 1936, 28, 1313–1315.