investigation of the potential use of (iils) immobilized ionic liquids in shale gas sweetening

Refinery Process Stream Purification Refinery Process Catalysts Troubleshooting Refinery Process Catalyst Start-Up / Shutdown Activation Reduction In-situ Ex-situ Sulfiding Specializing in Refinery Process Catalyst Performance Evaluation Heat & Mass Balance Analysis Catalyst Remaining Life Determination Catalyst Deactivation Assessment Catalyst Performance Characterization Refining & Gas Processing & Petrochemical Industries Catalysts / Process Technology - Hydrogen Catalysts / Process Technology – Ammonia Catalyst Process Technology - Methanol Catalysts / process Technology – Petrochemicals Specializing in the Development & Commercialization of New Technology in the Refining & Petrochemical Industries

Web Site: www.GBHEnterprises.com

GBH Enterprises, Ltd.

Investigation of the Potential Use of (IILs) Immobilized Ionic Liquids in Shale Gas Sweetening

Case Study: #01521017GB/H Process Information Disclaimer

Information contained in this publication or as otherwise supplied to Users is believed to be accurate and correct at time of going to press, and is given in good faith, but it is for the User to satisfy itself of the suitability of the Product for its own particular purpose. GBHE gives no warranty as to the fitness of the Product for any particular purpose and any implied warranty or condition (statutory or otherwise) is excluded except to the extent that exclusion is prevented by law. GBHE accepts no liability for loss, damage or personnel injury caused or resulting from reliance on this information. Freedom under Patent, Copyright and Designs cannot be assumed.



The Potential Use of (IILs) Immobilized Ionic Liquids in Shale Gas Sweetening BACKGROUND: Ionic liquids are salts of nitrogen- or phosphorus-containing organic cations coupled with inorganic anions. Because of the asymmetry of the cations and size differences between them and the anions, they do not pack readily into a crystal, and consequently are liquids at room temperature. Typical ionic liquids are shown in the following Figure.

Due to the unique chemical physical properties of ionic liquids, they have been called, green solvents. Especially room temperature ionic liquids (RTILs), such as those based on N,Ndialkylimidazolium ions, are interesting solvents for catalytic reactions, for example:



Other examples include:

Ionic liquids are non-volatile and non-flammable, eliminating the hazards associated with volatile organic compounds (VOCs). In addition, the properties of ionic liquids may be tuned by varying the identities of the cations and anions, thereby tailoring the solvent to a specific application. The cations and anions can be varied to give different degrees of lipophilicity, and hence different solvating properties. They are usually air and water stable. They consist mainly of discrete dissociated ions or as strongly structured liquids, with the electric conductivity of salts. They have practically zero vapor pressures, which means that in terms of emissions they are ideal replacements for volatile organic solvents. The downside is that they themselves cannot be purified by distillation. The reaction products in ionic liquids can either be distilled off or extracted with water or alkanes. There appear to be no immediate industrial applications but BP is said in 2001 to be on the point of using ionic-liquid Friedel–Crafts alkylation technology to upgrade an existing process. The technology apparently applies to the Heck reaction:



The ionic liquids show excellent extraction capabilities and allow catalysts to be used in a biphasic system for convenient recycling. For example, the hydrovinylation of styrene with ethene can be carried out successfully using an ionic liquid and supercritical CO2 as solvent (Eq. 10-15). The ionic liquid dissolves the metal organic complex catalyst and sc-CO2 facilitates mass transfer and continuous processing.

IFP France has developed dimerization, hydrogenation, isomerization, and hydroformylation reactions without conventional solvents. For butene dimerization a commercial process exists. There is formed a biphasic system with the catalyst in the IL phase, which is immiscible with the reactants and products. This system can be extended to a number of organometallic catalysts. A variety of other reactions such as acylation of toluene, anisole, and chlorobenzene to give selectively p-isomer, alkylations, etc. have been conducted with ionic liquids.

Immobilized Ionic Liquids in Shale Gas Sweetening The oxidation of thiols to disulfides with cobalt II phthalocyanine catalyst in an ionic liquid at room temperature. The ionic liquid is there to solubilize the Co catalyst ( which is insoluble in all other solvents ) plus it enhances the rate of reaction. The reaction mechanism is said to go via the dimerization of RS radicals, formed via reaction of RS- with Co(I)-02. The question is what sort of benefit an immobilized ionic liquid could bring in this type of reaction ?



CONCEPT PROPOSAL: One possible approach would be to solubilize a Co(II) catalyst in the surface ionic liquid layer. This would create a "homogeneous catalyst" complex dissolved in a solid, supported ionic liquid. The Co(II) catalyst would be considered as bound within the ionic liquid layer. This overall concept would therefore be, a solid IIL catalyst containing the Co ( or other ) oxidation catalyst in a fixed bed process using air as the primary oxidant. In any sweetening reaction involving IILs there would be competing reactions to consider, such as oxidation of any sulfides to sulfones and strong adsorption of S species onto the IIL. RELATED: ExxonMobil and Lyngby University ( Topsoe ) have used this approach in hydroformylation where the active rhodium catalyst is dissolved in an immobilized ionic liquid and is reported to behave like a homogeneous system. A copy of the ExxonMobil paper is attached to explain this concept more clearly. For completeness I have summarized other uses of IILs in oxidation reactions below, and attached several papers for your convenience:

Recent work done by a European researcher on the use of IILs in selective oxidation reactions has concentrated on "difficult" alkane activation reactions such as the oxidation of cyclohexane to adipic acid. For this work he has developed a new class of water stable phosphonium based IILs using BF4-, PF6- and Cr as anions. The contribution of the IIL is its superacidity which activates the oxidant ( such as H2O2 or O2) and enables the reaction to proceed under relatively mild conditions of 80'C in a single phase containing cyclohexane/acetonitrile/30%aq. H202. Another European technology licensor is looking at the use of IILs in the oxidation of sulfur species in ultra low sulfur diesel. In this reaction they are again activating H202 and oxidizing, for example, sulfides to sulfones/sulfoxides.



REFERENCES: 1. Industrial Catalysis A Practical Approach Second, Completely Revised and Extended Edition, Prof. Dr. Jens Hagen University of Applied Sciences Mannheim Windeckstrasse , WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany 2. INDUSTRIAL ORGANIC CHEMICALS, Second Edition; HAROLD A. WITTCOFF Scientific Adviser, Nexant ChemSystems Inc.; Vice President of Corporate, Research, General Mills, Inc. (ret.) BRYAN G. REUBEN Professor Emeritus of Chemical Technology London South Bank University JEFFREY S. PLOTKIN Director, Process Evaluation and Research Planning Program, Nexant ChemSystems Inc. Published by John Wiley & Sons, Inc., Hoboken, New Jersey

ATTACHMENTS 1. Supported Ionic Liquid Catalysis – A New Concept for Homogeneous Catalysis Christian P. Mehnert,† Raymond A. Cook,† Nicholas C. Dispenziere,‡ Edmund J. Mozelski, ‡ and Mobae Afeworki† †ExxonMobil Research and Engineering Company, Corporate Strategic Research, and ‡ExxonMobil Chemical Company, Basic Chemicals & Intermediates Technology, 1545 Route 22 East, Annandale, NJ 08801, USA, [email protected]. 2. Supported ionic liquid phase (SILP) catalysts for hydroformylation A. Riisager1, K.M. Eriksen1, R. Fehrmann1, P. Wasserscheid2 1 Interdisciplinary Research Center for Catalysis (ICAT) and Department of Chemistry, Technical University of Denmark, DK-2800 Lyngby, Denmark. 2 Institut für Technische Chemie und Makromolekulare Chemie, RWTH, D-52074 Aachen, Germany. 3. Extractive Desulfurization and Denitrogenation of Fuels Using Ionic Liquids Shuguang Zhang, Qinglin Zhang,† and Z. Conrad Zhang* Akzo Nobel Chemicals Inc., 1 Livingstone Avenue, Dobbs Ferry, New York 10522



4. Hydroformylation of 1-hexene with rhodium in non-aqueous ionic liquids : how to design the solvent and the ligand to the reaction. Frédéric Favre, Hélène Olivier-Bourbigou,* Dominique Commereuc and Lucien Saussine Institut Français du pétrole, 1-4 Avenue de Bois-Préau, 92852 Rueil- Malmaison, France. E-mail: [email protected] Received (in Cambridge, UK) 10th May 2001, Accepted 7th June 2001 First published as an Advance Article on the web 6th July 2001 5. Catalytic reactions in ionic liquids, Roger Sheldon Laboratory of Organic Chemistry and Catalysis, Delft University of Technology, Julianalaan 136, Delft. BL-2628, The Netherlands. E-mail: [email protected]. Received (in Cambridge, UK) 10th August 2001, Accepted 11th September 2001. First published as an Advance Article on the web 18th October 2001

Supported Ionic Liquid Catalysis –

A New Concept for Homogeneous Catalysis

Christian P. Mehnert,† Raymond A. Cook,† Nicholas C. Dispenziere,‡ Edmund J. Mozelski, ‡ and Mobae Afeworki†

†ExxonMobil Research and Engineering Company, Corporate Strategic Research, and ‡ExxonMobil Chemical Company, Basic Chemicals & Intermediates Technology, 1545 Route 22 East, Annandale, NJ

08801, USA, [email protected]. Recently ionic liquids have attracted significant attention in the scientific literature. These ionic phases are salts that have melting points at ambient temperatures and can be utilized as liquid solvent media for a wide variety of chemical processes. Unlike conventional solvent systems these liquids exhibit low vapor pressure, tunable polarity, and high thermal stability. Depending on the choice of the ionic fragments a reaction environment can be designed to accommodate the catalysis and the separation of a chemical process in the most efficient way. By combining solid support materials with the advantages of ionic liquids we were successful in the preparation of surface-immobilized ionic liquid phases. In the new concept of supported ionic liquid catalysis (SILC) a homogeneous catalyst complex is dissolved in a multiple layer of an ionic liquid on a heterogeneous support (Fig. 1). This layer serves as the reaction phase in which a homogeneous catalyst is dissolved. Although the resulting material is a solid, the active species is dissolved in the immobilized ionic liquid phase and performs like a homogeneous catalyst. The presentation will focus on the investigation and characterization of supported ionic liquid catalysis for hydroformylation reactions. A comparison study will evaluate the differences of biphasic hydroformylation catalysis in ionic liquid media and conventional solvent systems. A high-pressure NMR study of hydroformylation reactions in ionic liquid systems will be presented to complement the catalyst evaluations.

Figure 1. Complex HRh(CO)(tppts)3 is immobilized in a multiple layer of ionic liquid on the

surface of a heterogeneous support material and performs like a homogeneous catalyst.

P

SO3Na

HRh(CO)

R RC

O

H

RhRh Rh

NNMe Bu

BF43 3

Support Material

Surface ImmobilizedIonic Liquid Complex

Supported Ionic Liquid PhaseCatalyst: Organic Phase

CO/H2

Supported ionic liquid phase (SILP) catalysts for hydroformylation A. Riisager1, K.M. Eriksen1, R. Fehrmann1, P. Wasserscheid2

1 Interdisciplinary Research Center for Catalysis (ICAT) and Department of Chemistry, Technical University of Denmark, DK-2800 Lyngby, Denmark. 2 Institut für Technische Chemie und Makromolekulare Chemie, RWTH, D-52074 Aachen, Germany. Introduction Room-temperature ionic liquids (IL's) receive significant interest due to their potential application as solvents in catalytic processes. So far, work has mainly involved improvement of current two-phase aqueous-organic catalytic systems by substituting water with IL as the solvent for immobilization of organometallic catalysts. This approach has been particular successful for Rh-catalyzed, liquid biphasic hydroformylation where several groups have reported good to excellent results for the conversion of C5-C8 and longer-chained olefins to the corresponding aldehydes. In most of these applications the IL [bmim][PF6] (bmim: 1-butyl-3-methyl-imidazolium) was used generating an effective medium for catalyst separation and recycling, provided the Rh-catalysts were modified with ligands containing charged groups. These charged ligands ensured a high relative catalyst affinity for the polar IL-phase compared to the lipophilic product phase.

Only minor attention has been given to catalysts made by "heterogenization" of ionic liquids on solid supports despite their potential for fixed bed gas phase reactions leading to advantageous industrial process design. To date only a few applications have been reported [1] where acidic chlorometallate ionic liquids on silica or coal were used as Lewis acid catalysts for Friedel-Craft (F-C) reactions. Moreover, tethered catalysts made of e.g. Lewis-acidic [bmim]-chloroaluminate complexes immobilized on supports via covalent anchoring have been used as catalytic active sites for F-C [2] and paraffin alkylations [3]. Only very recently a short communication has appeared on liquid biphasic hydroformylation [4] using an immobilized catalyst system composed of a supported ionic liquid phase. Results and Discussion Preliminary kinetic results have been obtained from continuous biphasic gas-liquid and liquid-liquid phase hydroformylation using novel [bmim][X]/silica (X = e.g. PF6 or octyl-SO4) supported ionic liquid phase (SILP) catalysts containing Rh-complexes of the charged ligands 1, 2 and 3. As an example the applicability of the SILP concept for continuous liquid biphasic hydroformylation of oct-1-ene using Rh-2 catalyst is here illustrated.

A monotonic increase in the activity was observed until steady-state was reached after 4 hours (Fig. 1) remaining unchanged for at least 3 additional hours (end of experiment). At steady-state a final n/iso ratio of 2.6 was reached which is in the anticipated range using a mono-dentate phosphine ligand. However, fluctuations were observed initially most likely due to a pre-activation reaction. ICP-AES analysis of outlet samples taken at steady-state demonstrated that Rh metal leaching was negligible (≤ 0.7%, detection limit).

In comparison to the conventional liquid-liquid biphasic catalysis in organic/IL mixtures the new supported ionic liquid phase (SILP) catalyst systems offer the significant advantage of very efficient IL use and short diffusion distances due to the highly dispersed ionic liquid catalyst solution. Due to these advantages we believe that other applications of organometallic catalysts using ILs as solvents could be reconsidered in form of SILP catalyst systems. References [1] M.H. Valkenberg, C. deCastro and W.F. Hölderich, Appl. Catal. A, 215 (2001) 185. [2] M.H. Valkenberg, C. deCastro and W.F. Hölderich, Green Chem., 4 (2002) 88 and cited

references; M.H. Valkenberg, E. Sauvage, C.P. deCastro-Moreira and W.F. Hölderich, WO 0132308 (2000) to ICI, UK.

[3] E. Benazzi, H. Olivier, Y. Chauvin, J.F. Joly and A. Hirschauer, Abst. Pap. Am. Chem. Soc., 212 (1996) 45; E. Benazzi, A. Hirschauer, J.F. Joly, H. Olivier and J.Y. Bernhard, EP 0553009 (1993) to IFP, France.

[4] C.P Mehnert, R.A. Cook, N.C. Dispenziere and M. Afeworki, J. Am. Chem. Soc., 124 (2002) 12932.

Fig. 1 Hydroformylation of oct-1-ene pre-saturated with CO/H2 gas using SILP Rh-2 catalyst. (CO:H2) = 1; p(CO/H2) = 14 atm; T = 115 °C; LHSV = 16 h-1. ●: TOF, ∆: n/iso ratio.

SEPARATIONS

Extractive Desulfurization and Denitrogenation of Fuels Using IonicLiquids

Shuguang Zhang, Qinglin Zhang,† and Z. Conrad Zhang*

Akzo Nobel Chemicals Inc., 1 Livingstone Avenue, Dobbs Ferry, New York 10522

Two types of ionic liquids, 1-alkyl-3-methylimidazolium [AMIM] tetrafluoroborate and hexafluo-rophosphate and trimethylamine hydrochloride (AlCl3-TMAC), were demonstrated to bepotentially applicable for sulfur removal from transportation fuels. EMIMBF4 (E ) ethyl),BMIMPF6 (B ) butyl), BMIMBF4, and heavier AMIMPF6 showed high selectivity, particularlytoward aromatic sulfur and nitrogen compounds, for extractive desulfurization and denitroge-nation. The used ionic liquids were readily regenerated either by distillation or by waterdisplacement of absorbed molecules. The absorbed aromatic S-containing compounds werequantitatively recovered. Organic compounds with higher aromatic π-electron density werefavorably absorbed. Alkyl substitution on the aromatic rings was found to significantly reducethe absorption capacity, as a result of a steric effect. The cation and anion structure and size inthe ionic liquids are important parameters affecting the absorption capacity for aromaticcompounds. At low concentrations, the N- and S-containing compounds were extracted fromfuels without mutual hindrance. AlCl3-TMAC ionic liquids were found to have remarkably highabsorption capacities for aromatics.

Introduction

Sulfur present in transportation fuels leads to SOxemission to air and inhibits the performance of pollutioncontrol equipment on vehicles. To minimize the negativehealth and environmental effects from automobile ex-hausts, increasing regulatory pressures are imposed onoil refineries to reduce the sulfur levels of the fuels,1-4

with the ultimate goal of zero emissions. While conven-tional hydroprocessing catalysts have been highly ef-fective for the reduction of sulfur levels, further im-provement of the hydrodesulfurization (HDS) efficiencyis limited to increasingly severe operating conditions atescalated cost. Not only does the energy consumptionbecome intensive, but also more severe conditionsrequired result in an increased hydrogen consumption,which causes undesired side reactions. When gasolineis desulfurized at higher pressure, many olefins aresaturated, resulting in lowered octane numbers. Highertemperature processing also leads to increased cokeformation and subsequent catalyst deactivation.5

The reactivity of organosulfur compounds over HDScatalysts depends on the molecular structures of S-containing compounds.6,7 The aliphatic organosulfurcompounds are very reactive in conventional hydrotreat-ing processes, and they can be completely removed fromthe fuels without much difficulty. The aromatic sulfurcompounds including thiophenes, benzothiophenes, andtheir alkylated derivatives, however, are generally more

difficult to convert over HDS catalysts. Therefore, thearomatic sulfur compounds present the most difficultchallenges to the HDS processes.

Alternative technologies are of particular interest inproviding potential solutions for sulfur-free clean fuels.For example, reactive adsorption8 and extraction withorganic solvents have been studied.9 The extractivedesulfurization (EDS) is an attractive alternative be-cause the process is applicable at ambient conditionswithout special equipment requirements. Besides thelow energy consumption, hydrogen consumption andhandling are also eliminated. In addition, the processdoes not change the chemical structure of the fuelcomponents. The organosulfur components can be re-covered at higher concentration following the extractionprocess if the solvents chosen for such a process can beregenerated. Therefore, the extractive solvents shouldbe sufficiently selective for absorption of sulfur com-pounds at high capacity without affecting the olefincontents. In addition, the solvents must be readilyregenerated following the extraction step.

Ionic liquids have been studied for applications re-lated to green chemical processes, such as liquid/liquidextractions, gas separations, electrochemistry, andcatalysis.10-18 Ionic liquids are typically nonvolatile,nonflammable, and thermally stable.19 In general, ionicliquids have higher density than organic liquids andwater. Therefore, many ionic liquids exist as a separatephase when in contact with organic and aqueous phases.These features make it possible to readily recycle theionic liquids for multiple extractions without additionalenvironmental concern. The ionic liquids based ontetrafluoroborate and hexafluorophosphate are knownto be moisture-insensitive. With short-chain 1-alkyl, the

* To whom correspondence should be addressed. Tel.:1 (914) 674 5034. Fax: 1 (914) 693 1782. E-mail:[email protected].

† Current address: Millennium Cell Inc., One IndustrialWay West, Eatontown, NJ 07724.

614 Ind. Eng. Chem. Res. 2004, 43, 614-622

10.1021/ie030561+ CCC: $27.50 © 2004 American Chemical SocietyPublished on Web 12/18/2003

former is water miscible and the latter is water im-miscible, even though a small amount of water (∼1%)can be dissolved in the latter. The melting points ofEMIMBF4 (1-ethyl-3-methylimidazolium tetrafluorobo-rate) and BMIMPF6 (1-butyl-3-methylimidazolium hexa-fluorophosphate) are both close to 5 °C. The BMIMBF4has a melting point of about -80 °C.19 As liquids at roomtemperature, these compounds are thermally stable upto about 300 °C, in the absence of strong acid. Forexample, Holbrey and Seddon’s thermogravimetric studyshowed that BMIMBF4 had a small weight loss of 3.5wt % between 280 and 320 °C when heated at 10 °C/min under nitrogen, and no further degradation wasobserved until 360 °C.20 Trimethylamine hydrochloride(TMAC) and AlCl3 based ionic liquids are easy toprepare and have low cost and low melting points.Although basic chloroaluminate molten salts at a nar-row AlCl3 percentage are in a liquid state at roomtemperature,12 they are less attractive to serve asextractive solvents because of their high viscosity. Thus,our focus is mainly on the EDS efficiency of acidicAlCl3-TMAC ionic liquids for transportation fuels.

In this study, we studied in detail the molecularabsorption properties of several water-sensitive ionicliquids (AlCl3-TMAC) and water-insensitive ionic liq-uids (AMIMBF4 and AMIMPF6) for various fuel com-ponents as well as for desulfurization efficiencies fromcommercial fuel samples. The absorption capacities ofwater-insensitive ionic liquids for N-containing com-pounds were also evaluated because these compoundsseverely inhibit the conversion efficiency of HDS cata-lysts, even at N compound concentrations below <30ppm.21-24 The possible effects of N-containing aromaticcompounds on the extractive removal of sulfur com-pounds were determined.

Experimental Section

Preparation of EMIMBF4. EMIMBF4 ionic liquidwas prepared by mixing equal moles of 1-ethyl-3-methyl-1H-imidazolium chloride (Aldrich) and lithiumtetrafluoroborate (Aldrich). Details of the preparationwere described elsewhere.25 The general structures ofthe ionic liquids used in this work with 1-alkyl-3-methylimidazolium (AMIM) cations are shown in Chart1.

Preparation of BMIMPF6. A BMIMPF6 ionic liquidwas prepared by mixing 1-butyl-3-methylimidazoliumchloride, [BMIM]Cl, and LiPF6 (Aldrich) in acetonitrilefollowed by filtration to remove a LiCl precipitate anddistillation to remove acetonitrile. [BMIM]Cl was ob-tained by refluxing equal molar amounts of 1-meth-ylimidazole and 1-chlorobutane in a flask while heatingand stirring at about 70 °C for 48 h.26 Another ionicliquid sample, prepared by the reaction of 1-butyl-3-methylimidazolium chloride and HPF6, was obtainedfrom University of Alabama.27

Preparation of BMIMBF4. A BMIMBF4 ionic liquidwas prepared by mixing 1-butyl-3-methylimidazoliumchloride and LiBF4 (Aldrich) in acetonitrile followed byfiltration to remove a LiCl precipitate and distillationto remove acetonitrile. One sample of BMIMPF6 and two

other ionic liquids, 1-methyl-3-octylimidazolium tet-rafluoroborate (MOIMBF4), and 1-hexyl-3-methylimi-dazolium hexafluorophosphate (HMIMPF6), were ob-tained from Aldrich Chemicals.

Preparation of Trimethylammonium Chloroalu-minate Ionic Liquids. Two acidic trimethylammo-nium chloroaluminate ionic liquids were prepared withAl-TMAC ratios of 1.5 and 2.0, respectively. The ratiowas carefully chosen at two levels so room temperatureionic liquid can be obtained with different acid strengthsfor the study.

(a) 2.0:1.0 AlCl3-TMAC Ionic Liquid. Aluminumtrichloride (2 mol) was added slowly to trimethyl-ammonium chloride salt (TMAC, 1 mol), both fromAldrich, in a glovebox under dry nitrogen. The reactionbetween the two solids was exothermic. A light-brown-ish liquid was formed. This liquid was stirred for 5 h.It has a density of 1.4-1.5 g/cm-3 at room temperature.The product was stable as a liquid at room temperatureunder a dry atmosphere. The formation of this type ofionic liquid is expressed in reaction (1).

(b) 1.5:1.0 AlCl3-TMAC Ionic Liquid. The sameprocedure as that described above for the preparationof 2.0:1.0 AlCl3-TMAC ionic liquid was followed, exceptthat a molar ratio of 1.5 for AlCl3-TMAC was used. Alight-yellow liquid was formed. This particular ionicliquid can be viewed as an equal mixture of (CH3)3NH+-Al2Cl7

- and (CH3)3NH+AlCl4-.

Measurement of the Absorption Capacity andS-Removal Efficiency. The model compounds 2-me-thylpentane, 1-hexene, methylcyclopentane, benzene,toluene, trimethylbenzene, thiophene, 2-methylthiophene,isobutyl mercaptan, dibenzothiophene (DBT), and 4,6-dimethyldibenzothiophene (DMDBT) were selected torepresent typical types of molecules in gasoline anddiesel fuels. A few model fuels with about 1000 ppmsulfur were prepared by dissolving a certain amount ofDBT or DMDBT in n-dodecane (n-C12). Model fuelscontaining nitrogen compounds were also made byadding pyridine or piperidine to n-C12 with or withoutsulfur-containing compounds.

The absorption capacity of an ionic liquid for a specificmodel compound was measured at room temperatureby adding an excess amount of the model compounddropwise to a glass vial containing the ionic liquid toform a two-phase system. After absorption equilibriumwas established, excess model compound in the upperphase was carefully separated with a pipet from theionic liquid phase. The amount of absorbed modelcompound in the ionic liquid phase was measured bythe weight gain. No detectable ionic liquid was foundin the upper phase.

The absorption selectivity for thiophene and toluenewas measured from a mixture containing both at nearlyequal concentration by weight. The concentrations ofabsorbed toluene and thiophene in the ionic liquid phasewere measured by NMR spectroscopy after the equilib-rium absorption.

In the present work, ionic liquids AlCl3-TMAC,EMIMBF4, BMIMBF4, MOIMBF4, BMIMPF6, andHMIMPF6 were also applied for sulfur removal fromtransportation fuels. Gasoline and diesel fuel sampleswith different S contents shown in Table 1 wereobtained from a commercial source. To make sure an

Chart 1. AMIMBF4 and AMIMPF6 Ionic Liquids

(CH3)3NHCl + 2AlCl3 f (CH3)3NH+Al2Cl7- (1)

Ind. Eng. Chem. Res., Vol. 43, No. 2, 2004 615

extraction equilibrium was reached, all single extrac-tions were conducted for 30 min with a weight ratio(organic phase/ionic liquid) of 5:1, except where other-wise indicated at room temperature. It has been verifiedthat shaking the absorption vial for 30 min is more thansufficient to establish the equilibrium. The organicphase was then separated from the ionic liquid phasefor analysis.

Analytical Procedures. Quantitative elementalanalysis of sulfur was conducted on a Bruker S4Explorer wavelength-dispersive X-ray fluorescence spec-trometer. Hydrocarbon and aromatics analyses werecarried out on an HP 6890 gas chromatograph (GC)/mass spectrometer (MS) using an OV-1 column (30 m× 0.32 mm i.d. × 5 µm). The GC/MS was also used toobtain S-compound and N-compound contents in themodel fuels containing both.

A Varian Inova 500 MHz NMR spectrometer and aJEOL GSX 270 MHz NMR spectrometer were used toverify the structures of the ionic liquids. NMR analyses(13C quantitative and attached proton test) were con-ducted to obtain the sulfur content in the ionic liquidphase after absorption. Some ionic liquids with absorbedcompound(s) were characterized using Fourier trans-form infrared (FTIR) and X-ray photoelectron spectro-scopic analyses.

Results

Sulfur-Compound Absorption Equilibrium. Toestablish the time needed to reach absorption equilib-rium, a model fuel containing 990 ppm S in the form ofDBT was treated with BMIMPF6. Sulfur analysis wasconducted on the fuels treated for 5, 10, 20, and 30 min.

As shown in Figure 1, the equilibrium was reachedafter 10 min of contact between the model fuel phaseand the ionic liquid phase.

Absorption Capacities for Model OrganosulfurCompounds. A previous work has established that theAMIMBF4 and AMIMPF6 ionic liquids, with the 1-alkylbeing ethyl and butyl, have negligible absorption foralkanes and very low absorption for olefins, in compari-son to aromatic compounds with and without sulfur.28

The absorption capacities of typical BF4-- and PF6

--based ionic liquids for the simple sulfur compounds arecompared in Figure 2.

As shown in Figure 2, for thiophene absorption,BMIMPF6 has the highest absorption capacity, followedby BMIMBF4. EMIMBF4 has the lowest absorptioncapacity among the three. NMR analysis of the absorbedthiophene in these ionic liquids indicated that thestructure and size of both anion and cation in the ionicliquids have a strong effect on their absorption capac-ity,29 consistent with the results of measurement based

Table 1. Sulfur Concentrations of Commercial Fuels

sampletotal sulfur

(ppm) sampletotal sulfur

(ppm)

low-sulfur gasoline 250 low-sulfur diesel 220high-sulfur gasoline 820 high-sulfur diesel 12122

Figure 1. Sulfur content versus contact time.

Figure 2. Absorption capacity for thiophenes and alkylthiol.

616 Ind. Eng. Chem. Res., Vol. 43, No. 2, 2004

on mass balance. Methyl substitution of thiophenedecreased the absorption capacity, similar to the alky-lated benzene molecules.28 The absorption for isobut-ylthiol is much lower because of the lack of aromaticity.

The absorption capacities of the AlCl3-TMAC ionicliquids are shown in Figure 3. The absorption of theionic liquids for the saturated hydrocarbons is negli-gible, and that for the olefin is very low. It is evidentthat these ionic liquids have remarkably high absorptioncapacity for the aromatic compounds and the absorptioncapacity is clearly influenced by the nature and stericeffects of the absorbed molecules. The observed sterichindrance is more significant for the alkylated thiophenethan the alkylated benzene. The absorption capacity of1.5:1.0 AlCl3-TMAC ionic liquid for toluene is about58% of that for benzene, and that for 2-methylthiopheneis only about 15% of that for thiophene. It appears that1.5:1.0 AlCl3-TMAC ionic liquid is more sensitive tosteric hindrance than 2.0:1.0 AlCl3-TMAC ionic liquid.

It was further observed that, upon addition ofthiophene or 2-methylthiophene into the AlCl3-TMACionic liquids, the color of the mixture changed im-mediately to brown and then gradually to dark brown.The ionic liquids interacted with 2-methyl-1-propaneth-iol mildly and led to the formation of a solid phase inthe ionic liquid phase and a colorless organic phase.

The mixing of aromatics with an ionic liquid resultedin the formation of a light-yellowish solution as anorganic phase. The saturated hydrocarbons did notinteract with the ionic liquid. Upon mixing, a clear two-phase solution was formed. The absorption capacity of2.0:1.0 AlCl3-TMAC ionic liquids for aromatics wasslightly higher than that of 1.5:1.0 AlCl3-TMAC (Figure3).

Absorption Efficiency for DBT and DMDBTfrom Model Fuels. The absorption efficiencies of the

three ionic liquids BMIMPF6, EMIMBF4, and BMIMBF4for DBT or DMDBT from two model fuels are comparedin Figure 4.

Like the case for thiophene absorption, EMIMBF4having the smallest cation among the three showedagain the least extractive removal of DBT and DMDBTfrom dodecane. On a weight basis, BMIMBF4 appearedto be slightly more effective than BMIMPF6 for extract-ing DBT or DMDBT from n-C12. However, on a molarbasis, the order is reversed, with a slight differencebetween the two. Interestingly, DMDBT was moredifficult to remove than DBT because of methyl substi-tution.

Competitive Absorption of Thiophene and Tolu-ene in the Ionic Liquids. The competitive absorptionselectivity of two model compounds, toluene and thio-phene, from their model mixture, was measured usingBMIMPF6 and BMIMBF4 ionic liquids. As shown inFigure 5, the ionic liquid BMIMPF6 showed remarkablyhigher absorption for both thiophene and toluene, ascompared to the BMIMBF4 ionic liquid. A total of about1.7 mol of thiophene and toluene was absorbed per mole

Figure 3. Absorption capacity of AlCl3-TMAC ionic liquids for model compounds.

Figure 4. Extraction of DBT/DMDBT from n-C12 using differentionic liquids.


of BMIMPF6 ionic liquid. The total absorbed thiopheneand toluene per mole of BMIMBF4 was about 0.7 mol.The thiophene-toluene ratio in BMIMPF6 was 1.6, andthat in BMIMBF4 was 2.1. Even though the totalabsorption capacity of BMIMBF4 was lower than thatof BMIMPF6, the former was more selective for thethiophene absorption than the latter.

Absorption of N-Containing Compounds. Severalmodel N compounds were studied to assess the ap-plicability of ionic liquids for the removal of organoni-trogen compounds. The model aromatic N compounds,pyridine and 2-methylpyridine (2-picoline), were foundto be completely miscible with BMIMBF4. The corre-sponding saturated compounds, piperidine and 2-me-thylpiperidine, showed limited absorption in the ionicliquid, as shown in Table 2. It is clear that absorptionof aromatic N compounds is favored.

Extractive Removal of Both Organonitrogenand Organosulfur Compounds. The absorption re-sults by the BMIMBF4 ionic liquid for the modelcompounds, DBT with either pyridine or piperidine fromdodecane, are given in Table 3.

When a model fuel consisted of n-C12 and anothercompound, either DBT, pyridine, or piperidine, the ionicliquid removed 12% S (DBT) in n-C12, 45% N (pyridine)in n-C12, and 9% N (piperidine) in n-C12. The amountremoved from each model fuel is much less than theabsorption capacity for the corresponding pure modelcompound by the ionic liquid, reflecting a partitioningof the model compounds in both the ionic liquid and thedodecane phases. The most effective extraction was the

removal of pyridine from dodecane. This is not surpris-ing because pure pyridine is fully miscible with the ionicliquid.

For the two model fuels, which contained eitherpyridine or piperidine and DBT, the most remarkableobservation is that the extraction of DBT and of theN-containing compounds by the ionic liquid was notmutually affected.

Single Extraction of Sulfur Compounds fromGasoline and Diesel Fuels. Table 4 shows experi-mental results of two AlCl3-TMAC ionic liquids for Sremoval from the high-sulfur gasoline sample.

About 20% sulfur removal from the high-sulfur gaso-line was achieved in a single contact with the 1.5:1.0AlCl3-TMAC ionic liquid at a gasoline to ionic liquidratio of 180 by weight. The 2.0:1.0 AlCl3-TMAC ionicliquid, which is more acidic, appears to be inferior tothe 1.5:1.0 AlCl3-TMAC ionic liquid for sulfur removalon the basis of equal weight of the ionic liquids.

The effect of the ratio of ionic liquid to fuel on theextraction efficiency is shown in Figure 6. The 1.5:1.0AlCl3-TMAC ionic liquid is effective for sulfur removalfrom diesels and high-sulfur gasoline at a low ionicliquid usage. An increase in the ratio of ionic liquid tofuel did not result in a proportional increase in Sremoval. Water and sulfur in diesel samples were allremoved by the AlCl3-TMAC ionic liquids. However,because chloroaluminate-based ionic liquid is moisture-sensitive, irreversible loss of ionic liquid was alsoobserved as a result of the formation of solid materialsupon absorption.

Ionic liquid treatments were found to remove almostcompletely the color of the gasoline samples, even at lowsulfur removal efficiency. On the other hand, the colorof the diesel samples after the treatments became moreintense.

Multicycle Extractions of Sulfur Compoundsfrom Gasoline. It has been established that the chainlength of the alkyl group in the 1-alkyl-3-methylimmi-dazolium cation has a pronounced effect on the absorp-tion of aromatic sulfur compounds,28 with increased

Figure 5. Competitive absorption of thiophene and toluene inBMIMPF6 and BMIMBF4.

Table 2. Absorption Capacities of BMIMBF4 forN-Containing Compoundsa

absorption capacity mol/mol of BMIMBF4

pyridine fully misciblepiperidine 0.722-picoline fully miscible2-methylpiperidine 0.23

a Based on 2 g of model compound in 1 g of ionic liquid.

Table 3. Absorption of S- and N-Containing Compounds by BMIMBF4

before treatment (ppm) after treatment (ppm) S or N removal (%)

model fuel S N S N S N

DBT in n-C12 747 660 12pyridine in n-C12 779 425 45DBT and pyridine in n-C12 764 779 660 460 14 41piperidine in n-C12 661 601 9DBT and piperidine in n-C12 764 723 677 658 11 9

Figure 6. Sulfur removal efficiency with 1.5:1.0 AlCl3-TMACionic liquid at room temperature.


absorption capacity at increased chain length. There-fore, MOIMBF4 and HMIMPF6, two ionic liquids withlong alkyl chains, were used for the extractive removalof sulfur compounds from a commercial gasoline samplecontaining about 820 ppm sulfur. Figure 7 shows thecontents of sulfur and aromatics after various cycles oftreatments. Using MOIMBF4, the first extraction re-duced the sulfur level to 700 ppm. Because the ionicliquid can be regenerated after equilibrium absorption(see below), multiple extraction cycles were applied forthis fuel sample. After 10 treatment cycles, a samplewith 320 ppm sulfur was obtained with a lighter colorthan that of the original fuel sample. At the same time,the aromatics concentration in the sample dropped from14.7 to 7.9 wt % after the 10 extraction cycles. TheHMIMPF6 liquid was slightly less effective on extrac-tion.

Multiple Extractions for Sulfur Removal fromDiesel Fuel. The two ionic liquids, MOIMBF4 andHMIMPF6, were also tested for sulfur removal from acommercial diesel fuel that contained 1.2 wt % sulfurand 7.9 wt % aromatics. With MOIMBF4, the sulfurlevel was lowered to 0.9 wt % and the aromatics contentto 3.8 wt % after 10 treatments. The ionic liquid becamebrown after use. The sulfur content in the diesel sampletreated with HMIMPF6 for 10 cycles was 1.0 wt %. Theused HMIMPF6 turned dark.

FTIR and XPS Characterization of AbsorbedModel Compounds. FTIR spectra (not shown) of the1.5:1.0 AlCl3-TMAC ionic liquid after benzene absorp-tion indicated a simple addition of the spectral featuresof benzene and ionic liquid that shows that benzene wassimply dissolved in the ionic liquid phase without astrong chemical interaction.

The absorption of thiophene and 2-methylthiophenein the ionic liquid produced a dark solid in the ionicliquid phase. FTIR study of the liquid products showedthe spectral feature of thiophene, with additional char-acteristic IR bands of alkylation products. The absorbed2-methylthiophene in the ionic liquid also showedalkylation relative to the starting material. Detailedcharacterization of the alkylation product is beyond thescope of this paper.

Elemental analysis by XPS of the dark solids formedthrough the thiophene-ionic liquid reaction revealedthe presence of sulfur in the solid phase and thatelements of N and Al are at a ratio corresponding tothe original ionic liquid formulation.

Regeneration of Used Ionic Liquid and Recov-ery of Absorbed Compounds. For a thiophene-saturated EMIMBF4 phase, the absorbed thiophene wasreleased into a separated phase upon addition of waterbecause of the fact that EMIMBF4 is miscible withwater. Water was then vaporized from the ionic liquidphase under a nitrogen flow at 110 °C for about 3 h.The EMIMBF4 ionic liquid was nearly quantitativelyrecovered.

Because BMIMPF6 has little miscibility with water,its regeneration was carried out by direct distillationafter saturated absorption of thiophene. The ionic liquidwas fully regenerated after heating at 110 °C for 3 hunder nitrogen. The absorbed thiophene recovered fromdistillation corresponds to the amount absorbed. NMRanalyses indicated that the ionic liquids, EMIMBF4 andBMIMPF6, maintained their original structures afterthe regenerations.

Regeneration of the used AlCl3-TMAC ionic liquidsby distillation was found to be infeasible because of theformation of the dark solids during their contact withgasoline and diesel. The sulfur-containing compoundscannot be removed from the used ionic liquid undervacuum at 80-110 °C for about 6 h. However, it wasfound that the solid material was soluble in methanoland acetone but only slightly soluble in carbon tetra-chloride.

Discussions

[AMIM]BF4 and [AMIM]PF6 Ionic Liquids. Therapid establishment of absorption equilibrium for theorganosulfur compounds between the fuel phase and theionic liquid phase, as shown in Figure 1, suggests thatsuch a process is rather simple to carry out. Becausethe AMIMBF4 and AMIMPF6 ionic liquids used in thiswork are insensitive to water and are liquid at ratherlow temperatures, the extraction is widely applicableat ambient conditions.

In general, the AMIMBF4 and AMIMPF6 ionic liquidshave a strong propensity for absorbing compounds witharomatic nature from the main aliphatic fuels. Theirabsorption capacities decrease as the aromatics becomehindered by the alkyl groups in the aromatic rings. Forexample, the absorption capacity decreases in the orderof benzene > toluene > xylene > cumene, as reportedin an earlier work.28 As shown in Figure 2, the methylgroup in 2-methylthiophene significantly lowered theabsorption capacity with respect to thiophene.

The strong affinity of the ionic liquids for the aromat-ics, organosulfur and organonitrogen compounds, isconceivably related to the high polarity of the ionicliquids. The aromatic molecules with highly delocalizedelectron density can be readily polarized through theirinteraction with the ionic liquids. It follows then thatlinear alkanes, cyclic alkanes, and olefins are barelyabsorbed. Similarly, alkylthiols and saturated alky-lamines are slightly absorbed, as observed through thiswork (Figures 2 and 3 and Table 2).

The structural features of ionic liquids play animportant role in the absorption of organic molecules,particularly for the aromatic compounds. For example,as shown in Figure 2, EMIMBF4 has the lowest absorp-

Figure 7. Sulfur and aromatics content in gasoline treated fordifferent cycles (aromatics: toluene, ethylbenzene, and xylene).

Table 4. Sulfur Removal from a High-Sulfur GasolineSample Using AlCl3-TMAC

sulfur aftertreatment (ppm)

sulfurremoval (%)

2.0:1.0 AlCl3-TMAC 700 151.5:1.0 AlCl3-TMAC 660 20


tion capacity for thiophene, while BMIMBF4 has alargely increased absorption capacity for it. With thesame BF4

- anion, BMIM is a larger cation than EMIMby extending the alkyl group chain length. The increaseof the cation size by the substitution of a longer alkylgroup to the imidazolium ring was responsible for thenearly 2-fold increase of the absorption capacity forthiophene. A similar phenomenon also holds withincreasing anion size. In this case, the effect of PF6

-

(with a diameter of 2.4 Å) as a larger anion is comparedto that of the BF4

- anion (with a diameter of 2.2 Å). Anincrease in the anion size led to an additional increasein the absorption capacity, as evidenced in Figure 2, bycomparing the thiophene absorption by BMIMBF4 andBMIMPF6 ionic liquids. The effect of the anion sizebecame most pronounced when BMIMPF6 and BMIMBF4were applied to a mixture of toluene and thiophene, asshown in Figure 5. The total absorbed toluene andthiophene in BMIMPF6 is about 2.4 times that inBMIMBF4. A similar trend is observed for the absorp-tion of 2-methylthiophene. The effects of cation andanion sizes on the interaction of absorbed thiophene andthe ionic liquids were further confirmed by NMRstudy.29

These results clearly point to the nature of theinteraction between absorbed aromatic compounds andthe ionic liquids. Molecules with highly polarizableπ-electron density preferably insert into the dynamicmolecular structure of the ionic liquids. The drivingforce for the molecular insertion is the favorable elec-tronic interaction of polarized aromatic molecules withthe charged ion pairs of ionic liquids. On the other hand,the insertion of aliphatic molecules of little polarizableelectronic structure would only weaken the Columbicinteraction of the ion pairs of the ionic liquids. There-fore, such an insertion is not favored. Alkyl substitutionon the aromatic ring may effectively disturb the molec-ular interaction in preferred orientation.

The higher thiophene-toluene ratio in BMIMBF4than in BMIMPF6 (as shown in Figure 5) could berationalized by the larger anion size in BMIMPF6.Because toluene is a larger molecule than thiophene,in addition to the steric hindrance from the methylgroup, it is relatively easier to be accommodated inBMIMPF6 than in BMIMBF4.

The effect of cations and anions of the ionic liquidsand the steric effect of absorbed compounds on theabsorption capacity were further displayed when thethree ionic liquids, EMIMBF4, BMIMPF6, and BMIMBF4,were applied to extract DBT and DMDBT from themixture with n-C12, as shown in Figure 4.

The absorption capacity measurements of BMIMPF6and BMIMBF4 for thiophene from a model mixture withtoluene (Figure 5) showed that the presence of toluenesignificantly reduced the absorbed amount of thiopheneas compared to a pure model thiophene compound(Figure 2). However, it is remarkable to note that theabsorption of toluene by the ionic liquids was onlyslightly reduced by the absorption of thiophene. Forexample, the absorption capacities for pure toluene byBMIMPF6 and BMIMBF4 are 0.78 and 0.3 mol/mol ofion liquid, respectively.25 In equilibrium with a modelmixture containing a nearly equal amount of tolueneand thiophene, the amount of absorbed toluene is 0.65mol/mol of BMIMPF6 and 0.23 mol/mol of BMIMBF4.It is known that the methyl group on the aromaticbenzene ring is an electron-donating group. Therefore,

the interaction of toluene and the ionic liquid would beexpected to be strong. It is likely that the absorbedtoluene inhibits the absorption of thiophene as a resultof steric hindrance of toluene. The relative strength ofthe interaction of toluene and thiophene with the ionicliquids remains to be determined.

When BMIMBF4 was applied for the extractive re-moval of DBT, pyridine, and piperidine from modelfuels, the results showed that DBT and the N com-pounds were independently absorbed into the ionicliquid without noticeable mutual hindrance. It is likely,in this case, that the ionic liquid was not saturated bythe S and N compounds because of the low concentrationof the model compounds in dodecane. Again, the absorp-tion by the ionic liquid for saturated alkylamine fromdodecane was very low. However, the extractive removalefficiency for aromatic nitrogen compounds, in this casepyridine, was remarkably high. The results indicatedthat the ionic liquids were particularly selective foraromatic N-containing compounds from fuels.

The water-insensitive ionic liquids were readily re-generated by two methods. One was by wetting thesaturated ionic liquids with water. Water as a smallmolecule with strong polarity has stronger interactionwith ionic liquids than polarized aromatic compounds.As a result, aromatic sulfur compounds were quantita-tively repelled from the ionic liquids. The absorbedwater can be removed to regenerate ionic liquids. Wateror other small polar molecules in fuels were shown tofavorably compete with organosulfur compounds forabsorption, leading to reduced absorption efficiency forthe organosulfur compounds. Another method of ionicliquid regeneration is by direct distillation. This methodis applicable for the removal of absorbed molecules attemperatures within the ionic liquid stability range.

Even though the absorptive removal of S compoundsis not very high in a single extraction because of theextremely low concentration of S compounds in thefuels, the feasibility for ionic liquid regeneration andreuse makes water-insensitive ionic liquids attractivefor processes involving multiple cycles. As shown inFigure 7, multiple-cycle extractive removal was dem-onstrated to be an effective process.

The results obtained on sulfur removal from gasolineand diesel samples suggest that the S removal fromdiesel is more difficult than that from gasoline. Thisobservation is likely related to a less favored equilibriumabsorption by ionic liquid in contact with the dieselphase of heavier molecules. The partitioning of aromaticcompounds in ionic liquids was reduced in the presenceof a heavy organic solvent, such as diesels.

The removal of saturated S-containing compoundsfrom fuels by ionic liquids is much less effective thanthe removal of aromatic S-containing compounds. Theresidual S compounds in the fuels after extractiveremoval by ionic liquids may mainly consist of saturatedones. As stated in the Introduction, conventional HDScatalysts are highly effective for the reduction of satu-rated organosulfur compounds. Therefore, the EDScould be a complementary process to the HDS.

AlCl3-TMAC Ionic Liquids. In the chloroaluminateionic liquids with a ratio of AlCl3-TMAC between 1 and2, the predominant Lewis acidic species present is wellestablished and known to be Al2Cl7

-. The Al3Cl10- is a

minor species,30 and AlCl4- as a neutral partner species

coexists through equilibrium (2) where Al2Cl7- is the


Lewis acid and Cl- is the Lewis base. Indeed, Al2Cl7-

was reported to catalyze alkylation and acylation reac-tions.18

In this study, [(CH3)3NH]+ is the cation, which is notexpected to largely change the nature of the anion. Infact, AlCl3-TMAC ionic liquids were found to effectivelycatalyze the alkylation of aromatics with olefins,12 whichcould explain our observation of the formation of alky-lation products when thiophene was in contact withAlCl3-TMAC ionic liquids.

Both 1.5:1.0 and 2.0:1.0 AlCl3-TMAC ionic liquidshave nearly equal (and the highest) absorption capacityfor thiophene. This indicates that the acidity changedoes not have a significant effect on the sulfur removalefficiency with these ionic liquids. The difference inamount absorbed by 2.0:1.0 and 1.5:1.0 AlCl3-TMACfor benzene is not as pronounced as that for alkylatedaromatics. Therefore, the more pronounced steric hin-drance for 1.5:1.0 AlCl3-TMAC ionic liquid could beattributed to the smaller AlCl4

- anion, which accountedfor half of the total chloroaluminate anions. The largerAl2Cl7

- anion appeared to be the cause of the reducedsteric effect for 2.0:1.0 AlCl3-TMAC ionic liquids.

The highest absorption capacity observed with PF6--

based ionic liquids for thiophene was 3.5, about half ofthe absorption capacity of AlCl3-TMAC ionic liquidsobserved in the present study.

A high capacity for sulfur removal from the diesel andhigh-sulfur gasoline (Figure 6) was achieved at a lowratio of ionic liquid to fuels. Therefore, multiple extrac-tions at a low ionic liquids-to-fuels ratio might be aviable approach. The low S-removal efficiency observedfor the low-S gasoline could be related to distributionof a low concentration of aromatic S compounds in thesample. For example, the ratio of sulfur concentrationto aromatics in low-S gasoline is about 5.8 au comparedto 424 au for high-S gasoline. Although AlCl3-based ionicliquids are effective for the removal of S-containingcompounds, contact between AlCl3-based ionic liquidsand thiol-containing compounds resulted in the forma-tion of dark precipitates. Thus, the application of AlCl3-based ionic liquids is limited to the absorption of certainaromatic compounds such as DBT.

Conclusions

Ionic liquids EMIMBF4, BMIMPF6, and BMIMBF4and other heavier AMIMPF6 ones showed remarkableselectivity for the absorption of aromatics and aromaticS- and N-containing molecules from transportationfuels. These ionic liquids are moisture-insensitive,thermally stable under the distillation conditions, andreadily regenerated for reuse. The absorbed aromaticS-containing compounds were quantitatively recoveredduring the regeneration. The Lewis acidic AlCl3-TMACionic liquids were found to have remarkably highabsorption capacities for aromatics, particularly sulfur-containing aromatic compounds, but their regenerationis problematic. The results suggest that compounds withhigher aromatic π-electron density are favorably ab-sorbed. A methyl group on the aromatic rings was foundto significantly reduce the absorption capacity, possiblybecause of a steric effect. The cation and anion structureand size in the ionic liquids are important parametersaffecting the absorption capacity for aromatic com-

pounds. At low concentrations, the N- and S-containingcompounds were extracted from fuels without mutualhindrance.

Acknowledgment

We thank the Akzo Nobel Multi-BU Program forproviding financial support. Analytical support given byDr. Gary Darsey and Evan Chen for elemental sulfuranalysis and by Dr. Biing-Ming Su for NMR analysis isgreatly appreciated.

Literature Cited

(1) Min, W. A Unique Way to Make Ultra Low Sulfur Diesel.Korean J. Chem. Eng. 2002, 19, 601.

(2) http://www.epa.gov/otaq/tr2home.html.(3) http://www.epa.gov/otaq/tr2home.html.(4) http://www.dieselnet.com/standards/eu/ld.html.(5) Babich, I. V.; Moulijn, J. A. Science and technology of novel

processes for deep desulfurization of oil refinery streams: a review.Fuel 2003, 82, 607.

(6) Ma, X.; Sakanishi, K.; Mochida, I. Hydrodesulfurizationreactivities of various sulfur compounds in diesel fuel. Ind. Eng.Chem. Res. 1994, 33, 218.

(7) Meille, V.; Schulz, E.; Lemaire, M.; Vrinat, M. Hydrodes-ulfurization of Alkyldibenzothiophenes over a NiMo/Al2O3Cata-lyst: Kinetics and Mechanism. J. Catal. 1997, 170, 29.

(8) Meier, P. F.; Reed, L. E.; Greenwood, G. J. RemovingGasoline Sulfur. Hydrocarbon Eng. 2001, 1, 26.

(9) Funakoshi, I.; Aida, T. Process for recovering organic sulfurcompounds from fuel oil. U.S. Patent 5,753,102, 1998.

(10) Howard, K. A.; Mitchell, H. L.; Waghore, R. H. Liquid saltextraction of aromatics from process feed streams. U.S. Patent4,359,596, 1982.

(11) Boate, D. R.; Zaworotko, M. J. Organic Non-quaternaryClathrate Salts for Petroleum Separation. U.S. Patent 5,220,106,1993.

(12) Sherif, F. G.; Shyyu, L.; Greco, C. C. Linear AlkylbenzeneFormation Using Low Temperature Ionic Liquid. U.S. Patent5,824,832, 1998.

(13) Koch, V. R.; Nanjundiah, C.; Carlin, R. T. HydrophobicIonic Liquids. U.S. Patent 5,827,602, 1998.

(14) Silvu, S. M.; Suarcz, P. A. Z.; de Souza, R. F.; Doupont, J.Selective Linear Dimerization of 1,3-Butadiene by PalladiumCompounds Immobilized into 1-n-Butyl-3-methyl ImidazoliumIonic Liquids. Polym. Bull. 1998, 40, 401-405.

(15) Carmichael, A. J.; Haddletton, D. M.; Bon, S. A. F.; Seddon,K. R. Copper(I) Mediated Living Radical Polymerisation in an IonicLiquid. Chem. Commun. 2000, 1237-1238.

(16) Carlin, R. T.; Wilkes, J. S. Complexation of cp2MCl2 in aChloroaluminate Molten-salt-relevance to Homogeneous Ziegler-Natta Catalysis. J. Mol. Catal. 1990, 63, 125-129.

(17) Goledzinowski, M.; Birss, V. I.; Galuszka, J. Oligomeriza-tion of Low-molecular-weight Olefins in Ambient TemperatureMolten Salts. Ind. Eng. Chem. Res. 1993, 32, 1795-1797.

(18) Boon, J. A.; Levisky, J. A.; Pflug, J. L.; Wilkes, J. S.Friedel-Crafts Reactions in Ambient-temperature Molten Salts.J. Org. Chem. 1986, 51, 480-483.

(19) http://bama.ua.edu/∼rdrogers/.(20) Holbrey, J. D.; Seddon, K. R. The Phase Behavior of

1-Alkyl-3-methylimidazolium Tetrafluoroborates; Ionic Liquid andIonic Liquid Crystals. J. Chem. Soc., Dalton Trans. 1999, 2133-2139.

(21) Laredo, G. C.; Altamirano, E.; De los Reyes, J. A. InhibitionEffects of Nitrogen Compounds on the Hydrodesulfurization ofDibenzothiophene: Part 2. Appl. Catal. A 2003, 243 (2), 207-214.

(22) Ho, T. C. Property-reactivity Correlation for HDS ofMiddle Distillates. Appl. Catal. A 2003, 244 (1), 115.

(23) van Looij, F.; van der Laan, F.; Stork, W. H. J.; DiCamillo,D. J.; Swain, J. Key Parameters in Deep Hydrodesulfurization ofDiesel Fuel. Appl. Catal. A 1998, 170, 1.

(24) Egorova, M.; Prins, R. I. Effect of N-Containing Moleculeson the Hydrodesulfurisation of Dibenzothiophene. Prepr. Symp.sAm. Chem. Soc., Div. Fuel Chem. 2002, 47 (2), 445.

2AlCl4- S Al2Cl7

- + Cl- (2)


(25) Zhang, S.; Zhang, Z. C. Selective Sulfur Removal fromFuels Using Ionic Liquids at Room Temperature. Prepr. Symp.sAm. Chem. Soc., Div. Fuel Chem. 2002, 47 (2), 449.

(26) Visser, A. E.; Swatloski, R. P.; Rogers, R. D. pH-DependentPartitioning in Room-Temperature Ionic Liquids Provides A Linkto Traditional Solvent Extraction Behavior. Green Chem. 2000, 2(Feb), 1.

(27) Bosmann, A.; Datsevich, L.; Jess, A.; Lauter, A.; Schmitz,C.; Wasserscheld, P. Deep Desulfurization of Diesel Fuel byExtraction with Ionic Liquids. Chem. Commun. 2001, 2494.

(28) Zhang, S.; Zhang, Z. C. Novel Properties of Ionic Liquidsin Selective Sulfur Removal from Fuels at Room Temperature.Green Chem. 2002, 4, 376-379.

(29) Su, B.; Zhang, S.; Zhang, Z. C. Structural Elucidation ofThiophene Interaction with Ionic Liquids by Multinuclear NMRSpectroscopy. J. Phys. Chem. 2003, submitted for publication.

(30) Wilkes, J. S.; Levisky, J. A.; Wilson, R. A.; Hussey, C. L.Dialkylimidazolium Chloroaluminate Melts: A New Classof Room-Temperature Ionic Liquids for Electrochemistry,Spectroscopy, and Synthesis. Inorg. Chem. 1982, 21, 1263-1264.

Received for review July 7, 2003Revised manuscript received November 3, 2003

Accepted November 5, 2003

IE030561+


Com

munication

ww

w.rsc.org/chem

comm

CH

EMCO

MM

Hydroformylation of 1-hexene with rhodium in non-aqueous ionicliquids : how to design the solvent and the ligand to the reaction

Frédéric Favre, Hélène Olivier-Bourbigou,* Dominique Commereuc and Lucien Saussine

Institut Français du pétrole, 1-4 Avenue de Bois-Préau, 92852 Rueil-Malmaison, France.E-mail: [email protected]

Received (in Cambridge, UK) 10th May 2001, Accepted 7th June 2001First published as an Advance Article on the web 6th July 2001

A wide range of ionic liquids based on imidazolium andpyrrolidinium cations and weakly coordinating anionsproved to be efficient solvents for the biphasic rhodiumcatalyzed hydroformylation of 1-hexene; the reaction rateand regioselectivity, and the retention of the rhodium can beoptimized by fitting the nature of the anions and cations ofthe ionic liquid and the modified phosphite or phosphineligands.

Ionic liquids are good solvents for transition-metal complexesin many homogeneously catalyzed reactions, e.g. olefin hydro-genation, hydroformylation, oligomerization and Pd mediatedcarbon–carbon coupling reactions.1 In many cases the reactionproducts are very weakly soluble in the ionic phase so that thecatalyst can be separated by simple decantation and recycled.The aqueous two-phase catalysis concept, which has beenalready applied industrially for propene hydroformylation,2 canthen be extended to substrates and ligands that are poorlysoluble or non stable in water. Higher olefin Rh-hydro-formylation has been performed using different 1-butyl-3-methylimidazolium room-temperature liquid salts as sol-vents, in the presence of phosphine ligands. The main difficultyis to immobilize the rhodium catalyst in the ionic liquid phasewhile maintaining its activity and selectivity. A solution is tomodify the neutral phosphine ligands with ionic groups.3Thanks to their chemical and physical versatility,4 ionic liquidscan be specially designed to fit with the ligand and the operatingconditions that provide the best performances in catalysis.

In this communication, for the first time we report the effectof the nature of the cations and anions of the ionic liquids on theRh-catalyzed hydroformylation of 1-hexene. We also provideour preliminary study on the performances of differentphosphorus ligand–ionic liquid systems.

We have prepared a wide range of ionic liquids by varying thenature of the cation e.g. 1,3-dialkylimidazolium, 1,2,3-trialk-ylimidazolium and N,N-dialkylpyrrolidinium and the nature ofthe anion e.g. BF4

2, PF62, CF3CO2

2, CF3SO32 (OTf2) and

N(CF3SO2)22 (NTf2

2). The BF42, NTf2

2 and PF62 ionic

liquids were prepared by anion exchange starting fromimidazolium or pyrrolidinium chloride. The CF3SO3

2 andCF3CO2

2 salts were prepared by direct methylation of 1-alkyli-midazole or 1-alkylpyrrolidine with the corresponding methylesters.5 We have measured the solubility of 1-hexene in theseionic liquids (Fig. 1). For a given anion, e.g. CF3CO2

2, thesolubility of 1-hexene increases upon increasing the length ofthe alkyl chain of the 1,3-dialkylimidazolium e.g. 1-butyl-3-methylimidazolium (BMI+) vs. 1-hexyl-3-methylimidazo-lium (HMI+). Methylation of the C(2) atom of the imidazoliumring tends to decrease the solubility of 1-hexene e.g. 1-butyl-2,3-dimethylimidazolium BDMI+NTf2

2 vs. BMI+NTf22. No

significant differences are observed by changing the 1-butyl-3-methylimidazolium cation for N,N-butylmethylpyrrolidinium(BMP+). For a same cation, e.g. BMI+, the solubility of1-hexene increases as follows: BF4

2 < PF62 < OTf2 <

CF3CO22 < NTf2

2.In a first series of experiments, we performed 1-hexene

hydroformylation using these different ionic liquids as solvents

for the Rh(CO)2(acac) precursor associated with the sodium saltof monosulfonated triphenylphosphine (TPPMS) (Fig. 1). Theresults reveal that there is a correlation between the reactionrates (TOF min21) and the solubility of 1-hexene in ionic

Fig. 1 Turnover frequency as a function of 1-hexene solubility in the ionicliquids. Reaction conditions: Rh(CO)2(acac) 0.075 mmol, 1-hexene/Rh =800, TPPMS/Rh = 4 , heptane was used as internal standard, CO/H2 (molarratio) = 1, P(CO/H2) = 2 MPa, T = 80 °C, TOF determined at 25%conversion of 1-hexene.

Scheme 1

Scheme 2 Synthesis of ligand 2. Reagents and conditions: i, KOBut, refluxfor 3 h in THF; ii, Me3O+BF4

+ in CH2Cl2, 278 °C.

This journal is © The Royal Society of Chemistry 2001

1360 Chem. Commun., 2001, 1360–1361 DOI: 10.1039/b104155j

liquids. In all cases, the selectivity in aldehydes is > 97%, theremainder being isomerized hexenes. Surprisingly, in theNTf2

2 based ionic liquids, lower TOF are obtained despite therelatively good solubility of 1-hexene in these media. Assuggested by the determination of ionic liquid relative polarity,6the NTf2

2 based salts could be more coordinating than the PF62

salts. In all cases, the n/i ratio is not affected by the nature of thesolvent.

In a second series of experiments, we have synthesized themonosubstituted guanidinium triphenylphosphine ligands 1aand 1b,7 and the pyridinium diphenylethylphosphine ligand 2(Scheme 1). Ligand 2 was prepared in two steps according toScheme 2. Ligands 1a and 1b show good solubility in the ionicliquid BMI+BF4

2 . They give good selectivities towards thelinear aldehydes. Similar catalytic performances were obtainedby using 1a or 1b (Table 1, entries 1 and 2). However, theretention of the Rh in the BMI+BF4

2 phase was more efficientwith 1a (the Rh content in the organic phase was lower than thedetection limit according to ICP analysis for 1a, while the levelwas 0.8% of the initial Rh for 1b). Ligand 2 (entry 3) presentshigher reaction rates and higher selectivity towards aldehydesthan 1a and 1b. However, the leaching of the Rh in the organicphase was found to be higher for 2 (2% of the initial Rh).

Phosphites and bisphosphites are well known ligands for Rh-hydroformylation to afford higher reaction rates.8 Because oftheir instability toward hydrolysis, examples of their use inaqueous two-phase hydroformylations are rare.9 Ionic liquidsoffer suitable alternative solvents. We describe here the first useof phosphite based ligands for the biphasic hydroformylation of1-hexene in ionic liquids. Ligand 3, a mixture of tetra-butylammonium salt of the mono- di- and tri-sulfonatedtriphenylphosphites, has been prepared by transesterification oftriphenylphosphite with the tetrabutylammonium salt of p-hydroxyphenylsulfonic acid.9 In the reaction with the ligand 3,using BMI+PF6

2 as the solvent, good catalytic activity isobserved (entry 4). The selectivity for the linear aldehyde ismuch higher than the selectivity obtained with phosphineligands (enties 1–3). The use of the modified phosphite 3 limitsthe loss of the Rh in the organic phase (leaching is 2% of theinitial Rh used). At the end of the run, the organic phase isdecanted and separated from the ionic liquid which is reused(entry 5 and 6). Despite a loss of activity which could be

ascribed to a partial degradation of the Rh active catalyst duringthe separation, the n/i ratio remains high after two recyclings.

In conclusion, it is shown that thanks to the great versatilityof ionic liquids, it is possible to optimize Rh-hydroformylationperformances by adjusting the nature of the anions and cationspresent in the solvent and the nature of the ligands. Phosphiteligands, which are unstable in an aqueous two-phase system,can be used. The problem of Rh leaching can be minimized bythe modification of phosphorus ligands with cationic (guanidin-ium or pyridinium) or anionic (sufonate) groups. By adjustingthe ligand and the ions of the solvent, excellent Rh retention hasbeen achieved.

Notes and references1 For a review see: H. Olivier, in Aqueous-Phase Organometallic

Catalysis, ed. B. Cornils and W.A. Herrmann, Wiley-VCH, Weinheim,1998, p. 553. T. Welton, Chem. Rev., 1999, 99, 2071; P. Wasserscheidand W. Keim, Angew. Chem., Int. Ed., 2000, 39, 3772; J. Dupont, C. S.Consorti and J. Spencer, J. Braz. Chem., 2000, 11, 337; J. D. Holbrey andK. R. Seddon, Clean Prod. Processes, 1999, 1, 223.

2 E. G. Kuntz, Fr. Pat., 2314910, 1975 (to Rhône-Poulenc); B. Cornils andE. Wiebus, CHEMTECH, 1995, 25listlistr, 33.

3 Y. Chauvin, L. Mussmann and H. Olivier, Angew. Chem., Int. Ed., 1995,34, 2698; C. C. Brasse, U. Englert, A. Salzer, H. Waffenschmidt and P.Wasserscheid, Organometallics, 2000, 19, 3818; P. Wasserscheid, H.Waffenschmidt, P. Machnitzki, K. W. Kottsieper and O. Stelzer, Chem.Commun., 2001, 451.

4 A. J. Carmichael, C. Hardacre, J. D. Holbrey, K. R. Seddon and M.Nieuwenhuyzen, Electrochem. Soc. Proceedings, Molten Salts XII, ed.P. C. Trulove, H. C. De Long, G. R. Stafford and S. Deki, TheElectrochemical Society, Pennington, NJ, 2000, vol. 91-41, p. 209.

5 P. Bonhôte, A. Dias, N. Papageorgiou, K. Kalyanasundaram and M.Grätzel, Inorg. Chem., 1996, 35, 1168.

6 A. J. Carmichael and K. Seddon, J. Phys. Org. Chem., 2000, 13, 591.7 A. Hessler, O. Stelzer, H. Dibowski, K. Worm and F. P. Schmidtchen,

J. Org. Chem., 1997, 62, 2362; P. Machnitzki, M. Tepper, K. Wenz, O.Stelzer and E. Herdtweck, J. Organomet. Chem., 2000, 602, 158.

8 P. C. J. Kamer, J. N. H. Reek and P. W. N. M. van Leeuwen, in RhodiumCatalyzed Hydroformylation, ed. P.W.N.M. van Leeuwen and C. Claver,Kluwer Academic Publishers, Netherlands, 2000, p. 35.

9 B. Fell, G. Papadogianakis, W. Konkol, J. Weber and H. Bahrmann,J. Prakt. Chem., 1993, 335, 75.

Table 1 Hydroformylation with different ligand–ionic liquid systemsa

Entry Ligand L L/Rh Ionic liquidReactiontime/min

Conversionb

(%)Aldehydesc

(mol %) n/id TOFe/min21

1 1a 10 BMI+BF42 180 77 74 3.7 3

2 1b 7 210 83 78 4 33 2 4 180 87 96 2.6 44 3 9.5 BMI+PF6

2 180 96 88 12.6 45f 3 240 85 89 11.2 26g 3 330 42 88 11.7 1

a Reaction conditions: Rh(CO)2(acac) 0.075 mmol, 1-hexene/Rh = 800, CO/H2 (molar ratio) = 1 ; P(CO/H2) = 2 MPa, T = 80 °C, heptane (internalstandard) = 2 mL, 1-hexene = 7.5 mL, ionic liquid = 4 mL. b Conversion = [(initial 1-hexene) 2 (1-hexene after reaction)]/(initial 1-hexene). c The otherproducts are 2- and 3-hexenes. d Linear to branched aldehyde ratio. e Mol of aldehydes per mol Rh per minute at 25% conversion. f Recycling of 4. g Recyclingof 5.

Chem. Commun., 2001, 1360–1361 1361

Feature Article

ww

w.rsc.org/chem

comm

CH

EMCO

MM

Catalytic reactions in ionic liquids

Roger Sheldon

Laboratory of Organic Chemistry and Catalysis, Delft University of Technology, Julianalaan 136, DelftBL-2628, The Netherlands. E-mail: [email protected]

Received (in Cambridge, UK) 10th August 2001, Accepted 11th September 2001First published as an Advance Article on the web 18th October 2001

The chemical industry is under considerable pressure to replacemany of the volatile organic compounds (VOCs) that arecurrently used as solvents in organic synthesis. The toxic and/orhazardous properties of many solvents, notably chlorinatedhydrocarbons, combined with serious environmental issues,such as atmospheric emissions and contamination of aqueouseffluents is making their use prohibitive. This is an importantdriving force in the quest for novel reaction media. Curzons andcoworkers,1 for example, recently noted that rigorous manage-ment of solvent use is likely to result in the greatest improve-ment towards greener processes for the manufacture ofpharmaceutical intermediates. The current emphasis on novelreaction media is also motivated by the need for efficientmethods for recycling homogeneous catalysts. The key to wasteminimisation in chemicals manufacture is the widespreadsubstitution of classical ‘stoichiometric’ syntheses by atomefficient, catalytic alternatives.2 In the context of homogeneouscatalysis, efficient recycling of the catalyst is a conditio sine quanon for economically and environmentally attractive processes.Motivated by one or both of the above issues much attention hasbeen devoted to homogeneous catalysis in aqueous biphasic3,4

and fluorous biphasic5 systems as well as in supercritical carbondioxide.6 Similarly, the use of ionic liquids as novel reactionmedia may offer a convenient solution to both the solventemission and the catalyst recycling problem.7,8

Historical developmentWhat are ionic liquids? Quite simply, they are liquids that arecomposed entirely of ions. Molten sodium chloride, forexample, is an ionic liquid but a solution of sodium chloride inwater is an ionic solution. The term molten salts, which was

previously used to describe such materials, evokes an image ofhigh-temperature, viscous and highly corrosive media. The termionic liquid, in contrast, implies a material that is fluid at (orclose to) ambient temperature, is colourless, has a low viscosityand is easily handled, i.e. a material with attractive propertiesfor a solvent. Room temperature ionic liquids are generally saltsof organic cations, e.g. tetraalkylammonium, tetraalkylphos-phonium, N-alkylpyridinium, 1,3-dialkylimidazolium andtrialkylsulfonium cations (Fig. 1).

In order to be liquid at room temperature, the cation shouldpreferably be unsymmetrical, e.g. R1 and R2 should be differentalkyl groups in the dialkylimidazolium cation. The meltingpoint is also influenced by the nature of the anion (seeTable 1).

Room temperature ionic liquids are not new. Ethylammon-ium nitrate, which is liquid at room temperature (but usuallycontains 200–600 ppm water) was first described in 1914.9 Inthe late 1940s, N-alkylpyridinium chloroaluminates werestudied as electrolytes for electroplating aluminium. Thesesystems were reanimated by the groups of Hussey,10 Oster-young11 and Wilkes12 in the late 1970s. The first examples ofionic liquids based on dialkylimidazolium cations were reportedin the early 1980s by Wilkes and coworkers.12 They contained

Roger Sheldon was born in Nottingham (UK) in 1942. Afterreceiving a PhD in Organic Chemistry from Leicester Uni-versity (1967) he spent two years as a postdoc with ProfessorJay Kochi in the USA. From 1969–1980 he was with ShellResearch in Amsterdam and from 1980–1990 he was R&DDirector of DSM Andeno. In 1991 he moved to his preentposition as Professor of Organic Chemistry and Catalysis at theDelft University of Technology. His research interests arefocused on the application of catalytic methodologies—homo-geneous, heterogeneous and enzymatic—in organic synthesis,particularly in relation to fine chemicals production. He haswidely promoted the concepts of E factors and atom efficiencyfor assessing the environmental impact of chemical processes.He is the author of ca. 300 scientific publications, numerouspatents and three books on the subject of catalysis andchirotechnology. He is the Editor-in-Chief of Journal ofMolecular Catalysis B: Enzymatic and Chairman of theEditorial Board of Green Chemistry. Among other distinctionshe was recently awarded a Doctor Honoris Causa from theRussian Academy of Sciences.

Fig. 1 Structures of ionic liquids.

Table 1 Melting points of some dialkylimidazolium salts

R X mp/°C

Me Cl 125Et Cl 87n-Bu Cl 65Et NO3 38Et AlCl4 7Et BF4 6Et CF3SO3 29Et (CF3SO3)2N 23Et CF3CO2 214n-Bu CF3SO3 16

This journal is © The Royal Society of Chemistry 2001

DOI: 10.1039/b107270f Chem. Commun., 2001, 2399–2407 2399

chloroaluminate anions (AlCl42 or Al2Cl72) and proved to beuseful catalysts/solvents for Friedel–Crafts acylations.13 How-ever, a serious obstacle for widespread use of these ionic liquidsis the high reactivity of the chloroaluminate anion towardswater.

The first example of the new ionic liquids, that currently arereceiving so much attention as novel media for homogeneouscatalysis, ethylmethylimidazolium tetrafluoroborate(emimBF4)† was reported by Wilkes et al. in 1992.14 Thesynthesis of the corresponding hexafluorophosphate followedshortly thereafter.15 In contrast to the chloroaluminate salts thefluoroborates and hexafluorophosphates are stable towardshydrolysis. Subsequently, 1,3-dialkylimidazolium salts con-taining a wide variety of anions, e.g. CF3SO3

2, [CF3SO2]2N2,CF3CO2

2, CH3CO22, PhSO3

2 and many more have beenprepared.16

Ionic liquids can be prepared by direct quaternisation of theappropriate amine or phosphine. Different anions can subse-quently be introduced by anion exchange. It is beyond the scopeof this review to discuss in detail the synthesis of ionic liquidsand the reader is referred to excellent reviews for manydetails.17–20 It is important to note, however, that ionic liquids,owing to their non-volatile nature, cannot be purified bydistillation. Consequently, they should be produced in highpurity. For example, if synthesis involves exchange of chlorideions it is important that no chloride ions remain in the productas they may seriously impede catalysis by strongly coordinatingto low valent transition metal complexes.

The hydrophilicity/lipophilicity of an ionic liquid can bemodified by a suitable choice of anion, e.g. bmimBF4 iscompletely miscible with water while the PF6 salt is largelyimmiscible with water. The lipophilicity of dialkylimidazoliumsalts, or other ionic liquids, can also be increased by increasingthe chain length of the alkyl groups.21

Ionic liquids containing ‘fluorous ponytails’ have even beendescribed.22 When these are added to conventional ionic liquidsthey facilitate emulsification with perfluorocarbons. Thisprovides the possibility of performing (catalytic) reactions inionic liquids/perfluorocarbon biphasic systems.

Catalysis in ionic liquids: general considerationsRoom temperature ionic liquids exhibit many properties whichmake them potentially attractive media for homogeneouscatalysis:

4 They have essentially no vapour pressure, i.e. they do notevaporate and are easy to contain.

4 They generally have reasonable thermal stability. Whiletetraalkylammonium salts have limited thermal stability,owing to decomposition via the Hoffmann elimination,emimBF4 is reportedly stable up to 300 °C and emim-(CF3SO2)2N up to 400 °C.16a In other words many ionicliquids have liquid ranges of more than 300 °C, compared tothe 100 °C liquid range of water.

4 They are able to dissolve a wide range of organic, inorganicand organometallic compounds.

4 The solubility of gases, e.g. H2, CO and O2, is generally goodwhich makes them attractive solvents for catalytic hydro-genations, carbonylations, hydroformylations, and aerobicoxidations.

4 They are immiscible with some organic solvents, e.g.alkanes, and, hence, can be used in two-phase systems.Similarly, lipophilic ionic liquids can be used in aqueousbiphasic systems.

4 Polarity and hydrophilicity/lipophilicity can be readilyadjusted by a suitable choice of cation/anion (see earlier) andionic liquids have been referred to as ‘designer solvents’.7

4 They are often composed of weakly coordinating anions, e.g.BF4

2 and PF62 and, hence, have the potential to be highly

polar yet non-coordinating solvents. They can be expected,therefore, to have a strong rate-enhancing effect on reactionsinvolving cationic intermediates.

4 Ionic liquids containing chloroaluminate ions are strongLewis, Franklin and Brønsted acids. Protons present inemimAlCl4 have been shown to be superacidic withHammett acidities up to 218.23 Such highly acidic ionicliquids are, nonetheless, easily handled and offer potential asnon-volatile replacements for hazardous acids such as HF inseveral acid-catalysed reactions.

Faced with these numerous potential benefits one maywonder if ionic liquids have any problems associated with theiruse. Atmospheric emissions may not be an issue but, when usedon an industrial scale, small amounts of ionic liquids willinevitably find their way into the environment via the proverbial‘mechanical losses’. So, what is known about their potentialenvironmental impact? A cursory examination of the literaturereveals a dearth of information regarding the biodegradabilityand toxicity of ionic liquids. A prerequisite for industrial use is,therefore, the generation of appropriate data to enable theassessment of the potential environmental impact of ionicliquids.

Another question which arises in any discussion of ionicliquids as reaction media pertains to the isolation of solublereaction products. Volatile products can be separated bydistillation. Non-volatile products, on the other hand, can beseparated by solvent extraction. Although this seems para-doxical—using an ionic liquid to avoid atmospheric emissionsand subsequently using a volatile organic solvent to extract theproduct—it could have environmental benefits. For example,substituting an environmentally unacceptable solvent by anionic liquid as the reaction medium, followed by extraction witha more benign organic solvent would constitute an environ-mental benefit. In this context it is worth noting the use ofsupercritical carbon dioxide to extract products from ionicliquids, which is currently the focus of attention.24 Quiteremarkably, scCO2 is highly soluble (up to 60 mol%) inbmimPF6 while the latter is insoluble in scCO2. Naphthalene,for example, was recovered quantitatively from bmimPF6 byscCO2 extraction, without any contamination of the extract bythe ionic liquid.

One can envisage various scenarios for catalysis in and/or byionic liquids:

4 Monophasic systems in which the catalyst and substrate aredissolved in the ionic liquid.

4 Monophasic systems in which the ionic liquid acts as boththe solvent and the catalyst, e.g. dialkylimidazolium chloro-aluminates as Friedel–Crafts catalysts (see later).

4 Biphasic systems in which the catalyst resides in the ionicliquid and the substrate/product in the second phase or viceversa.

4 Mono- or biphasic systems in which the anion of the ionicliquid acts as a ligand for the homogeneous catalyst, e.g. asulfonated phosphine ligand (see later).

4 Triphasic systems comprising, for example, an ionic liquid,water and an organic phase in which the catalyst resides inthe ionic liquid, the substrate and product in the organicphase and salts formed in the reaction are extracted into theaqueous phase, e.g. in Heck reactions (see later).

The first example of homogeneous transition metal catalysisin an ionic liquid is the platinum catalysed hydroformylation ofethene in tetraethylammonium trichlorostannate, described byParshall in 1972.25 This ionic liquid (referred to as a molten saltback in those days) has a melting point of 78 °C. These resultswere largely ignored for two decades. The potential of ionicliquids as novel media for homogeneous catalysis became morewidely appreciated largely due to the pioneering studies andextensive promotion of the groups of Seddon17 and Chauvin andOlivier-Bourbigou18 and, more recently, the groups of Welton19

2400 Chem. Commun., 2001, 2399–2407

and Keim and Wasserscheid.20 In the last five years their use asnovel media for, inter alia, catalytic hydrogenations, hydro-formylations, isomerisations, olefin dimerisations, oligomerisa-tions and polymerisations and Heck couplings, has been rapidlyexpanding. The salient features of these studies will be reviewedin the ensuing discussion, with emphasis on their potential asclean synthetic methodologies.

HydrogenationThe first example of catalytic hydrogenation in an ionic liquidwas reported by Chauvin et al. in 1995.26 A solution of thecationic [Rh nbd(Ph3P)2]PF6 complex [nbd = norbornadiene(bicyclo[2.2.1]hepta-2,5-diene)] in bmimPF6 or bmimSbF6 wasshown to be an effective catalyst for the biphasic hydrogenationof pent-1-ene. Reaction rates were up to five times higher thanin acetone as solvent which was attributed to the formation of anunsolvated cationic rhodium(III) dihydride complex with twofree coordination sites in the nonsolvating ionic liquid. Incontrast, poor results were obtained with bmimBF4 which wasascribed to the presence of trace amounts of strongly coordinat-ing chloride ions in their sample of this ionic liquid. The catalystsolution in the ionic liquid could be reused with rhodium lossesbelow the detection limit of 0.02%.

Similarly, advantage was taken of the biphasic system toperform the selective hydrogenation of cyclohexadiene. Thesolubility of cyclohexadiene in bmimSbF6 is about five timesthat of cyclohexene and, hence, the latter was obtained in 98%selectivity at 96% conversion.

Dupont and coworkers27 performed the biphasic hydro-genation of cyclohexene with Rh(cod)2BF4 (cod = cycloocta-1,5-diene) in ionic liquids. They observed roughly equal rates(turnover frequencies of ca. 50 h21) in bmimBF4 and bmimPF6

(presumably their bmimBF4 was chloride-free).The same group showed that RuCl2(Ph3P)3 in bmimBF4 is an

effective catalyst for the biphasic hydrogenation of olefins, withturnover frequencies up to 540 h21.28 Similarly, (bmim)3-Co(CN)5 dissolved in bmimBF4 catalysed the hydrogenation ofbutadiene to but-1-ene, in 100% selectivity at completeconversion.28

More recently, the ruthenium-catalysed hydrogenation ofsorbic acid to cis-hex-3-enoic acid (Reaction 1) was achieved ina biphasic bmimPF6–methyl tert-butyl ether system.29

(1)

The ruthenium cluster, [H4Ru(h6-C6H6)4] [BF4]2 inbmimBF4 was shown to be an effective catalyst for thehydrogenation of arenes, to the corresponding cycloalkanes, at90 °C and 60 bar.30 The cycloalkane product formed a separatephase which was decanted and the ionic liquid phase, containingthe catalyst, could be repeatedly recycled.

Enantioselective hydrogenation in ionic liquids is of partic-ular interest as it could provide a means for facile recycling ofmetal complexes of expensive chiral ligands. In their originalstudy Chauvin et al.26 reported that [Rh(cod)(2)-(diop)]PF6

catalysed the enantioselective hydrogenation of a-acetamido-cinnamic acid to (S)-phenylalanine, in 64% ee, in a biphasicbmimSbF6–isopropyl alcohol (Reaction 2). The observed

(2)

enantioselectivity is what one would expect with diop which isnot a particularly good ligand for this reaction. The product,contained in the isopropyl alcohol, could be separated quantita-tively and the recovered ionic liquid, containing the catalyst,reused.

Similarly, Dupont and coworkers31 extended their studies ofruthenium-catalysed hydrogenations in ionic liquids to enantio-selective reactions. The chiral [RuCl2(S)-BINAP]2NEt3 com-plex was shown to catalyse the asymmetric hydrogenation of2-phenylacrylic acid and 2-(6-methoxy-2-naphthyl)acrylic acidin bmimBF4–isopropyl alcohol. The latter afforded the anti-inflammatory drug, (S)-naproxen, in 80% ee (Reaction 3). The

(3)

product could be quantitatively separated and the recoveredionic liquid catalyst solution recycled several times without anysignificant change in activity or selectivity.

An interesting recent development is the use of a biphasicionic liquid–supercritical CO2 for catalytic hydrogenation32,33

and other processes (see later). Tumas and coworkers32 showedthat the catalytic hydrogenation of olefins could be conducted ina biphasic bmimPF6–scCO2 system. The ionic liquid phasecontaining the catalyst was separated by decantation and reusedin up to four consecutive batches.

Jessop and coworkers33 extended this concept to theasymmetric hydrogenation of tiglic acid (Reaction 4) and theprecursor of the antiinflammatory drug ibuprofen (Reaction 5)using Ru(OAc)2(tolBINAP) as the catalyst.

(4)

(5)

They found that Reaction 4 was more selective in abmimPF6–water biphasic mixture while Reaction 5 gave poorenantioselectivities in the wet ionic liquid. In this case the bestresult (85% ee) was obtained using methanol as cosolvent at 100bar H2 pressure. In both cases the product was separated byscCO2 extraction when the reaction was complete. The differentsolvent effects observed with the two substrates was assumed tobe due to the solubility of H2 in the reaction mixture. Thehydrogen concentration dependence of asymmetric catalytichydrogenation with ruthenium BINAP complexes is known tobe dependent on the substrate.34 Class I substrates such as theibuprofen precursor give higher enantioselectivities at high H2

concentration while class II substrates, exemplified by tiglicacid, give higher enantioselectivities at low H2 concentra-tions.

HydroformylationHydroformylation of propene in an aqueous biphasic system,using a water-soluble rhodium complex of the sodium salt oftrisulfonated triphenylphosphine (tppts) forms the basis of theRuhr Chemie Rhone Poulenc process for the manufacture ofbutanal.35 Unfortunately this process is limited to C2 to C5

olefins owing to the very low solubility of higher olefins inwater. Hence, one can envisage that the use of an appropriateionic liquid could provide the basis for biphasic hydro-formylation of higher olefins.

Chem. Commun., 2001, 2399–2407 2401

As noted earlier, Parshall showed, in 1972, that platinum-catalysed hydroformylations could be performed in tetra-ethylammonium trichlorostannate melts.25 More recently, Waf-fenschmidt and Wasserscheid36 studied the platinum-catalysedhydroformylation of oct-1-ene in bmimSnCl3 (Reaction 6)

(6)

which is liquid at room temperature. Despite the limitedsolubility of oct-1-ene in the ionic liquid, high activities (TOF =126 h21) were observed together with a remarkably highregioselectivity (n/iso = 19). The product was recovered byphase separation and no leaching of platinum was observed.

The ruthenium- and cobalt-catalysed hydroformylation ofinternal and terminal olefins in molten tetra-n-butylphosphon-ium bromide was reported by Knifton in 1987.37 More recently,the rhodium-catalysed hydroformylation of hex-1-ene wasconducted in molten phosphonium tosylates, e.g. Bu3PEt+TsO2and Ph3PEt+TsO2 having melting points of 81–83 °C and94–95 °C, respectively, at 120 °C and 40 bar.38 Advantage wastaken of the higher melting points of these 'ionic liquids' todecant the product from the solid catalyst medium at roomtemperature.

Chauvin and coworkers26 investigated the rhodium-catalysedbiphasic hydroformylation of pen-1-tene in bmimPF6. Highactivities (TOF = 333 h21 compared with 297 h21 in toluene)were observed with the neutral Rh(CO)2(acac)–Ph3P as thecatalyst precursor but some leaching of the catalyst into theorganic phase occurred. This could be avoided by usingRh(CO)2acac with tppts or tppms (monosulfonated triphenyl-phosphine) as the catalyst precursor, albeit at the expense of rate(TOF = 59 h21 with tppms). Higher activities (TOF = 810h21) and high regioselectivity (n/iso = 16) were observed in thebiphasic hydroformylation of oct-1-ene in bmimPF6 usingcationic cobaltocenium diphosphine ligands but some catalystleaching ( < 0.5%) was observed.39

Better results were obtained with cationic guanidine-mod-ified diphosphine ligands containing a xanthene backbone.40

Xanthene-based diphosphine ligands with large bite angles (P–metal–P ~ 110°) are known to give high selectivities (!98%)towards the linear aldehyde.41 Biphasic hydroformylation ofoct-1-ene, using rhodium complexes of these ligands inbmimPF6 (Reaction 7), afforded high regioselectivities (ca. 20)

(7)

and the catalyst could be recycled ten times (resulting in anoverall turnover number of 3500) without detectable ( < 0.07%)leaching of Rh to the organic phase.

The group of Olivier-Bourbignou42 has recently explored theuse of a wide range of ionic liquids, based on imidazolium andpyrrolidinium cations and weakly coordinating anions, for thebiphasic hydroformylation of hex-1-ene catalysed by rhodiumcomplexes of modified phosphine and phosphite ligands. Thelatter are, in contrast, unstable in aqueous biphasic media. Therate and regioselectivity could be optimized by choosing asuitable combination of cation, anion and phosphine orphosphite ligand. Rhodium leaching was minimised by mod-ification of the ligands with cationic (guanidinium or pyr-idinium) or anionic (sulfonate) groups.

Another interesting recent development is the rhodium-catalysd biphasic hydroformylation of oct-1-ene in bmimPF6–

scCO2 in a continuous flow process.43 Because of the lowsolubility of Rh–tppms and Rh–tppts complexes in the ionicliquid, [pmim]+Ph2PC6H4SO3

2 (pmim = 1-propyl-3-methyl-imidazolium) was synthesised and used together withRh2(OAc)4 as the catalyst precursor. Aldehydes were producedat a constant rate for 72 h albeit with moderate regioselectivity(n/iso = 3.8). Analysis of recovered products revealed that < 1ppm Rh is leached into the organic phase.

The monophasic hydroformylation of methylpent-3-enoate inbmimPF6 has been reported.44 The linear aldehyde product(Reaction 8) is a precursor of adipic acid in an alternative

(8)

butadiene-based route. The product was removed by distillation(0.2 mbar/110 °C) and the ionic liquid recycled ten timeswithout significant loss in activity.

AlkoxycarbonylationMuch less attention has been focused on carbonylation reactionsin ionic liquids. The biphasic palladium-catalysed alkoxy-carbonylation of styrene (Reaction 9) in bmimBF4–cyclohex-

(9)

ane has been reported.45 Very high regioselectivities (!99.5%iso) were obtained, using PdCl2(PhCN)2 in combination with(+)-neomenthyldiphenylphosphine and toluene-p-sulfonic acid,under mild conditions (70 °C and 10 bar).

More recently, the palladium-catalysed alkoxycarbonylationand amidocarbonylation of aryl bromides and iodides inbmimBF4 and bmimPF6 has been described.46 Enhancedreactivities were observed compared to conventional media andthe ionic liquid–catalyst could be recycled.

Olefin dimerisation and oligomerisationThe nickel-catalysed dimerisation of lower olefins in ionicliquids containing chloroaluminate anions is probably the mostinvestigated reaction in ionic liquids.18,26,47–49 As early as 1990the group of Chauvin at the Institut Francais du Petrole (IFP)reported the nickel-catalysed dimerisation of propene inbmimAlCl4.47 The catalyst precursor consisted of L2NiCl2 (L =Ph3P or pyridine) in combination with EtAlCl2 (bmimCl–AlCl3–EtAlCl2 = 1+1.2+0.25). The active catalyst is a cationicnickel(II) complex, [LNiCH2CH3]+AlCl42, formed by reactionof L2NiCl2 with EtAlCl2. Since ionic liquids promote thedissociation of ionic metal complexes it was envisaged that theywould have a beneficial effect on this reaction.18

This proved to be the case: at 5 °C and atmospheric pressureproductivities ( > 250 kg dimers per g Ni) much higher thanthose observed in organic solvents were achieved.18,47,48 Themixture of dimers obtained, containing 2,3-dimethylbutene asthe major component (83%), has commercial importance as theprecursor of octane boosters for reformulated gasoline. It hasbeen produced since the mid-seventies by the IFP ‘Dimersol’process (25 units worldwide with a production of 3.4 3 106 tonsper annum) in a single-phase solvent-free medium.

The methodology was subsequently extended to the dimer-isation of butenes. The isooctene product constitutes thefeedstock for the manufacture of isononanols (plasticizers) by

2402 Chem. Commun., 2001, 2399–2407

hydroformylation. A productivity of > 100 kg dimers per g Niwas obtained at 10 °C.

Conducting these dimerisations as biphasic reactions in ionicliquids affords several benefits:18 a better selectivity to dimers(owing to their low solubility in the ionic liquid), a better use ofthe catalyst components and, hence, reduced disposal costs,substantially reduced reactor size, no corrosion and broaderscope (to less reactive, higher olefins).

The nickel-catalysed biphasic dimerisation of olefins in ionicliquids is being offered for licensing by IFP under the acronym‘Difasol process’ and is likely to be the first large scaleapplication of biphasic catalysis in ionic liquids.

Wasserscheid and Keim20,50 have developed alternative,alkylaluminium-free nickel catalysts for the linear dimerisationof but-1-ene. A turnover frequency of 1240 h21 and a dimerselectivity of 98% (64% linearity) was observed at 25 °C. Morerecently, the same group reported51 the use of cationic nickelcomplexes for the biphasic oligomerisation of ethene to highera-olefins in bmimPF6 (Reaction 10). The products separated as

(10)

a clear and colourless layer and the catalyst-containing ionicliquid could be recycled with leaching below the detection limit(0.1%).

The palladium-catalysed dimerisation of butadiene to octa-1,3,6-triene and octa-1,3,7-triene is industrially important as theproducts have a wide range of applications, e.g. as comonomersand in the synthesis of plasticizers, adhesives and fragrances.Since these octatrienes rapidly polymerise in the presence of air,separation of the product from the homogeneous catalyst, e.g.by distillation, presents a serious problem. This would seem,therefore, to be an attractive target for biphasic catalysis in anionic liquid.

Dupont and coworkers52 have reported that PdCl2–Ph3P (1:4)catalyses the biphasic dimerisation of butadiene (Reaction 11)

(11)

in bmimX (X = BF4, PF6 or CF3SO3) at 70 °C, affording octa-1,3,6-triene in 100% selectivity.

The same group reported53 that when the reaction isperformed with (bmim)2PdCl4 in bmimBF4–H2O (roughly 1+1v/v) selective telomerisation resulted to afford octa-2,7-dien-1-ol (Reaction 12). As is also observed in other mono- and

(12)

biphasic telomerisations of butadiene, the presence of carbondioxide was essential for high activity. Using 5 bar CO2 aturnover frequency of 204 h21 and a selectivity of 89% (11%octa-1,3,6-triene) at 49% conversion was observed at 70 °C.Interestingly, the reaction is monophasic under the reactionconditions but cooling the mixture to 5 °C produces two phasesand the ionic liquid phase can be separated and recycled. Thismethodology is a potentially attractive alternative to theaqueous biphasic telomerisation of butadiene developed byKuraray.3,4

Heck reactionsThe Heck and related C–C coupling reactions are of majorimportance in organic synthesis and are finding wide applica-tion in the manufacture of fine chemicals.54 The first example ofa Heck coupling in an ionic liquid was reported by Kaufmann etal. in 1996.55 Butyl trans-cinnamate was produced in high yieldby reaction of bromobenzene with butyl acrylate in moltentetraalkylammonium and tetraalkylphosphonium bromide salts(Reaction 13). No formation of palladium metal was observed

(13)

and the product was isolated by distillation from the ionicliquid.

Herrmann and Böhm56 subsequently showed that moltenBu4NBr (mp 103 °C) is a particularly suitable reaction mediumfor Heck reactions, affording superior results compared withcommonly used organic solvents such as DMF. For example, inthe reaction of bromobenzene with styrene, using diiodobis(1,3-dimethylimidazolin-2-ylidine)palladium(II) as the catalyst, theyield of stilbene was increased from 20% in DMF to 99% inBu4NBr under otherwise identical conditions. The product wasseparated by distillation and the catalyst containing ionic liquidrecycled up to 13 times without significant loss of activity.

Seddon and coworkers57 performed Heck couplings inbmimPF6 or n-hexylpyridinium PF6 using PdCl2 or Pd(OAc)2–Ar3P as the catalyst and Et3N or NaHCO3 as the base.

For example, Pd(OAc)2–Ph3P-catalysed coupling of 4-bro-moanisole with ethyl acrylate (Reaction 14) in bmimPF6 at140 °C, afforded ethyl 4-methoxycinnamate in 98% yield.

(14)

The high solubility of the catalyst in the ionic liquid allowsfor product isolation by extraction into a hydrocarbon, e.g.hexane or toluene. Furthermore, if water is added a triphasicsystem is obtained in which the salt formed in the reaction,Et3NHBr, is extracted into the aqueous phase.

It was also noted that palladium complexes of imidazolyli-dene carbenes, formed by reaction of the base with theimidazolium cation, may be implicated in these reactions.57

This was later confirmed by Xiao and coworkers58 whoobserved a significantly enhanced rate of the Heck coupling inbmimBr compared to the same reaction in bmimBF4. Thisdifference could be explained by the formation of the corre-sponding palladium–carbene complexes (which were isolatedand characterised) in the former but not in the latter. Theisolated carbene complexes were shown to be active catalystswhen redissolved in bmimBr. Presumably, the formation of thecarbene in bmimBr can be attributed to the stronger basicity ofbromide compared to tetrafluoroborate.

The Heck arylation of electron-rich enol ethers generallyleads to a mixture of regio isomers owing to competitionbetween cationic and neutral pathways, leading to a- and b-arylation, respectively (Reaction 15). The ionic pathway isfavoured with aryl triflates but these are less available and muchmore expensive than the corresponding chlorides and bromides.The ionic pathway would also be expected to be favoured byconducting the reaction in an ionic liquid and this proved to bethe case. Thus, Xiao and coworkers59 achieved > 99% selectiv-ity to the a-arylation product in the Heck coupling of1-bromonaphthalene to butyl vinyl ether in bmimBF4 (Reaction15). In contrast, the same reaction in toluene, acetonitrile, DMFor DMSO afforded mixtures of the a- and b-regio isomers. Arange of 4-substituted bromobenzenes were similarly shown togive a-/b-regioselectivities of > 99%.

Chem. Commun., 2001, 2399–2407 2403

(15)

Similarly, palladium-catalysed Stille60a and Negishi60b cou-plings and nickel-catalysed coupling of aryl halides61 inbmimBF4 and bmimPF6, respectively, have also been de-scribed.

Palladium-catalysed allylic substitutionPalladium-catalysed allylic substitution by carbon nucleophilesconstitutes another synthetically useful method for the genera-tion of C–C bonds. These reactions have also been performed inionic liquids, both in a mono- and biphasic system, usingPd(OAc)2–Ph3P (with K2CO3 as base) in bmimBF4 (Reaction16)62 and PdCl2–tppts in bmimCl–methylcyclohexane (Reac-tion 17),63 respectively. In the latter case distinct advantages

(16)

(17)

compared to the corresponding aqueous biphasic system werenoted: an order of magnitude higher activity and improvedselectivity owing to suppression of the competing reaction withwater as a nucleophile and a much decreased phosphonium saltformation (by reaction of tppts with the Pd–allyl complex).

Reaction (16) has also been performed using chiral ferroce-nylphosphine complexes of palladium, in bmimPF6.64 Theproduct was obtained in moderate enantioselectivity (62–74%ee), which was higher than that observed in conventionalsolvents.

Catalytic oxidationsConsidering the commercial importance of catalytic oxidations,and the fact that ionic liquids are expected to be relatively inerttowards autoxidation with O2, surprisingly little attention hasbeen devoted to performing such reactions in ionic liquids. TheNi(acac)2-catalysed aerobic oxidation of aromatic aldehydes, tothe corresponding carboxylic acids, in bmimPF6 has beendescribed.65 However, rather high (3 mol%) catalyst loadingswere used and this can hardly be considered a challengingoxidation.

The methyltrioxorhenium (MTO)-catalysed epoxidation ofolefins with the urea–H2O2 adduct (UHP) in emimBF4 has beenreported.66 Both the UHP and the MTO are soluble in emimBF4

and the medium remains homogeneous throughout the reaction.It should be noted, however, that the substrates were generally

highly reactive olefins and when the more challenging dec-1-ene was used, a long reaction time (72 h) was needed formoderate conversion (46%) using 2 equivalents of oxidant.When 30% aq. H2O2 was used as the oxidant this led to ringopening of sensitive epoxides.

Asymmetric Jacobsen-Katsuki epoxidation, with NaOClcatalysed by a chiral Mn Schiff’s base complex has beenconducted in bmimPF6.67 However, dichloromethane wasrequired as a cosolvent, as the ionic liquid solidifies at thereaction temperature (0 °C), which nullifies one of the primaryincentives for using an ionic liquid. The ionic liquid, containingthe catalyst, could be recovered and recycled 4 times albeit witha significant loss in yield.

A more recent, and very exciting development, is theelectroassisted biomimetic activation of molecular oxygen by achiral Mn Schiff's base complex in bmimPF6 described byGaillon and Bedioui.68 Evidence was provided for the formationof the highly reactive oxomanganese(V) intermediate that couldtransfer its oxygen to an olefin. This would appear to offerpotential for clean, electrocatalytic oxidations with molecularoxygen in ionic liquid media.

Lewis and Brønsted acid catalysis in ionic liquidsIonic liquids containing chloroaluminate (AlCl42, Al2Cl72)anions are strong Lewis acids and if protons are present they aresuperacidic (see earlier). Coupled with the fact that they arerelatively easy to handle this makes these materials attractivenon-volatile alternatives for standard Lewis acid catalysts, suchas AlCl3, and hazardous Brønsted acids such as HF. The ionicliquid can function as both a catalyst and a solvent for acidcatalysed processes. Since Lewis and Brønsted acid-mediatedprocesses generally involve cationic intermediates, e.g. carben-ium and acylium ions, one would also expect to see substantialrate enhancements in ionic liquids. Indeed, some of the firstreactions to be studied in ionic liquids were Friedel–Craftsalkylations and acylations. Wilkes and coworkers69 showed thationic liquids derived from the reaction of emimCl with AlCl3exhibit a wide range of Lewis acidity depending on the molarratio of reactants. A 1+1 mixture affords the tetrachloro-aluminate, emimAlCl4, which is referred to as being neutral andis not active as a Friedel–Crafts catalyst. In contrast, the 2+1adduct, emimAl2Cl7 is strongly acidic and was shown to be veryactive in Friedel–Crafts alkylations and acylations.69 Forexample, a mixture of benzene, acetyl chloride and emimAl2Cl7in the molar ratio 1.1+1.0+0.5 (i.e. less than a stoichiometricamount of the ionic liquid) afforded complete conversion toacetophenone in less than 5 minutes at room temperature.Spectral evidence suggested the formation of the free acyliumcation, CH3CO+, in the ionic liquid medium.

The Friedel–Crafts alkylation of benzene with long chain a-olefins is used industrially for the manufacture of more than twomillion tons of linear alkylbenzenes worldwide. The productsare the precursors of the corresponding alkylbenzene sulfonateswhich are widely used as surfactants. Traditionally the reactionis performed using liquid HF or AlCl3 as the catalyst. Theproduction of linear alkylbenzenes using chloroaluminate ionicliquids has been described.17a The potential to retrofit existinginstallations with the ionic liquid catalyst offers enormousbenefits with regard to reduced catalyst consumption, ease ofproduct separation and elimination of caustic quenchingassociated with catalyst leaching. Chloroaluminate ionic liquidsmodified with HCl were recently shown70 to give higher ratesand more favourable product distributions in Friedel–Craftsalkylations, which was attributed to the superacidities of thesemedia. In this context it is also worth mentioning the work ofHölderich et al.71 who showed that ionic liquids immobilised oninorganic supports (SiO2, Al2O3, TiO2, ZrO2) are effectivecatalysts for Friedel–Crafts alkylation of aromatics. Activities

2404 Chem. Commun., 2001, 2399–2407

were higher than those observed with a conventional zeolitecatalyst and no leaching of the ionic liquid from the surface wasobserved. Reactions were performed in batch, continuousliquid-phase and continuous gas-phase operation. For example,alkylation of benzene with dodecene afforded the mono-alkylated product in 98% selectivity at 99% conversion at80 °C.

Seddon and coworkers72 studied the Friedel–Crafts acylationof toluene, chlorobenzene and anisole with acetyl chloride inemimAl2Cl7 and obtained excellent regioselectivities to thepara isomer (Reaction 18). Similarly, the fragrance chemical,traseolide, was obtained in 99% yield as a single isomer(Reaction 19). It should be noted, however, that the question of

(18)

(19)

product recovery from the reaction medium still needs to beaddressed in these systems. As in conventional AlCl3-promotedacylations the ketone product forms a strong complex with thechloroaluminate ionic liquid.

Lanthanide triflates, in particular Sc(OTf)3, have been widelystudied as water-tolerant Lewis acids in a variety of transforma-tions, including Friedel–Crafts alkylations and acylations.73

Song and coworkers74 have recently shown that Sc(OTf)3

catalyses the Friedel–Crafts alkylation of aromatics with olefinsin hydrophobic ionic liquids, e.g. bmimPF6 and bmimSbF6. Incontrast, no reaction was observed in common organic solvents,water or hydrophilic ionic liquids such as bmimBF4 orbmimOTf. For example, reaction of benzene with cyclohexene(Reaction 20) afforded cyclohexylbenzene in 92% yield at

(20)

> 99% cyclohexene conversion in bmimSbF6 at 20 °C for 12 h.The product formed a separate layer and, after phase separation,the ionic liquid phase, containing the catalyst, was recycled toafford 92% yield of cyclohexylbenzene at > 99% conversion.Although high catalyst loadings (20 mol%) were used the easeof separation and recycling of the catalyst offers potentialenvironmental and economic benefits.

The same group has recently reported75 that Sc(OTf)3

catalyses Diels–Alder reactions in bmimX (X = BF4, SbF6 orOTf), in this case at much lower catalyst loadings (0.2 m%). Incontrast to the Friedel–Crafts alkylation (see above) the productdid not form a separate phase and was recovered by extractionwith ether. It was shown, however, that the ionic liquidcontaining the catalyst could be recycled eleven times without

loss of activity (Reaction 21). Furthermore, improved endo/exoselectivities were observed with cyclic dienes.

(21)

Another reaction catalysed by Lewis acids is the cycloaddi-tion of carbon dioxide to epoxides, affording cyclic carbonates.It was recently reported that this reaction is catalysed bybmimBF4.76 When propene oxide was allowed to react withCO2 (2.5 MPa) at 110 °C for 6 h in the presence of bmimBF4

(2.5 mol%), propene carbonate was obtained in 100% yield(Reaction 22).

(22)

The propene carbonate was distilled from the reactionmixture and the ionic liquid catalyst recycled four times withonly a minor loss in activity.

Protons in acidic ionic liquids have acidities greater thanthose of H2SO4 or liquid HF.8,17,18 They are, for example, ableto protonate benzene to the cyclohexadienyl cation. The acidicionic liquid, bmimAl2Cl7, catalyses the alkylation of ethenewith isobutane18 whereas neither HF nor H2SO4 is effective forthis reaction. The major product is 2,3-dimethylbutane(75–86%) which has a higher octane number than the productsof alkylation of higher olefins. The product forms a separateupper phase and the ionic liquid is readily recycled.

Olefin oligomerisation and polymerisation is also promotedby these strongly acidic ionic liquids, e.g. high molecularweight polyisobutene is readily obtained from isobutene.17 Thecatalytic activity and degree of polymerisation is determined bythe chain length of the alkyl group in the 1-alkyl-3-methylimi-dazolium or N-alkylpyridinium cation which provides a con-venient mechanism for controlling the product distribution.Polyisobutene is a commercially important lubricant and theionic liquid process has several advantages compared with theconventional process which employs a supported or dissolvedAlCl3 catalyst. The polymer forms a separate phase and isreadily separated and reused which obviates the need for anaqueous wash and, hence, reduces waste and costs. Moreover,the process can be easily retrofitted into existing productionfacilities.

The electrophilic nitration of aromatics, using a variety ofnitrating agents, has also been investigated in ionic liquidsolvents.77 It was noted that nitration in ionic liquids constitutesa useful alternative to classical methods owing to easier productisolation and recovery of the ionic liquid and avoidance ofproblems associated with the neutralisation of large quantititiesof acid.

Many synthetically important rearrangement reactions arecatalysed by Brønsted or Lewis acids and, hence, may benefitfrom using acidic ionic liquids as solvents and/or catalysts. Forexample, the Beckmann rearrangement of ketoximes in ionicliquids, in the presence of catalytic amounts (20 mol%) ofphosphorus compounds (e.g. P2O5) was recently reported.78

The ionic liquid tributylhexylammonium bis(trifluorome-thylsulfonyl) amide was shown79 to be a useful solvent for theacid-catalysed cyclotrimerisation of veratryl alcohol. Themethodology obviates the need for large quantities of organicsolvent and strongly dehydrating acids and provides thecyclotriveratrylene (CTV) in high yield and purity. CTV is ofinterest as a supramolecular host compound that complexes avariety of guest molecules.

Chem. Commun., 2001, 2399–2407 2405

Biocatalysis in ionic liquidsAttention has recently been focused on the use of enzymes inionic liquids. It was already noted in 198480 that the enzymealkaline phosphatase is relatively stable in a 4+1 (v/v) mixtureof triethylammonium nitrate and water. More recently, Lye andcoworkers81 reported a two-phase biotransformation in whichbmimPF6 acts as a reservoir for the substrate while thebiocatalyst—whole cells of Rhodococcus R312—is present inthe aqueous phase. Shortly thereafter two reports of enzymaticconversions in an ionic liquid medium appeared.82,83

Erbeldinger et al.82 reported the thermolysin-catalysedsynthesis of Z-aspartame (Reaction 23) in bmimPF6 containing

(23)

5% (v/v) water. The enzyme displayed excellent stability whensuspended in the ionic liquid and the activity was equal to thatobserved in ethyl acetate–water. A small amount of the enzyme( < 3.2 mg mL21) which dissolved in the ionic liquid wascompletely inactive.

We showed83 that Candida antarctica lipase is able tocatalyse a variety of transformations—transesterification, am-moniolysis and perhydrolysis—in bmimBF4 or bmimPF6 in theabsence of water. Reactions were performed with the free(NOVO SP525) or immobilized enzyme (Novozyme 435).Reaction rates were comparable with or better than thoseobserved in conventional organic media. For example, thereaction of octanoic acid with ammonia, in bmimBF4 at 40 °C,gave complete conversion to octanamide in 4 days compared to17 days for the same conversion using ammonium carbamate inmethyl isobutyl ketone.84

The epoxidation of cyclohexene by peroxyoctanoic acid,generated in situ by Novozyme 435-catalysed reaction ofoctanoic acid with 60% aqueous H2O2, proceeded smoothly inbmimBF4 (Reaction 24).

(24)

Subsequently, other groups have reported on lipase-catalysedenantioselective transesterification of chiral alcohols (Reaction25) in ionic liquids.85–87 Kragl and coworkers85 investigated the

(25)

kinetic resolution of 1-phenylethanol with nine different lipasesin ten different ionic liquids. Good activities and, in many cases,improved enantioselectivities were observed compared with thesame reaction in MTBE (methyl tert-butyl ether). Rates and/orenantioselectivities were dependent on the anion as was also

observed by Itoh and coworkers.86 It was also shown that theproduct could be extracted with ether and the ionic liquid(bmimPF6), containing the suspended enzyme, could berecycled.86 Similarly, Kim and coworkers87 observed markedlyenhanced enantioselectivities in Candida antarctica and Pseu-domonas cepacia lipase-catalysed transesterifications of chiralalcohols in bmimBF4 and bmimPF6.

Based on these initial studies the use of enzymes in ionicliquids would appear to open up a new field of nonaqueousenzymology. Ionic liquids could have added benefits forperforming biotransformations with highly polar substrates, e.g.carbohydrates and amino acids, which are sparingly soluble incommon organic solvents. We are currently investigating thescope with regard to type of enzyme, transformation and theeffect of the structure of the ionic liquid.

Concluding remarksCatalysis in ionic liquids is an exciting and burgeoning area ofresearch (many of the references in this review are from 2000and 2001) which holds considerable potential for industrialapplication. The use of ionic liquids as reaction media forcatalytic transformations or, in some cases, as the catalyst itselfcan have a profound effect on activities and selectivities.Furthermore, precision tuning of reactions can be achieved by asuitable combination of cation and anion, i.e. they are truly‘designer solvents’. In the majority of cases studied the ionicliquid, containing the catalyst could be readily recycled. Theyprovide a medium for performing clean reactions with mini-mum waste generation. As was remarked by Seddon8 theycould, quite literally, revolutionise the methodology of syn-thetic organic chemistry.

Notes and references† We have adopted the abbreviations used by many (but not all) authors fordialkylimidazolium cations, viz., emim for ethylmethylimidazolium, bmimfor butylmethylimidazolium, etc.

1 A. D. Curzons, D. J. C. Constable, D. N. Mortimer and V. L.Cunningham, Green Chem., 2001, 3, 1.

2 For leading references see: R. A. Sheldon, Pure Appl. Chem., 2000, 72,1233; R. A. Sheldon, Chem. Ind. (London), 1997, 12.

3 Aqueous Phase Organometallic Catalysis, eds. B. Cornils and W. A.Herrmann, Wiley-VCH, Weinheim, 1998.

4 G. Papadogianakis and R. A. Sheldon, Specialist Periodical ReportsCatalysis, Vol. 13, Senior Reporter J. J. Spivey, Royal Society ofChemistry, Cambridge, 1997, pp. 114–193.

5 I. T. Horvath and J. Rabai, Science, 1994, 266, 72.6 Chemical Synthesis using Supercritical Carbon Fluids, ed. P. G. Jessop

and W. Leitner, VCH/Wiley, Weinheim, 1999; W. Leitner, Top. Curr.Chem., 1999, 206, 107.

7 M. Freemantle, C&EN, March 30, 1998, p. 32; May 15, 2000, p. 37;January 1, 2001, p. 21; H. Carmichael, Chem. Brit., January 2000, p.36.

8 K. Seddon, Molten Salt Forum, 1998, 5–6, 53.9 P. Walden, Bull. Acad. Imper. Sci. (St. Petersburg), 1914, p. 1800; cited

in S. Sugden and H. Wilkins, J. Chem. Soc., 1929, p. 1291.10 C. L. Hussey, Adv. Molten Salt Chem., 1983, 5, 185.11 H. L. Chum, V. R. Kock, L. L. Miller and R. A. Osteryoung, J. Am.

Chem. Soc., 1975, 97, 3264; J. Robinson and R. A. Osteryoung, J. Am.Chem. Soc., 1979, 101, 323.

12 J. S. Wilkes, J. A. Levinsky, R. A. Wilson and C. A. Hussey, Inorg.Chem., 1982, 21, 1263.

13 J. A. Boon, J. A. Levinsky, J. L. Pflug and J. S. Wilkes, J. Org. Chem.,1986, 51, 480.

14 J. S. Wilkes and M. J. Zaworotko, J. Chem. Soc., Chem. Commun.,1992, 965.

15 J. Fuller, R. T. Carlin, H. C. De Lang and D. Haworth, J. Chem. Soc.,Chem. Commun., 1994, 299.

16 (a) P. Bonhôte, A. P. Dias, N. Papageorgiou, K. Kalyanasundaram andM. Grätzeb, Inorg. Chem., 1996, 35, 1168; (b) V. R. Koch, C.Nanjundiah, G. B. Appetecchi and B. Scrosati, J. Electrochem. Soc.,1995, 142, 116; (c) D. R. Macfarlane, J. Golding, S. Forsyth and G. B.Deacon, Chem. Commun., 2001, 1430.

2406 Chem. Commun., 2001, 2399–2407

17 J. D. Holbrey and K. R. Seddon, Clean Prod. Process., 1999, 1, 223;K. R. Seddon, J. Chem. Technol. Biotechnol., 1997, 68, 351.

18 Y. Chauvin and H. Olivier-Bourbigou, CHEMTECH, September 1995,p. 26.

19 T. Welton, Chem. Rev., 1999, 99, 2071.20 P. Wasserscheid and W. Keim, Angew. Chem., Int. Ed., 2000, 39,

3772.21 J. G. Huddleston, A. E. Visser, W. M. Reichert, H. D. Willauer, G. A.

Broker and R. D. Rogers, Green Chem., 2001, 3, 156; see also S. V.Dzyuba and R. A. Bartsch, Chem. Commun., 2001, 1466.

22 T. L. Merrigan, E. D. Bates, S. C. Dorman and J. H. Davis, Chem.Commun., 2000, 2051.

23 G. P. Smith, A. S. Dworkin, R. M. Pagni and S. P. Zingg, J. Am. Chem.Soc., 1989, 111, 525.

24 L. A. Blanchard, D. Hancu, E. J. Beckman and J. F. Brennecke, Nature,1999, 399, 28; L. A. Blanchard and J. F. Brennecke, Ind. Eng. Chem.Res., 2001, 40, 287.

25 G. W. Parshall, J. Am. Chem. Soc., 1972, 94, 8716.26 Y. Chauvin, L. Mussman and H. Olivier, Angew. Chem., Int. Ed. Engl.,

1995, 34, 2698.27 P. A. Z. Suarez, J. E. L. Dullins, S. Einloft, R. F. de Souza and J. Dupont,

Polyhedron, 1996, 15, 1217.28 P. A. Z. Suarez, J. E. L. Dullins, S. Einloft, R. F. de Souza and J. Dupont,

Inorg. Chim. Acta, 1997, 255, 207.29 S. Steines, B. Drießen-Hölscher and P. Wasserscheid, J. Prakt. Chem.,

2000, 342, 348.30 P. J. Dyson, D. J. Ellis, D. G. Parker and T. Welton, Chem. Commun.,

1999, 25.31 A. L. Monteiro, F. K. Zinn, R. F. de Souza and J. Dupont, Tetrahedron:

Asymmetry, 1997, 8, 177.32 F. Liu, M. B. Abrams, R. T. Baker and W. Tumas, Chem. Commun.,

2001, 433.33 R. A. Brown, P. Pollet, E. McKoon, C. A. Eckert, C. L. Liotta and P. G.

Jessop, J. Am. Chem. Soc., 2001, 123, 1254.34 R. Noyori, Asymmetric Catalysis in Organic Synthesis, Wiley, New

York, 1994.35 E. G. Kuntz, CHEMTECH, 1987, 570.36 H. Waffenschmidt and P. Wasserscheid, J. Mol. Catal. A: Chemical,

2000, 164, 61.37 J. F. Knifton, J. Mol. Catal., 1987, 43, 65.38 N. Karodia, S. Guise, C. Newlands and J. A. Anderson, Chem.

Commun., 1998, 2341.39 C. Brasse, U. Englert, A. Salzer, H. Waffenschmidt and P. Wassersc-

heid, Organometallics, 2000, 19, 3818.40 P. Wasserscheid, H. Waffenschmidt, P. Machnitzki, K. W. Kottsieper

and O. Steltzer, Chem. Commun., 2001, 451.41 P. W. N. M. van Leeuwen, P. C. J. Kamer, J. N. H. Reek and P. Dierkes,

Chem. Rev., 2000, 100.42 F. Favre, H. Olivier-Bourbigou, D. Commereuc and L. Saussine, Chem.

Commun., 2001, 1360.43 M. F. Sellin, P. B. Webb and D. J. Cole-Hamilton, Chem. Commun.,

2001, 781.44 W. Keim, D. Vogt, H. Waffenschmidt and P. Wasserscheid, J. Catal.,

1999, 186, 481.45 D. Zim, R. F. de Souza, J. Dupont and A. L. Monteiro, Tetrahedron

Lett., 1998, 39, 7071.46 E. Mizushima, T. Hayashi and M. Tanaka, Green Chem., 2001, 3, 76.47 Y. Chauvin, B. Gilbert and I. Guibard, J. Chem. Soc., Chem. Commun.,

1990, 1715.48 Y. Chauvin, S. Einloft and H. Olivier, Ind. Eng. Chem. Res., 1995, 34,

1149.49 Y. Chauvin, H. Olivier, C. N. Wyrvalski, L. C. Simon and R. F. de

Souza, J. Catal., 1997, 165, 275; see also L. C. Simon, J. Dupont andR. F. de Souza, J. Mol. Catal., 1998, 175, 215.

50 B. Ellis, W. Keim and P. Wasserscheid, Chem. Commun., 1999, 337.

51 P. Wasserscheid, C. M. Gordon, C. Hilgers, M. J. Muldoon and I. R.Dunkin, Chem. Commun., 2001, 1186.

52 S. M. Silva, P. A. Z. Suarez, R. F. de Souza and J. Dupont, PolymerBull., 1998, 40, 401.

53 J. E. L. Dullins, P. A. Z. Suarez, S. Einloft, R. F. de Souza, J. Dupont,J. Fischer and A. De Cian, Organometallics, 1998, 17, 815.

54 For a recent review see: I. P. Beletskaya and A. V. Cheprakov, Chem.Rev., 2000, 100, 3009.

55 D. E. Kaufmann, M. Nouroozian and H. Henze, Synlett, 1996, 1091.56 W. A. Herrmann and V. P. W. Böhm, J. Organometal. Chem., 1999,

572, 141; V. P. W. Böhm and W. A. Herrmann, Chem. Eur. J., 2000, 6,1017.

57 A. J. Carmichael, M. J. Earle, J. D. Holbrey, P. B. McCormac and K. R.Seddon, Org. Lett., 1999, 1, 997.

58 L. Xu, W. Chen and J. Xiao, Organometallics, 2000, 19, 1123.59 L. Xu, W. Chen, J. Ross and J. Xiao, Org. Lett., 2001, 3, 295.60 (a) S. T. Handy and X. Zhang, Org. Lett., 2001, 3, 233; (b) L. Sirieix,

M. Ossberger, P. Betzmeier and P. Knochel, Synlett, 2000, 1613.61 J. Howarth, P. James and J. Dai, Tetrahedron Lett., 2000, 41, 10 319.62 W. Chen, L. Xu, C. Chatterton and J. Xiao, Chem. Commun., 1999,

1247; L. Ross, W. Chen, L. Xu and L. Mao, Organometallics, 2001, 20,138.

63 C. de Bellefon, E. Pollet and P. Grenouillet, J. Mol. Catal. A: Chemical,1999, 145, 121.

64 S. Toma, B. Gotov, I. Kmentova and E. Solcaniova, Green Chem., 2000,2, 151.

65 J. Howarth, Tetrahedron Lett., 2000, 41, 6627.66 G. S. Owens and M. M. Abu-Omar, Chem. Commun., 2000, 1165.67 C. E. Song and E. J. Roh, Chem. Commun., 2000, 837.68 L. Gaillon and F. Bedioui, Chem. Commun., 2001, 1458.69 J. A. Boon, J. A. Levisky, J. L. Pflug and J. S. Wilkes, J. Org. Chem.,

1986, 51, 480.70 K. Qiao and Y. Deng, J. Mol. Catal. A: Chemical, 2001, 171, 81.71 C. DeCastro, E. Sauvage, M. H. Valkenberg and W. F. Hölderich,

J. Catal., 2000, 196, 86.72 C. J. Adams, M. J. Earle, G. Roberts and K. R. Seddon, Chem.

Commun., 1998, 2097.73 For reviews see: S. Kobayashi, Synlett, 1994, 689; S. Kobayashi, Eur. J.

Org. Chem., 1999, 15.74 C. E. Song, W. H. Shim, E. J. Roh and J. H. Choi, Chem. Commun.,

2000, 1695.75 C. E. Song, W. H. Shim, E. J. Roh, S. Lee and J. H. Choi, Chem.

Commun., 2001, 1122.76 J. Peng and Y. Deng, New J. Chem., 2001, 25, 639.77 K. K. Laali and V. J. Gettwert, J. Org. Chem., 2001, 66, 35.78 J. Peng and Y. Deng, Tetrahedron Lett., 2001, 42, 403.79 J. L. Scott, D. R. MacFarlane, C. L. Raston and C. M. Teoh, Green

Chem., 2000, 2, 123.80 D. K. Magnusson, J. W. Bodley and D. F. Adams, J. Sol. Chem., 1984,

13, 583.81 S. G. Cull, J. D. Holbrey, V. Vargas-Mora, K. R. Seddon and G. J. Lye,

Biotechnol. Bioeng., 2000, 69, 227; for a related study see: A. G. Fadeevand M. M. Meagher, Chem. Commun., 2001, 295.

82 M. Erbeldinger, A. J. Mesiano and A. J. Russell, Biotechnol. Progr.,2000, 16, 1131.

83 R. Madeira Lau, F. van Rantwijk, K. R. Seddon and R. A. Sheldon, Org.Lett., 2000, 2, 4189.

84 M. J. J. Litjens, A. J. J. Straathof, J. A. Jongejan and J. J. Heijnen, Chem.Commun., 1999, 1255.

85 S. H. Schofer, N. Kaftzik, P. Wasserscheid and U. Kragl, Chem.Commun., 2001, 425.

86 T. Itoh, E. Akasaki, K. Kudo and S. Shirakami, Chem. Lett., 2001,262.

87 K. W. Kim, B. Song, M. Y. Choi and M. J. Kim, Org. Lett., 2001, 3,1507.

Chem. Commun., 2001, 2399–2407 2407