Applied Catalysis A: General 222 (2001) 101–117

New developments in catalysis using ionic liquids

Charles M. Gordon∗
Department of Pure and Applied Chemistry, University of Strathclyde,
295 Cathedral Street, Glasgow G1 1XL, Scotland, UK

Abstract
Ionic liquids are low melting point salts that represent an exciting new class of reaction solvents for catalysis. Being
composed entirely of ions, they possess negligible vapour pressures, and the wide range of possible cations and anions means
that other solvent properties may be easily controlled. There is currently great interest in the use of these materials as solvents
for a wide range of applications, including catalysis. Many reactions show advantages when carried out in ionic liquids,
either with regard to enhanced reaction rates, improved selectivity, or easier reuse of catalysts. This review is intended to
bring the reader up to date on the developments involving ionic liquids in catalytic applications over the previous 18 months.
Recent spectroscopic investigations into the solvent properties of ionic liquids with relevance to catalysis are discussed first,
followed by a critical review of the major developments in transition metal, Lewis acid, and enzyme catalysed processes in
these solvents. Particular emphasis is given to the combination of ionic liquids with supercritical fluids, Pd-based catalysts,
and enzymes. Wherever possible, the results gained in ionic liquids are critically compared with those obtained using other
catalytic systems. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Ionic liquids; Biphasic catalysis; Pd catalysis; Supercritical fluids; Lewis acid catalysis; Biocatalysis

1. Introduction for catalysis, both biphasic and homogeneous. Ionic

liquids are simply liquids that are composed entirely
Developments in the field of catalysis are being of ions, but they may be distinguished from classical
reported constantly, in the form of new catalysts, molten salts by their low melting points: generally
novel catalytic reactions, and alternative methodolo- <100–150 ◦ C, although most described in this review
gies. Much of the pressure for this is driven by the are molten at room temperature. They are also gener-
economic requirement to develop systems in which ally much less corrosive than their high melting point
easy separation of products and reuse of catalyst is relatives. This means that they can be conveniently
possible, along with high reactivity and selectivity. A used in place of conventional organic solvents, for
significant advance in recent years has been the advent example, in catalytic processes as will be outlined in
of biphasic catalysis, where the catalyst is isolated in this review.
one phase and the product remains in another, thus The history of ionic liquids goes back to 1914,
allowing easy product isolation and catalyst reuse. when Walden reported the synthesis of ethylammo-
As part of this effort, an ever-increasing degree of nium nitrate (m.p. 12 ◦ C) [1]. This material is formed
interest has been focussed on ionic liquids as media simply by the reaction of ethylamine with concen-
trated nitric acid, but its discovery did not prompt any
∗ Tel.: +44-141-548-2285; fax: +44-141-548-4822. great amount of interest at the time. In the 60 or so
E-mail address: [Link]@[Link] (C.M. Gordon). years that followed, although the majority of work on

0926-860X/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 6 - 8 6 0 X ( 0 1 ) 0 0 8 3 4 - 1
102 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117

molten salts involved their use as high temperature also covering such diverse applications as separations
media for electrochemical and other studies, a steady [15] electrochemistry [16] photochemistry [17] and
stream of work also reported the use of lower melt- liquid crystals [18]. Overall, however, the potential
ing salts as solvents for organic reactions. Much of applications in biphasic catalysis arguably show the
this work involved the use of chloroaluminate-based greatest potential for finding industrial applications.
salts such as AlCl3 –NaCl eutectics and pyridinium The aim of this review is two-fold. Firstly, to high-
hydrochloride, and is summarised in an excellent light the papers that have appeared over the past 18
review by Pagni [2]. The use of salts based on organic months or so in the area of ionic liquid based bipha-
cations was quite limited in this period, however. In sic catalysis. In doing so, it is intended as a general
1967, for example, Swain et al. described the use of supplement to the excellent reviews of Welton [19],
tetra-n-hexylammonium benzoate as a solvent for ki- Holbrey and Seddon [20], and Wasserscheid and
netic studies and electrochemical reactions [3]. Most Keim [21], to which the reader is directed for a full
of the salts employed still displayed melting points coverage of the earlier literature. Secondly, three ar-
above room temperature, however. Probably the most eas potentially of great interest will be discussed in
significant discovery in this area, however, was by some detail, notably the use of supercritical fluids
Hurley [4] and Hurley and Weir [5] in 1948, who with ionic liquids, palladium catalysed reactions, and
developed room temperature liquid chloroaluminate biocatalysis in ionic liquids. These are areas which
melts for applications in aluminium electroplating. have attracted particular interest over the last year
Room temperature ionic liquids only really reached or so, but which are also potentially very exciting
a more general audience with the reopening of devel- areas of development for catalysis using ionic liquids
opment in this area by the groups of Osteryoung and as solvents. The remainder of the review will bring
co-workers [6,7] and Wilkes et al. [8] in the 1970s. up to date other areas of catalysis using other tran-
Through the 1980s chloroaluminate ionic liquids were sition metals and Lewis acids. Where possible, the
studied by, especially the groups of Hussey et al. [9] advantages or disadvantages of the ionic liquid-based
and Seddon et al. [10], as solvents for transition metal systems will be benchmarked against alternative sys-
complexes, mainly through electrochemical and spec- tems to allow a true comparison to be made, and an
troscopic investigations. The first report on the use assessment of the future potential will be assessed.
of this type of low melting ionic liquids as reaction
media for organic synthesis was in 1986, as combined
solvents and catalysts for Friedel–Crafts reactions 2. Synthesis and properties of ionic liquids
[11]. Their first applications as solvents for biphasic
catalysis came in 1990 by Chauvin et al., who reported The reviews mentioned in the previous section
the dimerisation of propene by nickel complexes provide an excellent source of information for much
dissolved in acidic chloroaluminate melts [12] and of the background on the preparation and properties
Osteryoung who reported the polymerisation of ethy- of ionic liquids, and so only more recent develop-
lene by Ziegler–Natta catalysts [13]. The main prob- ments will be discussed here. Almost all of the work
lem with the ionic liquids based on chloroaluminate described in this review has been carried out in neutral
anions, however, remained their sensitivity to water ionic liquids based on 1,3-dialkylimidazolium cations
and oxygen. In addition, these liquids are incompatible (Fig. 1), with 1-butyl-3-methylimidazolium [bmim]+
with many organic compounds such as alcohols and being probably the most common cation. Other cations
acetone. A major breakthrough occurred in 1992, with
the report by Wilkes’ group of the synthesis of a series
of air- and moisture stable imidazolium salts based
on anions such as [BF4 ]− and [PF6 ]− [14]. Since this
time, the number of different ionic liquids reported has
expanded enormously. Ionic liquids have been applied
to a wide range of applications, not only within the Fig. 1. Examples of typical 1-alkyl-3-methylimidazolium cations
fields of biphasic catalysis and organic synthesis, but and the abbreviations used to refer to them in this article.
 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117 103

employed to a lesser extent include 1-alkylpyridinium, Table 1

1,1-dialkylpyrrollidinium, tetraalkylammonium and Densities and viscosities of typical ionic liquids
tetraalkylphosphonium. The most common anions Ionic liquid Density at Dynamic viscosity
are [PF6 ]− and [BF4 ]− , particularly for catalytic ap- 20 ◦ C (g cm−3 ) at 20 ◦ C (cP)
plications, although in principle any anion could be [bmim][PF6 ] 1.37a,b 330c
employed. The synthesis of these salts is described [bmim][BF4 ] 1.24a,d 154d
elsewhere [19,21]. [emim][Tf2 N]e,f 1.52g,h 34g
[bmim][Tf2 N] 1.43g 52g
Ionic liquids of this type have displayed the useful
[emim][TfO] 1.39g,h 45g
combination of low melting point, along with high [bmim][TfO] 1.29g 90g
thermal and chemical stability necessary for a reac- a At 30 ◦ C.
tion solvent. They are also generally good solvents b See [22].
for transition metal complexes, but poorly miscible c Our work.
with non-polar organics such as alkanes and ethers, d See [23].
e emim = 1-ethyl-3-methylimidazolium.
allowing the formation of biphasic systems and the
f Tf = CF SO .
use of homogeneous catalysts. Some ionic liquids g See [24].
3 2

(for example those based on anions such as [PF6 ]− h At 22 ◦ C.

and [(CF3 SO2 )2 N]− ) are also immiscible with water,
introducing the possibility of forming triphasic sys-
tems or alternative separation methods. Not all of the by the combination of organic halide salts and Lewis
systems reported here are “typical” biphasic systems, acids (e.g. chloroaluminate melts) are added to the
where reactants and catalyst exist in two mutually im- equation, acidity and basicity can also be introduced.
miscible layers. In many cases, the reaction mixture When ionic liquids are compared with other sol-
is homogeneous, but when the reaction is complete vents widely used in biphasic catalysis, a number of
a biphasic system is created by the addition of an clear advantages can be seen. Firstly, transition metal
extraction solvent immiscible in the ionic liquid. complexes are generally soluble in ionic liquids with-
The other property unique to ionic liquids as out chemical modification of the type that is often
reaction solvents is their lack of measurable vapour needed for aqueous and fluorous systems. The ionic
pressure. This means that they emit no volatile organic liquids used in catalytic systems are also generally
compounds, and also introduces the additional possi- regarded as polar but weakly coordinating, a concept
bility of removal of products by distillation without that will be expanded upon later in this section. This
contamination by the solvent. It also facilitates the means that, although the catalysts are very soluble in
recycling of ionic liquids, which is likely to be a finan- the ionic liquid phase, a non-coordinating anion will
cial necessity in any industrial application given the mean that the active site is very accessible to organic
greater cost of these materials than conventional or- substrates. Furthermore, ionic liquids are generally
ganic solvents. Properties such as melting point, den- chemically inert towards both catalysts and reactive
sity, viscosity and solvation ability can be finely tuned intermediates, meaning that catalyst stability is not a
by alteration of the cation or anion [21]. In general, problem. Finally, the miscibility of ionic liquids can
ionic liquids have densities >1, and thus exist as the be controlled by altering their chemical makeup, thus
lower phase in most biphasic systems. They are also allowing “tuning” of the system towards formation of
generally quite viscous compared with conventional a biphasic system, as well as optimisation of reaction
organic solvents, which means that phase separation conditions.
from organic solvents is often more rapid than with Up until recently, little quantitative information
two solvents of similar viscosity. Examples of density existed regarding the nature of interactions between
and viscosity for typical imidazolium-based ionic liq- ionic liquids and different types of solute. Such infor-
uids are given in Table 1 [22–24]. From these data, it mation is obviously of great importance if we are to
can be seen that the density decreases, but the viscosity understand why certain types of reaction occur par-
increases as the length of the alkyl chain substituent on ticularly favourably in these media. A simplistic view
the imidazolium ring increases. When liquids formed might be that ionic liquids are composed entirely of
104 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117

ions, and are therefore, very polar. As will be ex- The probes discussed above give little indication of
plained below, this is at best a great oversimplification, the role of the anion in solvation, so an alternative
and in some cases incorrect. One of the major dif- probe molecule is required to quantify this effect. The
ferences between ionic liquids and a typical organic square planar Cu(II) salt 1 has been shown to give
solvent is that the former is a binary mixture of a good correlation between the donor number of a
two different species, and thus is likely to engage in solvent and the λmax value for the lowest energy d–d
a much wider range of solvent–solute interactions. band [28] arising from changes in the splitting of the
Recently, therefore, a number of research groups d-orbitals as the solvent coordinates at the axial sites
have carried out investigations using solvatochromic on the metal centre.
probes to try to gain more insight into this important
area. Studies using Nile Red [25] and Reichardt’s
dye [26] have suggested that ionic liquids based on
1-alkyl-3-methylimidazolium cations act as H-bond
donors to the same degree as short chain alcohols, and
can thus be regarded as being relatively polar. Previ-
ous studies on alkylammonium salts using Reichardt’s
dye have shown that tetraalkyl derivatives are rela- It has also been shown that this salt can be used to
tively non-polar, while those containing cations of the estimate the donor number of anions in solution [29].
type [NHx R(4−x) ]+ (x ≥ 1) are considerably more The position of λmax has been recorded for a range of
polar. The fluorescent probes 4-aminophthalimide ILs based on the [PF6 ]− , [TfO]− and [Tf2 N]− anions,
and 4-(N,N-dimethylamino)phthalimide have also and the results indicate that the strength of coordina-
been applied to investigations of the polarity of tion at the Cu(II) centre is entirely anion-dependent
ionic liquids [27]. This study supported the mea- [26]. The results gained using solvatochromic probes
surements with Reichardt’s dye [26] indicating that have recently been correlated with the reactivity of
the polarity of 1-alkyl-3-methylimidazolium based Ni-catalysed oligomerisation of ethylene carried out
ILs was similar to that of short chain alcohols, and in ionic liquids and other solvents [30]. Details of this
that the polarity decreased somewhat as the alkyl system are given in Section 5.
chain length increased. In these studies, changing the Clearly our understanding of the solvent properties
anion appeared to have little effect on the polarity of ionic liquids remains quite incomplete, as only
of the ionic liquid. 1-Butylpyridinium tetrafluorobo- a handful of solvatochromic probes have been em-
rate was also measured using these probes, and this ployed, and only for a small range of ionic liquids.
proved to exhibit a lower polarity than the imida- The concept of “polarity” in ionic liquids must be
zolium salts. One feature highlighted in this study considered carefully as the presence of both cations
was the influence of water on position of λmax , of and anions can lead to far more possible interactions
4-(N,N-dimethylamino)phthalimide which shifted than in conventional molecular solvents. Nevertheless,
from 526 nm in water-saturated [bmim][PF6 ] (wa- the results shown above have started to give a picture
ter content ∼0.324 M) to 512 nm (i.e. less polar) of imidazolium-based ionic liquids as media which
in [bmim][PF6 ] that had been dried by heating in are polar, but whose basicity can be controlled by the
vacuo for 24 h (water content ∼0.015 M). Another nature of the anion present. It is this unique combi-
example of this is shown by work from the author’s nation of factors that seems to make these materials
own research group, where the λmax of Reichardt’s so suitable for use in catalytic applications.
dye displayed a shift from 526 nm in water-saturated
[bmim][PF6 ] to 546.5 nm in dry solvent. This clearly
shows that for any physical measurements carried out 3. Supercritical fluids and ionic liquids
in ionic liquids, the water content should be estab-
lished. Clearly, the presence of water may also influ- One of the recurrent problems with the use of
ence the performance of an ionic liquid in catalytic ionic liquids for organic synthesis is the extraction of
applications. products, owing to the lack of volatility of the liquids
 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117 105

themselves. In biphasic systems this problem may be

alleviated, as the products form a separate phase from
the ionic liquid layer, which generally contains the
catalyst. If a truly biphasic system cannot be achieved,
organic products can usually be extracted using sol- (1)
vents that are immiscible with the ionic liquid (in many
cases diethyl ether or an alkane). Although leaching of The product was then extracted in an extremely pure
catalyst into the organic phase can be minimised, the form into scCO2 , without contamination from either
small mutual solubility of the two phases means that the ionic liquid or the Ru complex. Conversions of up
in many cases investigated to date, a degree of catalyst to 99% combined with ca. 90% ee were obtained, and
loss is observed when the ionic liquid/catalyst system the catalyst solution could be reused for four further
is recycled. One novel approach that reduces this pos- cycles with no apparent decrease in catalyst activity.
sibility is the use of supercritical CO2 (scCO2 ) as the The influence of addition of water, propan-2-ol and
extractant for the organic products. Blanchard et al. AgPF6 was also investigated, along with that of higher
first suggested the possibilities of this method [31] H2 pressures. No kinetic data were recorded in this
showing that naphthalene could be extracted quantita- study, however. The authors also reported the asym-
tively by scCO2 from [bmim][PF6 ]. Subsequent work metric hydrogenation of isobutylatropic acid, to give
by the same group has shown that a wide variety of the antiinflammatory drug ibuprofen (2, Eq. (2)).
solutes may be extracted in this manner, without any
contamination by the ionic liquid [32]. This is a result
of the observation that, although CO2 is extremely sol-
uble in ionic liquids, the reverse is not the case, with
no appreciable IL solubilisation in the CO2 phase. One
important observation was that the solubility of CO2
in ILs was very dependent on the water content. Water
saturated [bmim][PF6 ] can contain up to 2.3 wt.% wa-
ter, and has a CO2 solubility of only 0.13 mol fraction
compared with 0.54 mol fraction CO2 in dried (ca. (2)
0.15 wt.% water) [bmim][PF6 ] at 57 bar and 40 ◦ C
[33]. Thus, effective use of scCO2 as an extractant may In this case, poor enantioselectivity was observed
be limited to systems that can be kept water-free. Spec- in wet ionic liquid, but when methanol was added an
troscopic evidence that CO2 dissolves in [bmim][PF6 ] ee of 85% was observed using 100 bar H2 . This is
and [bmim][BF4 ] was also reported recently Kazarian higher than the enantioselectivity observed in aque-
et al. using ATR-IR [34]. These investigations gave an ous/organic biphasic systems.
estimate of 0.6 mol fraction for the solubility of CO2 Hydrogenation reactions have also been investi-
in [bmim][PF6 ] at 68 bar and 40 ◦ C. The first report gated in a combined scCO2 -ionic liquid system. Liu
of an attempt to combine scCO2 (and scC2 H6 ) and et al. report the hydrogenation of dec-1-ene and cyclo-
[bmim][BF4 ] for Rh colloid-catalysed hydrogena- hexene using Wilkinson’s catalyst RhCl(PPh3 )3 [37].
tion of arenes was not promising, however [35]. The The system remained biphasic throughout the reac-
authors suggested that the lack of activity towards tion, and for the hydrogenation of dec-1-ene a con-
hydrogenation might have resulted from deactivation version of 98% after 1 h was reported, corresponding
of the catalyst by traces of Cl− in the ionic liquid. to a turnover frequency (TOF) of 410 h−1 . The results
Following the studies reported above, three succes- suggest, however, that there is no reactivity advan-
ful reports of catalysis using combined IL/scCO2 tage with using scCO2 in place of hexane in biphasic
systems have been reported. Asymmetric hydrogena- reactions of this type. The same paper reports the
tion of tiglic acid using Ru(O2 CMe)2 ((R)-tolBINAP) RuCl2 (dppe)2 -catalysed (dppe = Ph2 PCH2 CH2 PPh2 )
proceeded with excellent yield and selectivity in hydrogenation of CO2 in the presence of dialky-
[bmim][PF6 ] (Eq. (1)) [36]. lamines to produce N,N-dialkylformamides. This
106 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117

Scheme 1.

reaction was chosen as it involves ionic carbamate the reaction system could be recycled two or three
intermediates, as shown in Scheme 1, which would times, after this time the catalyst was deactivated.
be expected to be much more soluble in the IL phase Greatly improved results were observed, however,
than in scCO2 . when [bmim][Ph2 PC6 H4 SO3 ] was used as the lig-
This proved to be the case, and the reaction and in place of P(OPh)3 . In this case, the activity
of di-n-propylamine in a mixed scCO2 –IL system of the catalyst remained high over 12 runs (TON =
resulted in complete amine conversion, along with 160–320 h−1 ), although the l:b ratio fell steadily
exclusive formation of N,N-di-n-propylformamide. from 3.7:1 to 2.5:1. No leaching of the rhodium
This product was very soluble in the ionic liquid was observed over the first nine cycles, but after
phase, partitioning only poorly into scCO2 and re- this time it became significant owing to formation
sulting in disappointing yields after the first reaction of scCO2 soluble [RhH(CO)4 ]. Finally, a continuous
cycle. It was found, however, that the isolated yield flow scCO2 /[bmim][PF6 ] system was used for the
of N,N-di-n-propylformamide increased from <5% in hydroformylation of oct-1-ene in conjunction with
the first reaction cycle to quantitative recovery by the [pmim]2 [PhP(C6 H4 SO3 )2 ] as the ligand. In this case,
third and fourth cycle, suggesting that the IL phase the reaction was slow for the first 8 h of reaction. After
becomes saturated with product over the first two this time, however, a linear plot of catalyst turnover
reaction cycles. versus time indicated that no catalyst decomposition
The final example reported to date describes the was occurring over periods up to 72 h. The l:b ratio
use of a continuous flow scCO2 /[bmim][PF6 ] sys- was reported as >3:1 over the whole reaction time,
tem for the rhodium catalysed hydroformylation of and <1 ppm Rh was identified in the product stream.
hex-1-ene [38]. With [Rh2 (OAc)4 ]/P(OPh)3 as the The initial reports discussed above indicate that
catalyst source, carrying out the reaction in pure the combination of ionic liquids with supercritical
[bmim][PF6 ] resulted in high yields but a very poor fluids represents an extremely promising combination
aldehyde selectivity (15.7%), and a poor linear: of techniques for biphasic catalysis. The problem of
branched (l:b) product ratio. When a combined scCO2 / catalyst leaching into the product stream is clearly
[bmim][PF6 ] system was employed, the reaction rate greatly reduced compared with traditional biphasic
fell, but much improved selectivity was observed systems. It is likely that many more examples will be
(82.3% aldehyde; 6.1:1, l:b). Unfortunately, although reported in the near future.

Scheme 2.
 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117 107

4. Pd catalysis versions were obtained using a range of arenes when

the quantity of dppp was doubled to 2 equiv. in the
Some of the most promising applications of biphasic ionic liquid reactions.
catalysis using ionic liquids have involved Pd-based In addition to the Heck reaction, a number of other
catalysts. Most of the reactions studied to date have palladium-catalysed processes have been reported
been C–C coupling reactions, and clear advantages in recently, principally cross-coupling reactions.
catalyst recycling and product separation have been Mathews et al. reported that Suzuki cross-coupling
previously shown to result when Heck reactions are reactions using Pd(PPh3 )4 as catalyst and [bmim][BF4 ]
carried out using both alkylammonium- [39–41] and as solvent gave excellent yields and TONs at room
imidazolium-based ionic liquids [42,43] as solvents. temperature (Eq. (3)) [47].

(3)

It has been suggested that this is in part a result of the In this system, it was found that optimum catalytic
in situ formation of palladium carbene complexes in activity was achieved by pre-heating the catalyst with
the imidazolium-based ionic liquids [44]. the aryl halide in [bmim][BF4 ] at 110 ◦ C, after which
A recent publication has reported that the use of the reaction was started by addition of the arylboronic
a Pd-benzothiazole carbene complex in combination acid and Na2 CO3 at room temperature. An extremely
with tetrabutylammonium bromide as solvent can large increase in reaction rate was observed compared
result in conversion of bromobenzene to butyl cin- with the conventional Suzuki conditions. The reaction
namate with a yield of 94% in only 10 min when of bromobenzene with phenylboronic acid under con-
sodium carbonate is used as the base [45]. Another ventional Suzuki conditions result gave an 88% yield
recent report on the Heck reaction in ionic liquids in 6 h (TON, 5 h−1 ), while the equivalent reaction in
describes how [bmim][BF4 ] has been successfully [bmim][BF4 ] gave 93% in 10 min (TON, 455 h−1 ).
used for the regioselective arylation of an electron The reaction products were typically extracted using
rich olefin, butyl vinyl ether [46]. Reactions of this diethylether, while the by-products (NaHCO3 and
type generally require the use of aryl triflates, silver NaXB(OH)2 (X = halide)) were removed using ex-
triflate, or thallium acetate in order to obtained good cess water. After this the ionic liquid/catalyst system
yields. The active catalyst was prepared in situ from was used for three further reaction cycles with no
Pd(OAc)2 and 1,3-bis(diphenylphosphino)propane decrease in either yield or TON.
(dppp), and the reaction in [bmim][BF4 ] was com- Palladium catalysed Negishi cross-couplings of
pared with that in four conventional organic solvents. organozinc reagents were achieved in 1-butyl-2,3-di-
Although a conversion of only 50% was observed, methylimidazolium tetrafluoroborate ([bdmim][BF4 ])
the ␣-arylated product was formed regioselectively using a novel ionic phosphine prepared by reaction of
(Scheme 2). PPh2 Cl with [bmim][PF6 ] (3) [48].
In DMF and DMSO, quantitative conversions of
the starting material were observed, but a significant
proportion of the ␤-arylated product was formed in
each case, while toluene and acetonitrile gave both
poor conversions and a mixture of products. Fur-
thermore, after the reaction in ionic liquid palladium
black was not observed, as was always the case with Yields of 70–92% were obtained using a variety
the conventional solvents. Finally, quantitative con- of substrates, the fastest reactions being observed
108 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117

for aryl iodides. Recycling of the catalyst/IL system

was attempted, but after the third cycle a decrease
in yield and increase in reaction time was observed,
suggesting that catalyst decomposition or leaching
was occurring. Handy et al. have reported a range of Scheme 3.
Stille couplings using PdCl2 (PhCN)2 /Ph3 As/CuI in
[bmim][BF4 ] [49]. Good yields and catalyst recycla- Palladium-catalysed allylic substitution in ionic
bility were reported for the reaction of ␣-iodenones liquids was first reported by Chen et al. in 1999 [50].
with vinyl and aryl stannanes, although the reaction The same group has also reported that the nature of
rates reported were considerably lower than those the phosphine ligand employed can have profound
obtained in N-methylpyrrolidinone. The formation of effects on the activity of the palladium catalyst [51].
diaryl compounds in good yields by the same reaction Electron donating phosphines such as PCy3 gave the
was also reported. The yields obtained in this reac- fastest rates, while use of electron acceptors such as
tion proved to be extremely dependent on the catalyst P(OPh)3 resulted in very low levels of conversion. A
employed. For example, the reaction of p-iodoanisole comparison between [bmim][BF4 ] and THF showed
and phenyl tributyltin proceeded with a yield of 82% that the ionic liquid allowed the use of a wider range
using the PdCl2 (PhCN)2 /Ph3 As/CuI catalyst system, of phosphines for the same reaction. This was as-
while Pd(PPh3 )4 gave a yield of 30%, and no product signed to mechanistic differences between the two
was observed with Pd(OAc)2 . With aryl bromides on solvents, with the ionic liquids enhancing charge sep-
the other hand, the PdCl2 (PhCN)2 /Ph3 As/CuI sys- aration in the intermediate allylpalladium complex,
tem gave only 35% yield, while Pd(PPh3 )4 gave the and thus reducing the likelihood of a reverse reaction
desired product in 90% yield. Competition reactions (Scheme 3).
indicated that the reaction remained chemoselective An enantioselective allylic substitution reaction
to the weaker C–I bond in all cases, however, as between (rac)-(E)-1,3-diphenyl-3-acetoxyprop-1-ene
indicated below (Eq. (4)). and dimethylmalonate has also been reported, us-
ing chiral Pd0 -ferrocenylphosphine complexes in
[bmim][PF6 ] (Scheme 4) [52]. The best yield (81%)
and ee value (74%) were obtained in a homoge-
neous system using a catalyst prepared in situ from
Pd2 (dba)3 ·CHCl3 and ligand 4.
On recycling, this system gave a somewhat reduced
yield, but a similar ee value. When the catalytic reac-
tion was carried out in a biphasic system using toluene
(4) or cyclohexane as the second phase, similar yields and

Scheme 4.
 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117 109

Scheme 5.

ee values to those from the homogeneous system were biphasic hydrogenation of a range of conjugated
obtained from the first reaction cycles, but recycling and non-conjugated dienes to form monoenes [54].
of the IL/catalyst system resulted in much reduced Good selectivity for hydrogenation of only one
enantioselectivities. double bond was observed in all cases, generally
Mizushima et al. have reported the palladium-cata- with trans-selectivity where applicable (Scheme 6).
lysed carbonylation of aryl halides in [bmim][PF6 ] Turnover number values were similar for both ionic
and [bmim][BF4 ] [53]. The reaction between a range liquids (TON 982 for [bmim][BF4 ], and 885 for
of alcohols and bromobenzene proved to give greatly [bmim][PF6 ]).
enhanced yields when [bmim][BF4 ] was present It should be noted, however, that for 1,3-butadiene
(82–91% depending on the alcohol) compared with the selectivity and TON values reported for reactions
the reaction in the absence of the IL (26–32%). Some- carried out in the ionic liquids differed little from those
what lower yields were obtained using [bmim][PF6 ]. obtained in neat diene and in CH2 Cl2 solution. Set
It was found that the catalyst/IL mixture could be against this, separation of the products and reuse of
recycled up to four times, although with decreasing the catalyst solution was clearly easier in the ionic
yield for each successive cycle. The recyclability of liquid reaction. In the hydrogenation of 1,3-butadiene,
the catalyst was improved by increasing the ratio of the catalyst solution was recycled 15 times without
PPh3 :Pd(OAc)2 from 4:1 to 20:1, a move that also in- apparent decrease in activity. The rate of conversion
creased the yield on the initial reaction to 99% in both also showed great dependence on H2 pressures.
ionic liquids. When the same reaction is carried out
under higher pressures of CO, double carbonylation
can occur, forming the ␣-ketoester (Scheme 5). With 5. Other transition metal catalysed reactions
Pri OH as the nucleophile, the single carbonylation
product was formed with much greater selectivity in Although ionic liquids are generally much better
[bmim][PF6 ] than in neat Pri OH. solvents for transition metal catalysts than the organic
When the nucleophile was NEt2 H, however, little products in biphasic systems, catalyst leaching can still
difference was observed in the presence of an ionic be a problem, particularly in reactions such as hydro-
liquid, although the catalyst could once again be re- formylation. As a result, reactions using novel cationic
cycled. guanidinium-modified phosphine ligands have been
Finally, Dupont et al. have reported the use of reported by Wasserscheid et al. [55]. Ligand 5 has
Pd(acac)2 (acac = acetylacetonate) dissolved in been used for the Rh-catalysed hydroformylation of
[bmim][BF4 ] and [bmim][PF6 ] for the selective oct-1-ene in [bmim][PF6 ].

Scheme 6.
110 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117

Although the conversions obtained were lower on

the first cycle than those using PPh3 as the ligand, solvent probe 1 only slightly more strongly than
much lower levels of leaching were observed (<0.07% CH2 Cl2 .
compared with 53%). This meant that the levels of Relatively few transition metal catalysed oxida-
conversion remained similar on recycling of the cata- tion reactions have been reported in ionic liquids,
lyst/IL mixture. The selectivity for the n-isomer was the first being the epoxidation of allylic ethers and
relatively low using 5, however. When the reaction was allylic arenes using a chiral Mn (III)(salen) complex
repeated using a specially prepared xanthene based [56]. More recently, Owens and Abu-Omar reported
bidentate phosphine ligand 6, the conversion steadily the epoxidation of alkenes and allylic alcohols using
increased as the catalyst was reused (from 10.6% on methyltrioxorhenium and urea hydrogen peroxide
the first run to 44.3% on the seventh run), and a much (UHP) in [bmim][BF4 ] [57]. In almost all cases, con-
better n:iso ratio of ca. 20:1 was obtained. Once again, versions and selectivity towards epoxide formation
little catalyst leaching was observed. were excellent. When aqueous hydrogen peroxide was
Alkene oligomerisation has been one of the more used in place of UHP, however, the diol was formed
widely studied of transition metal catalysed reac- almost exclusively. Although no direct comparison
tions in ionic liquids. Recently, the oligomerisation of with conventional solvents was made, reaction times
ethylene has been achieved using a cationic Ni-based were reported to be comparable. Also, no reference
catalyst both in a biphasic system where the cata- was made regarding whether the catalyst retained
lyst was immobilised in [bmim][PF6 ], and homo- activity when recycled.
geneously using CH2 Cl2 as the solvent [30]. Using Song et al. have reported the Cr(salen) catalysed
the ionic liquid, a much higher selectivity for linear asymmetric ring opening of epoxides in a range of
alk-1-enes was achieved than in any previous sys- [bmim]+ based ionic liquids [58]. Excellent yields and
tems. Moreover, the rate of reaction was nearly seven ee values were obtained using hydrophobic ionic liq-
times faster in [bmim][PF6 ] (TOF = 12712 h−1 ) than uids [bmim][PF6 ] and [bmim][SbF6 ] (see Scheme 7),
in CH2 Cl2 (TOF = 1852 h−1 ). Effectively no con- while in the water miscible salts [bmim][BF4 ] and
version was observed in butane-1,4-diol, however. [bmim][TfO], little or no product was formed under
This solvent gives a biphasic system, and is used the same conditions. This was perhaps surprising, as
successfully with neutral Ni complexes (for example the authors report that the catalyst was fully dissolved
the Shell SHOP ethylene oligomerisation process). in the latter solvents, while it remained as a suspen-
Studies using solvatochromic dyes indicated that, al- sion in the hydrophobic liquids.
though [bmim][PF6 ] has a polarity almost identical In order to aid catalyst immobilisation, the reaction
to butane-1,4-diol, it coordinates with Cu(II)-based was also carried out in a 5:1 mixture of [bmim][PF6 ]

Scheme 7.
 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117 111

and [bmim][TfO], giving a good combination of cata- mixture in the ionic liquid at room temperature, unlike
lyst activity and immobilisation (68% yield, 94% ee). the same system in toluene. Polymerisation with
This system allowed reuse of the catalyst over five excellent levels of conversion occurred at tempera-
cycles with no loss of activity or ee and with little tures down to 30 ◦ C, giving products of relatively low
leaching of catalyst. Unfortunately no comparison polydispersity and molecular mass. The high solubil-
with the reaction yield, rate or stereoselectivity in ity of the catalyst in the ionic liquid meant that sep-
conventional media was reported in this paper. aration of the polymer could be achieved simply and
In a final example of oxidation, Howarth has efficiently by washing with toluene (in which PMMA
reported the first example of conversion of aromatic is soluble but [bmim][PF6 ] immiscible), leaving the
aldehydes to aromatic alcohols in an ionic liquid ionic liquid/catalyst solution available for reuse. No
[59]. The catalyst used was [Ni(acac)2 ] (3 mol%), indication was given of the activity of the recycled
with O2 at atmospheric pressure the oxidant. The catalyst, however.
yields obtained were in the range 47–66%, lower than
those obtained using perfluorinated solvents, possibly
because of the higher solubility of O2 in these me- 6. Lewis acid catalysis
dia. In the latter system, however, it was necessary
to modify the catalyst with perfluorinated tails, thus Ionic liquids based on Lewis acids such as AlCl3
adding to the cost and complexity of the process. have been applied to a number of Lewis acid catalysed
Recycling of the catalyst/IL system for three fur- organic transformations, for example Friedel–Crafts
ther cycles without diminution of the yield was also reactions [11,62]. More recently, there have been a
reported. number of reports of reactions utilising neutral ionic
Moving on to other classes of reaction, the cy- liquids in combination with Lewis acids, in some
clodimerisation of 1,3-butadiene by iron complexes cases with extremely large effects on reactivity and se-
in [bmim][BF4 ] and [bmim][PF6 ] has been reported lectivity. One of these reports that 20 mol% Sc(OTf)3
[60]. In the latter solvent, a TOF of up to 1404 h−1 can catalyse the Friedel–Crafts alkylation of aromatic
was obtained, with 100% selectivity for 4-vinyl- compounds in a range of hydrophobic 1-alkyl-3-
1-cyclohexene. The catalyst, prepared by reduction methylimidazolium [PF6 ]− and [SbF6 ]− ionic liquids
of [Fe(NO)2 Cl]2 by Zn, Et2 AlCl or n-BuLi in situ, [63]. Conversions were reported to be effectively
could be reused several times without affecting per- quantitative, with isolated yields also high (65–93%).
formance. The same reaction with isoprene gave rise Considerable isomerisation of the hex-1-ene starting
to a mixture of cyclic dimers. material was observed in these reactions, as indicated
There have been few reports of transition metal by the formation of two different isomeric products
catalysed polymerisation reactions in ionic liquids, in solvent-dependent ratios of between 1.5:1 and 2:1
but Carmichael et al. have recently reported the Cu(I) for products 7 and 8, respectively (Scheme 8). Sur-
mediated living radical polymerisation of methyl- prisingly, in water miscible ionic liquids based on
methacrylate (MMA) in [bmim][PF6 ] [61]. The the [BF4 ]− and [TfO]− anions, as well as a range of
catalyst is based on a mixture of CuBr and N-propyl- conventional solvents, no reaction was observed at
2-pyridylmethanimine, which form a homogeneous all. The reaction between cyclohexene and benzene

Scheme 8.
112 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117

Scheme 9.

was carried out using the same ionic liquid/Sc(OTf)3 using Sc(OTf)3 as a catalyst extremely high yields and
system three times, with no decrease in yield. endo-selectivity can be obtained [69]. The use of only
Zulfiqar and Kitazume have reported the forma- 0.2 mol% Sc(OTf)3 in [bmim][PF6 ] resulted in a huge
tion of 2-methyl-2,3-dihydrobenzo[b]furan in 62% increase in rate compared with that obtained with the
yield via the Sc(OTf)3 -catalysed Claisen rearrange- same concentration of Lewis acid in CH2 Cl2 . Rather
ment/cyclisation sequential reaction of allyl phenyl surprisingly, similar rate enhancements were observed
ether in 8-ethyl-1,8-diazabicyclo[5,4,0]-7-undecenium even when the reaction was carried out in CD2 Cl2
trifluoromethanesulfonate ([EtDBU][TfO]) (Scheme 9) with just 1 equiv. of [bmim][PF6 ] present. Isolated
[64]. Similar results were gained using o- and p-tolyl yields of 71–96% were obtained using a range of
analogues. When the reaction was carried out in typical substrates, and in all cases studied, endo:exo
[bmim][PF6 ] and [bmim][BF4 ] much lower yields ratios were >99:1. The ionic liquid/Sc(OTf)3 mixture
were obtained, indicating that the nature of the cation was then reused for 10 subsequent reaction cycles
has a pronounced effect on the reaction rate in this without loss of activity. In a related study, Zulfiqar
system. No indication was given regarding how the and Kitazume reported the use of microencapsulated
reaction progressed in the absence of ionic liquid, Sc(OTf)3 for the catalysis of aza-Diels–Alder reac-
however. The reaction mechanism was confirmed tions in [emim][TfO] and [EtDBU][TfO] (Scheme 10)
by showing that 2-allylphenol also formed the same [70].
product when heated in a mixture of Sc(OTf)3 in Similar yields were obtained using both ionic
[EtDBU][TfO]. Reuse of the catalyst/ionic liquid liquids. The reaction could be carried out using both
mixture was achieved with no apparent decrease in preformed imines, and also with the imine formed
activity over three cycles. in situ by the reaction of an aldehyde and an amine
There have been a number of reports on the use in the ionic liquid. Both ionic liquid/catalyst systems
of ionic liquids as solvents for Diels–Alder reactions could be reused with no apparent decrease in activity.
[65–68]. One of these [66] noted that the use of Rosa et al. have reported the use of ionic liquids as
BF3 ·OEt2 and ZnI2 as Lewis acid catalysts resulted solvents for 1,4-diazabicyclo[2,2,2]-octane (DABCO)
in faster reaction rates and greatly enhanced stere- mediated Bayliss–Hillman reactions [71]. At room
oselectivities. A recent report has shown that when temperature, the reaction rate was enhanced by fac-
Diels–Alder reactions are carried out in ionic liquids tors of 14.1 and 33.6 compared with CH3 CN for

Scheme 10.
 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117 113

[bmim][BF4 ] and [bmim][PF6 ], respectively. When free alkylaluminium during the reaction. The most
catalytic quantities (20 mol%) of DABCO were em- remarkable observation was the large degree of rate
ployed, the rate enhancement in [bmim][PF6 ] was enhancement observed when successive reactions
reduced to 11.1 times faster than that in CH3 CN. were carried out using the same ionic liquid/nickel
Addition of a range of Lewis acids (which had pre- complex mixture. At −10 ◦ C, the productivity in-
viously been reported to increase the reaction rate) creased from 5 kg mol−1 h−1 for the first reaction, to
generally resulted in a reduction in both the rate of 96 kg mol−1 h−1 on the second cycle, and finally to
reaction and the yields obtained. 198 kg mol−1 h−1 on the third reaction cycle. Over
the same sequence, a decrease in the average molecu-
lar weight of the polyethylene products was observed,
7. Lewis acid-based ionic liquids with a bimodal distribution in runs 2 and 3. This was
assigned to an increase in the quantity of amorphous
Although ionic liquids based on Lewis acids such product formed in successive reactions.
as AlCl3 were the materials that really kick-started Aluminium(III) chloride-based ionic liquids
research into the area, this review clearly shows that immobilised on a range of silica and alumina sup-
such liquids have largely been supplanted by the ports have also been used as Lewis acid catalysts for
neutral ionic liquids for catalytic applications (and Friedel–Crafts alkylation reactions [73]. ZrO2 and
indeed most other applications). This is largely due TiO2 were found to immobilise the ionic liquid very
to the fact that this class of ionic liquids is gener- poorly, and were thus unsuitable for use as catalysts.
ally much more air- and moisture-sensitive than the Good conversions were reported for the reaction of
neutral salts, thus causing problems with handling dodecene with a range of aromatic compounds, gen-
and long-term stability. Furthermore, these liquids erally with good selectivity towards the formation of
are often incompatible with many classes of organic monoalkylated compounds. ICP analysis indicated
compound; for example, AlCl3 -based ionic liquids that negligible amounts of ionic liquid were leaching
react rapidly with acetone [19]. Despite this, some from the support, but deactivation of the catalyst was
new applications of such liquids have been reported observed when it was recycled. This was thought to
recently, as outlined below. Clearly the fact that the result from a combination of deactivation by water,
relative acidity or basicity of the reaction medium can as well as adsorption of reactants onto the catalyst
be modified over a wide range of values is potentially surface that caused blocking of the active sites. It was
a very useful for catalysis since the selectivity of reported that no reaction occurred when dodecene
reactions can be influenced by the acid strength of was added to the reaction before the aromatic com-
the catalyst. pound, while reversal of this order gave good levels
Ethylene has been efficiently polymerised us- of conversion, thus indicating that dodecene was the
ing a diiminenickel complex dissolved in acidic species causing catalyst deactivation. The reactions
[bmim] organochloroaluminate ionic liquids [72]. were also carried out in a continuous liquid-phase
The ionic liquids were formed using a mixture of reactor, and a slow deactivation of the catalyst with
ethylaluminium dichloride, aluminium(III) chloride time was observed. When the same reaction was at-
and 1-butyl-3-methylimidazolium chloride in a mo- tempted in a continuous gas-phase reactor, virtually
lar ratio of 0.32:1.00:1.00. The polymerisation re- no conversion was observed, however.
action occurred efficiently at temperatures ranging Deng et al. have reported the use of acidic 1-butyl-
from −10 to +10 ◦ C. Over this temperature range, pyridinium/AlCl3 -based ionic liquids as reaction
the activity of the catalyst increased with increasing media for esterification reactions [74]. In general,
temperature, but a decrease in the Tm value of the similar levels of conversion and selectivity were
resulting polymers with increasing reaction temper- observed to those obtained under conventional condi-
ature indicated that increased chain branching was tions using concentrated sulfuric acid as the catalyst.
occurring. The catalyst/ionic liquid mixture could be The esters formed were generally immiscible with
recycled, but trimethylaluminium had to be added for the ionic liquids, however, resulting in easy product
each successive cycle as the organic phase extracted isolation. The ionic liquids were recycled three times
114 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117

with some decrease in yield between the first and H2 O–[bmim][PF6 ] system, allowing easier separation
second reaction cycle. of the two phases when the reaction was complete.
Aluminium(III) chloride is not the only Lewis acid The same paper also reported the use of [bmim][PF6 ]
used to form ionic liquids. Wasserscheid et al. have for liquid–liquid extraction of erythromycin.
described the use of moderately Lewis acidic ionic Soon after the publication of this paper, Erbeldinger
liquids based on SnCl2 for the Pt catalysed hydro- et al. reported the use of [bmim][PF6 ] as a solvent for
formylation of functionalised and non-functionalised the thermolysin catalysed formation of Z-aspartame
olefins [75]. The Pt catalyst showed enhanced stability from carbobenzoxy-L-aspartane and L-phenylalanine
and selectivity in the ionic liquids for these reactions methyl ester hydrochloride (Eq. (5)) [77].

(5)

compared with conventional organic solvents. The Effectively quantitative conversions were reported
paper also reports a means for determining the Lewis after 50 h reaction, and rates of reaction were com-
acidity of chlorostannate ionic liquids using 119 Sn parable to those observed in conventional organic
NMR spectroscopy. solvents such as ethyl acetate. After reaction, the ionic
liquid could be successfully recycled for further reac-
tions without affecting the rate of yield of the process.
8. Biocatalysis Two features of this reaction should be noted. The
first is that the addition of a small (5% v/v) quantity
One of the most exciting recent developments in of water was required for efficient reaction. Secondly,
the use of ionic liquids for catalytic systems is the the thermolysin was observed to dissolve in the ionic
application of biocatalysis in these solvents. Only a liquid at concentrations up to 3.2 mg ml−1 , but was
small range of reactions has been reported to date, completely inactive at these concentrations. This
but the results reported suggest that this area could suggests that the enzyme is only active in suspension.
yield some exciting developments in the near fu- The first anhydrous system was reported by Madeira
ture. The first report was by Cull et al. in 2000 [76], Lau et al., who described the use of candida arctica
who reported the use of [bmim][PF6 ] as a solvent lipase B (CaLB) for transesterification, ammoniolysis,
for the biphasic hydration of 1,3-dicyanobenzene us- and epoxidation reactions [78]. In this case, the reac-
ing Rhodococcus 312. The reaction has been carried tion solvents were pure [bmim][PF6 ] or [bmim][BF4 ],
out in toluene solution, but this solvent can have a with no added water as in the previous two cases. The
damaging effect on the cell wall of the biological transesterification of ethyl butanoate with butan-1-ol
catalyst, as well as the known problems with toxi- occurred in 81% yield after 4 h in both ionic liquids
city and flammability. The reaction was carried out using a supported CaLB catalyst. This reaction was
in a biphasic H2 O–[bmim][PF6 ] system, and showed slightly faster than that observed in tert-butanol or
a lower initial rate of reaction than in an equivalent 1-butanol solution. The reaction using free CaLB in
H2 O–toluene system, but a slightly higher final yield. ionic liquids was much slower, but after 24 h con-
In the former system, the ionic liquid effectively acts versions around 80% were achieved. CaLB showed
as a reservoir for the organic substrate, while the cells lower activity in [bmim][BF4 ] than in tert-butanol
are present in the aqueous phase. The authors also in- for the ammoniolysis of ethyl octanoate, however.
dicated that less cell aggregation was observed in the Finally, CaLB catalysed epoxidation of cyclohexene
 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117 115

gave yields only slightly lower than those obtained solvents. Recycling studies on the CaLB-catalysed re-
in acetonitrile, previously identified as the optimum action showed that the ionic liquid and enzyme could
organic solvent for this process. be recycled twice without loss of enantioselectivity
The first example of enantioselective biocatalysis or reactivity.
in ionic liquids was reported by Schöfer et al. [79]. The area of biocatalysis combined with ionic liquids
This paper reports the screening of nine lipase sys- is clearly still in its infancy, with only a small range of
tems against ten different ionic liquids for the kinetic reactions and enzymes investigated to date. Probably
resolution of rac-1-phenylethanol by transesterifica- the most important finding of the preliminary reports
tion with vinyl acetate. The results gained were com- is the fact that many enzymes retain their activity in
pared with those obtained in methyl tert-butyl ether ionic liquids. As yet, no significant examples of rate
(MTBE), a widely used solvent for transesterification enhancement in ionic liquids have been reported, but
reactions. Quite different reactivity patterns were ob- the final two papers mentioned above clearly indicate
served in different ionic liquids, and these were differ- that advantages may be gained over reactions in or-
ent again from MTBE in many cases. As a result, no ganic solvents with regard to improvements in enan-
“best ionic liquid” was identified, but [bmim][Tf2 N] tioselectivity and reusability of the solvent-enzyme
appeared to give some useful behaviour. Most notably, systems. It is likely that many more examples will be
it gave excellent ee values (>98%) for five of the en- reported in the near future. As the range of examples
zyme systems, and high conversion (but little or no studied becomes greater, it will become necessary to
chiral resolution) with two others. Possibly the most investigate the nature of solvation of enzymes in ionic
significant observation, however, was the observation liquids.
that using this ionic liquid it was possible to remove
the products by distillation under reduced pressure, af-
ter which the enzyme-IL system could be reused three 9. Summary and conclusions
times with a loss in activity of <10% per cycle.
In a similar study, Kim et al. reported some trans-
This review shows that ionic liquids show great
esterification reactions of vinyl acetate with four
promise as reaction media for many types of catalysis.
different alcohol substrates in [bmim][BF4 ] and
The ability to control their properties very precisely
[bmim][PF6 ] [80]. The lipases employed were CaLB
sets them apart from other approaches such as aque-
(immobilised), and Pseudomonas cepacia (PCL,
ous or fluorous systems, although all of the techniques
native). The reactions were found to proceed with
remain complementary. Clearly ionic liquids remain
higher enantioselectivities than in THF and toluene
relatively costly compared with conventional organic
in all cases, with the hydrophobic [PF6 ]− salt usually
solvents or water, but this must be set against the fact
giving the best values. The PCL-catalysed reaction
that they are generally used in much smaller quantities,
of 1-chloro-3-phenoxypropan-2-ol with vinyl acetate
and are likely to be reused in most applications. Fur-
(Eq. (6)) was scaled up to a preparative level. Reac-
thermore, there remains a distinct lack of information
tions were carried out to just over 50% completion
regarding the role played by the ionic liquid in many
(48 h reaction time), giving unreacted substrates of
of the reactions discussed here. For example, why does
>99.5% ee in 43% yield, and to 46% completion
Sc(OTf)3 display such different levels of reactivity
(36 h) giving the acetylated product with >99.5% ee
in different types of ionic liquids? Also, many of the
in 42% yield.

(6)
In general, reaction rates were equivalent to, or reactions have been tested with only a small number
slightly lower than those observed in the organic of ionic liquids, where there are clearly many different
116 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117

combinations of cation and anion that could be tested. [11] J.A. Boon, J.A. Levisky, J.L. Pflug, J.S. Wilkes, J. Org. Chem.
Further information on the solvent properties of ionic 51 (1986) 480.
liquids is required, as are detailed mechanistic studies, [12] Y. Chauvin, B. Gilbert, I. Guibard, J. Chem. Soc., Chem.
Commun. (1990) 1715.
particularly where unexpected reactivity patterns are [13] R.T. Carlin, R.A. Osteryoung, J. Mol. Catal. 63 (1990) 125.
observed. [14] J.S. Wilkes, M.J. Zaworotko, J. Chem. Soc., Chem. Commun.
Ionic liquids will clearly not provide advantages in (1992) 965.
all systems, but improvements in reactivity or selectiv- [15] A.E. Visser, R.P. Swatloski, R.D. Rogers, Green Chem. 2
ity are observed in many of the cases discussed above (2000) 1.
[16] C.L. Hussey, in: G. Mamantov, A.I. Popov (Eds.),
when the appropriate combination of cation and an- Chemistry of Nonaqueous Solutions, VCH, Weinheim, 1994,
ion are selected. Despite this, the two factors that will pp. 227–276.
ultimately decide whether these systems are viable on [17] C.M. Gordon, A.J. McLean, Chem. Commun. (2000) 1395.
a larger scale are likely to be the ability to reuse the [18] C.M. Gordon, J.D. Holbrey, A.R. Kennedy, K.R. Seddon, J.
catalyst without a decrease in activity, and whether Mater. Chem. 8 (1998) 2627.
[19] T. Welton, Chem. Rev. 99 (1999) 2071.
the products can be separated efficiently without
[20] J.D. Holbrey, K.R. Seddon, Clean Prod. Processes 1 (1999)
contamination from the ionic liquid or catalyst. The 223.
combination of supercritical CO2 and ionic liquids [21] P. Wasserscheid, W. Keim, Angew. Chem., Int. Ed. 39 (2000)
looks like one very promising approach to attain this 3772.
goal. Finally, the recent work with enzymic catalysts [22] P.A.Z. Suarez, S. Einloft, J.E.L. Dullius, R.F. deSouza, J.
Dupont, J. Chim. Phys. 95 (1998) 1626.
in ionic liquids has presented an entirely new range of
[23] K.R. Seddon, A. Stark, M.-J. Torres, Pure Appl. Chem. 72
possibilities. While the success of transition metal and (2000) 2275.
Lewis acid catalysed processes in ionic liquids might [24] P. Bonhote, A.-P. Dias, N. Papageorgiou, K. Kalyanasun-
not seem too surprising, the very promising results daram, M. Grätzel, Inorg. Chem. 35 (1996) 1168.
gained to date in biocatalysis are possibly somewhat [25] A.J. Carmichael, K.R. Seddon, J. Phys. Org. Chem. 13 (2000)
591.
counterintuitive. It is hoped that developments like
[26] M.J. Muldoon, C.M. Gordon, I.R. Dunkin, J. Chem. Soc.,
these will attract further interest from scientists in Perkin Trans. 2 (2001) 433.
even more diverse disciplines into this exciting area. [27] S.N.V.K. Aki, J.F. Brennecke, A. Samanta, Chem. Commun.
(2001) 413.
[28] R.W. Soukup, K. Sone, Bull. Chem. Soc. Jpn. 60 (1987) 2286.
References [29] W. Linert, R.F. Jameson, A. Taha, J. Chem. Soc., Dalton
Trans. (1993) 3181.
[30] P. Wasserscheid, C.M. Gordon, C. Hilgers, M.J. Muldoon,
[1] P. Walden, Bull. Acad. Imper. Sci. (St. Petersburg) (1914) I.R. Dunkin, Chem. Commun. (2001) 1186.
1800. [31] L.A. Blanchard, D. Hancu, E.J. Beckman, J.F. Brennecke,
[2] R.M. Pagni, in: G. Mamantov, C.B. Mamantov, J. Braunstein Nature 399 (1999) 28.
(Eds.), Advances in Molten Salt Chemistry, Vol. 6, Elsevier, [32] L.A. Blanchard, J.F. Brennecke, Ind. Eng. Chem. Res. 40
New York, 1987, pp. 211–346. (2001) 287.
[3] C.G. Swain, A. Ohno, D.K. Roe, R. Brown, T. Maugh II, [33] L.A. Blanchard, Z. Gu, J.F. Brennecke, J. Phys. Chem. B
J. Am. Chem. Soc. 89 (1967) 2648. 105 (2001) 2437.
[4] F.H. Hurley, US Patent 2,446,331 (1948), Chem. Abstr. (1949) [34] S.G. Kazarian, B.J. Briscoe, T. Welton, Chem. Commun.
43, P7645b. (2000) 2047.
[5] F.H. Hurley, T.P. Weir Jr., J. Electrochem. Soc. 98 (1951) [35] R.J. Bonilla, B.R. James, P.G. Jessop, Chem. Commun. (2000)
207. 941.
[6] H.L. Chum, V.R. Koch, L.L. Miller, R.A. Osteryoung, J. Am. [36] R.A. Brown, P. Pollet, E. McKoon, C.A. Eckert, C.L. Liotta,
Chem. Soc. 97 (1975) 3264. P.G. Jessop, J. Am. Chem. Soc. 123 (2001) 1254.
[7] J. Robinson, R.A. Osteryoung, J. Am. Chem. Soc. 101 (1979) [37] F. Liu, M.B. Abrams, R.T. Baker, W. Tumas, Chem. Commun.
323. (2001) 433.
[8] J.S. Wilkes, J.A. Levisky, R.A. Wilson, C.L. Hussey, Inorg. [38] M.F. Sellin, P.B. Webb, D.J. Cole-Hamilton, Chem. Commun.
Chem. 21 (1982) 1263. (2001) 781.
[9] T.B. Scheffler, C.L. Hussey, K.R. Seddon, C.M. Kear, P.D. [39] D.E. Kaufmann, M. Nouroozian, H. Henze, Synth. Lett.
Armitage, Inorg. Chem. 22 (1983) 2099. (1996) 1091.
[10] D. Appleby, C.L. Hussey, K.R. Seddon, J.E. Turp, Nature [40] W.A. Herrmann, V.P.W. Böhm, J. Organomet. Chem. 572
323 (1986) 323 614. (1999) 141.
 C.M. Gordon / Applied Catalysis A: General 222 (2001) 101–117 117

[41] V.P.W. Böhm, W.A. Herrmann, Chem. Eur. J. 6 (2000) 1017. [62] C.J. Adams, M.J. Earle, K.R. Seddon, Chem. Commun.
[42] A.J. Carmichael, M.J. Earle, J.D. Holbrey, P.B. McCormac, (1998) 2097, and references therein.
K.R. Seddon, Org. Lett. 1 (1999) 997. [63] C.E. Song, W.H. Shim, E.J. Roh, J.H. Choi, Chem. Commun.
[43] J. Howarth, A. Dallas, Molecules 5 (2000) 851. (2000) 1695.
[44] L. Xu, W. Chen, J. Xiao, Organometallics 19 (2000) 1123. [64] F. Zulfiqar, T. Kitazume, Green Chem. 2 (2000) 296.
[45] V. Calo, A. Nacci, L. Lopez, N. Mannarini, J. Organomet. [65] J. Howarth, K. Hanlon, D. Fayne, P. McCormac, Tetrahedron
Chem. 41 (2000) 8973. Lett. 38 (1997) 3097.
[46] L. Xu, W. Chen, J. Ross, J. Xiao, Org. Lett. 3 (2001) 295. [66] M.J. Earle, P.B. McCormac, K.R. Seddon, Green Chem. 1
[47] C.J. Mathews, P.J. Smith, T. Welton, Chem. Commun. (2000) (1999) 23.
1249. [67] T. Fischer, A. Sethi, T. Welton, J. Woolf, Tetrahedron Lett.
[48] J. Sirieix, M. Ossberger, B. Betzemeier, P. Knochel, Synlett. 40 (1999) 793.
(2000) 1613. [68] C.W. Lee, Tetrahedron Lett. 40 (1999) 2461.
[49] S.T. Handy, X. Zhang, Org. Lett. 3 (2001) 233. [69] C.E. Song, W.H. Shim, E.J. Roh, S. Lee, J.H. Choi, Chem.
[50] W. Chen, L. Xu, C. Chatterton, J. Xiao, Chem. Commun. Commun. (2001) 1122.
(1999) 1247. [70] F. Zulfiqar, T. Kitazume, Green Chem. 2 (2000) 137.
[51] J. Ross, W. Chen, L. Xu, J. Xiao, Organometallics 20 (2001) [71] J.N. Rosa, C.A.M. Afonso, A.G. Santos, Tetrahedron 57
138. (2001) 4189.
[52] S. Toma, B. Gotov, I. Kmentová, E. Solcaniová, Green Chem. [72] M.F. Pinheiro, R.S. Mauler, R.F. de Souza, Macromol. Rapid
2 (2000) 149. Commun. 22 (2001) 425.
[53] E. Mizushima, T. Hayashi, M. Tanaka, Green Chem. 3 (2001) [73] C. DeCastro, E. Sauvage, M.H. Valkenburg, W.F. Hölderich,
76. J. Catal. 196 (2000) 86.
[54] J. Dupont, P.A.Z. Suarez, A.P. Umpierre, R.F. de Souza, J. [74] Y. Deng, F. Shi, S. Miao, J. Beng, K. Wang, J. Mol. Catal.
Braz. Chem. Soc. 11 (2000) 293. 165 (2001) 33.
[55] P. Wasserscheid, H, Waffenschmidt, P. Machnitzki, K.W. [75] P. Wasserscheid, H. Waffenschmidt, J. Mol. Catal. 164 (2000)
Kottsieper, O. Stelzer, Chem. Commun. (2001) 451. 61.
[56] C.E. Song, E.J. Roh, Chem. Commun. (2000) 837. [76] S.G. Cull, J.D. Holbrey, V. Vargas-Mora, K.R. Seddon, G.J.
[57] G.S. Owens, M.M. Abu-Omar, Chem. Commun. (2000) Lye, Biotechnol. Bioeng. 69 (2000) 227.
1165. [77] M. Erbeldinger, A.J. Mesiano, A.J. Russell, Biotechnol. Prog.
[58] C.E. Song, C.R. Oh, E.J. Roh, D.J. Choo, Chem. Commun. 16 (2000) 1129.
(2000) 1743. [78] R. Madeira Lau, F. van Rantwijk, K.R. Seddon, R.A. Sheldon,
[59] J. Howarth, Tetrahedron Lett. 41 (2000) 6627. Org. Lett. 26 (2000) 4189.
[60] R.A. Ligabue, J. Dupont, R.F. deSouza, J. Mol. Catal. A 169 [79] S.H. Schöfer, N. Kaftzik, P. Wasserscheid, U. Kragl, Chem.
(2001) 11. Commun. (2001) 425.
[61] A.J. Carmichael, D.M. Haddleton, S.A.F. Bon, K.R. Seddon, [80] K.-W. Kim, B. Song, M.-Y. Choi, M.-J. Kim, Org. Lett. 10
Chem. Commun. (2000) 1237. (2001) 1507.