Ionic Liquid-Based Microemulsions in Catalysis

The design and properties of surface-active ionic liquids that are able to form stable microemulsions with heptane and water are presented, and their promise as reaction media for thermomorphic palladium-catalyzed cross-coupling reactions is demonstrated.


■ INTRODUCTION
Microemulsions are thermodynamically stable, isotropic, optically transparent solutions composed of two intrinsically immiscible solvents in the presence of one or more surfactants, and have found wide application for catalytic reactions and processes. 1,2 They are particularly suitable as nanoreactors for many catalytic transformations due to their microstructure, which may vary continuously from spherical to cylindrical, tubular, and bicontinuous oil and water phases. 3 The nanometer-sized oil or water droplets provide highly dynamic nanoreactors, with the interface disintegrating and reforming on a time scale of milliseconds, which can promote favorable reaction kinetics. 4 Moreover, microemulsions can simultaneously co-solubilize large volumes of hydrophilic and hydrophobic compounds, which can overcome the solubilization barriers of the formerly immiscible liquids.
The formation of stable microemulsions between two intrinsically immiscible liquids, e.g., oil and water, requires the presence of amphiphiles to reduce the interfacial tension between the two phases to a very low value. However, microemulsions must not necessarily consist of organic solvents and water, alternative fluids such as ionic liquids have been investigated. 5 Different concepts can be, and have been, realized in this regard, including mostly nonaqueous microemulsions, where water is replaced with room-temperature ionic liquids, for example in the well-studied system {1-butyl-3-methylimidazolium tetrafluoroborate, [C 4 mim][BF 4 ]/toluene/Triton X-100}. 6 Other examples include water−ionic liquid microemulsions, e.g., {H 2 O/1-butyl-3-methylimidazolium hexafluorophosphate, [C 4 mim][PF 6 ]/Triton X-100}, 7−9 or even a microemulsion of two intrinsically immiscible ionic liquids, e.g., 1-octylimidazolium aprotic and protic ionic liquids with the anionic surfactant, 1,4-bis[(2-ethylhexyl)oxy]-1,4-dioxobutane-2-sulfonate (commonly known as docusate or [AOT] − ) 10 all of them stabilized by the addition of one or more conventional surfactants. Less attention has been focused on the use of ionic liquids as amphiphiles themselves, and only a few examples exist where surface-active ionic liquids are used to reduce the interfacial tension of two immiscible solvents to form stable microemulsions. 11 This is surprising, as the surfactant behavior is a natural consequence of the structure of ionic liquids featuring cations with a hydrophilic headgroup and a hydrophobic tail, i.e., the alkyl chain attached to the ions. 12,13 More recently, the properties of catanionic ionic liquids, containing both anions and cations with hydrophobic tails, have proved to be exceptional. 14,15 Surface-active ionic liquids have already been well explored for a number of applications in synthesis, 16,17 catalysis 18, 19 and separations 20,21 rendering them ideally suited for the development of targeted microemulsions, 22−26 and for their separation. 27 In this article, the design and synthesis of surface-active ionic liquids composed of both surface-active cations and anions that are capable of forming stable microemulsions with heptane and water are reported. These microemulsions were applied as reaction media for palladium-catalyzed Suzuki cross-coupling reactions, resulting in high reactivity even at low catalyst loadings, with simple product separation and successful catalyst recycling through thermomorphic catalysis.

■ RESULTS AND DISCUSSION
Ionic Liquid Synthesis. Our strategy for the preparation of ionic liquid-based stable microemulsions relied on the design of ionic liquids that are composed of surface-active imidazolium cations in combination with a dialkylphosphinate anion.
As it is well-known that imidazolium cations are far from innocent when used in transition metal catalysis, we selected two surface-active ionic liquids, namely [C 12 mim][( i C 8 ) 2 PO 2 ] and [C 12 dmim][( i C 8 ) 2 PO 2 ], for this study (Figure 1). Both ionic liquids were prepared according to the standard alkylation-metathesis methodologies in two-steps, initially providing the surface-active chloride ionic liquids [C 12 mim]Cl and [C 12 dmim]Cl. 28 The metathesis with the surface-active anion was performed with the sodium salt of bis(2,4,4trimethylpentyl)phosphinic acid, a cheap mining chemical sold under the brand name Cyanex 272. 29 As the prepared hydrophobic ionic liquids naturally tend to form emulsions, workup and purification after ion exchange was difficult and required the formation of a three-phase system after addition of water and hexane, with the aid of ethanol. This strategy allowed the repeated extraction of the ionic liquids with water to reduce the chloride content, and both ionic liquids were isolated with good purity and yields of 86 and 88%, respectively. Based on their dual surfactant nature, these ionic liquids have the ability to bring hydrophobic and hydrophilic solvents together as stable microemulsions that can form multiphase system phases (Figure 1, inset).
Phase Behavior. Detailed investigations on the phase behavior of water-heptane-ionic liquid mixtures at 25°C via titration and conductivity measurements showed that all four types of Winsor microemulsion systems are present in the ternary phase diagram for the surface-active ionic liquid [C 12 mim][( i C 8 ) 2 PO 2 ], indicating the typical phase behavior found for many mixtures of oil (here heptane), water, and ionic surfactants at variable salinity ( Figure 2). 30−32 According to the Winsor classification, there are four general types of microemulsion systems: At small concentration of a surfactant biphasic system of oil-in-water (o/w) microemulsions form with excess oil can be formed (Winsor I). Alternatively, the ionic liquid-rich water-in-oil (w/o) microemulsion may coexist with the ionic liquid-poor aqueous phase (Winsor II). Type III corresponds to a three-phase system where an ionic liquid-rich middle-phase coexists with both excess water and oil ionic liquid-poor phases (Winsor III), while Winsor type 4 is a singlephase isotropic solution which forms upon addition of higher amount of ionic liquid (Winsor IV). 33 Interestingly, the two-phase o/w microemulsion system (Winsor I) is interrupted by a small and isolated three-phase system with a bicontinuous middle phase (Winsor III). Moreover, a gelatin mesophase was identified via conductivity measurements carried out in order to study the microstructural transformations from a droplet microemulsion to a bicontinuous phase. Based on percolation theory, Clausse et al. 34 have identified different types of microemulsions, i.e., water-in-oil,

The Journal of Organic Chemistry
Article oil-in-water, and bicontinuous microemulsions via conductivity measurements in the optically clear phase region at variable water content. The variation of conductivity as a function of water content was further studied and gave information on the microstructural transformations from a droplet microemulsion to a bicontinuous phase. With increasing water content, the initial nonlinear increase of conductivity indicates the existence of a percolation phenomenon that is attributed to inverse microdroplet aggregation, while the following linear increase is due to the formation of a w/o microemulsion. 35 At a certain water concentration, a sharp decrease in conductivity indicates the formation of a bicontinuous mesophase (see Figure 3).
Catalytic Studies. After evaluation of suitable phase behavior, the application of these novel ionic liquid based microemulsions as reaction media was studied in palladiumcatalyzed cross-coupling processes that are widely used in the pharmaceutical industry for C−C bond formation. 36 The chosen Suzuki reaction, a cross-coupling reaction between aryl halides with boronic acids, 37 typically leads to the formation of substituted biaryl moieties that are present in many pharmaceutically active compounds, herbicides, new materials, polymers, liquid crystals, and ligands. 38 Despite extensive work and progress in ligand and catalyst design, the product separation and catalyst recovery can still provide an obstacle for large scale processing. Recent trends in this area have addressed the application of thermomorphic catalysis for this purpose, relying on temperature-dependent phase behavior for catalyst recycling, as shown in an elegant study by Schomacker and co-workers. 39 Here, the palladium-catalyzed Suzuki cross-coupling reaction between 4-bromoacetophenone and phenyl boronic acid was selected to evaluate the application of ionic liquid-based microemulsions as reaction media for transition metal catalysis ( Figure 4).
Initially, the optimization of reaction conditions for this model reaction was done with respect to base and catalyst loading (Table 1,  (1) Interestingly, the best yields under standard conditions (entries 7 and 9) correspond to largest values of γ with the smallest values of α. If this proves to be a general observation, it will lend itself to design experiment optimizations 40  On the contrast, a decrease in the water content, i.e., a change from α = 0.5 to α = 0.625 did not alter the yield significantly. Eventually, complete conversion and quantitative yield of 4-ethanoylbiphenyl was obtained for both ionic liquids [C 12 mim]-[( i C 8 ) 2 PO 2 ] and [C 12 dmim][( i C 8 ) 2 PO 2 ] at optimized conditions (IL/heptane/H 2 O 0.1/0.6/1, 0.18 mmol K 2 CO 3 ) in short reaction times (1 h) at a low catalyst loading of 0.1 mol% PdCl 2 .
As can be seen from Table 1, the yield of the reaction without ionic liquid in a simple biphasic mixture is around 70% (entry 12). The addition of a small amount of ionic liquid renders the biphasic mixture into a stable triphasic system with an intermediate microemulsion layer (Winsor III), which results in increased yields. In fact, the formation of the nanoreactors in the microemulsion where a polar reagent reacts with an oily substrate is a clear advantage, and results in high reactivity, even at room temperature (entry 8). The impact of microemulsion formation is also visible when comparing results in pure ionic liquid, pure heptane or pure water (entries 13− 15). Yields of the Suzuki reaction performed in a single solvent remained drastically below the results obtained with the microemulsion system; in fact no conversion was observed when the reaction as performed in the pure ionic liquid [C 12 mim][( i C 8 ) 2 PO 2 ] (entry 15). Eventually, a control experiment was performed in the presence of 0.01 mmol bis(2,4,4trimethylpentyl)phosphinic acid, as it is well-known that phosphinates can act as P,O ligands for palladium (entry 16). 41 While a good yield of 86% could be obtained with 10 mol% bis(2,4,4-trimethylpentyl)phosphinic acid, the values remained clearly below the yields obtained in the microemulsion system, thereby emphasizing the high reactivity obtained in the ionic liquid/heptane/H 2 O microemulsion system. Although both ionic liquids were able to form microemulsions, and are suitable reaction media for the chosen Suzuki reaction, [C 12    These findings demonstrate that the role of ionic liquid is not limited to the formation of a suitable reaction media. Instead, [C 12 mim][( i C 8 ) 2 PO 2 ] plays a dual role of surfactant and ligand, resulting in the outstanding reactivity even at room temperature. In case of [C 12 dmim][( i C 8 ) 2 PO 2 ], NHC complex formation between ionic liquid and palladium is clearly impossible, since the C(2) position is blocked with the bulky methyl group. In contrast, we observed the formation of finely dispersed palladium nanoparticles during the reaction that are well stabilized and immobilized in the intermediate microemulsion phase as proven with TEM images (see Figure 6). A similar effect was reported by Zhang et al. 7 who used the H 2 O/ TX-100/1-butyl-3-methylimidazolium hexafluorophosphate ([C 4 mim]PF 6 ) microemulsions as a medium for the in situ preparation of Pd nanoparticles, and several studies in literature focused on the application of Pd nanoparticles in microemulsion as efficient catalytic system for cross-coupling reactions. 44−46 Based on the optimized conditions (α and γ of 0.375 and 0.059, respectively), the substrate scope was expanded to

The Journal of Organic Chemistry
Article include various substituted starting materials with both electron-withdrawing and electron-donating functionalities (Table 2). Excellent isolated yields >90% were obtained for most products with aryl halides or boronic acids with electrondeficient substituents, which is in accordance with literature data for Suzuki couplings. Moreover, the products could be directly isolated by crystallization from the heptane phase. However, in the cases of less reactive chloroarenes, or sterically demanding starting materials, such as 2,6-dimethylbromobenzene or 2,6-dimethylphenylboronic acid, yields remained below expectation. The catalytic system seems to be less effective in C−C bond formation of chloroarenes, indicating that their conversion in Suzuki cross coupling is still a challenge. The preferential solubility of the starting materials, products and catalysts in different microemulsion layers, in combination with the temperature-dependent phase behavior, offers novel opportunities for catalyst recycling. 47 In the chosen reactions, phenylboronic acid and base are soluble in the lower aqueous phase, bromoacetophenone or the product 4-acetylbiphenyl is dissolved in the top heptane phase, whereas the transition metal catalysteither in a well-defined molecular state, or as a nanoparticle dispersionis preferentially located in the intermediate bicontinuous microemulsion layer (Figure 7).
The strategy outlined here for thermomorphic catalysis with catalyst recovery relies on a stepwise shift of phase boundaries. Initially, the triphasic microemulsion system is charged with catalyst and starting materials at room temperature. An increase in reaction temperature to 80°C results in a shift of phase boundaries to two phases, while, in case of [C 12 dmim]-[( i C 8 ) 2 PO 2 ], well-dispersed palladium nanoparticles are simultaneously formed. Once complete conversion is reached, the reaction is cooled to 60°C, resulting in the reformation into a stable triphasic system. At this lower temperature, the top organic layer can easily be separated. Eventually, the product crystallizes from the separated heptane phase at room temperature and can be easily isolated by filtration as colorless crystals in high purity.
Simultaneously, salts formed during the reaction can be removed with the aqueous phase, while the catalyst remains immobilized in the ionic liquid-rich middle phase. A consecutive run can then be started by addition of fresh starting materials and solvents, i.e., 4-bromoacetophenone, phenyl boronic acid, base, heptane, and water.
Again, considerable differences in catalyst recovery were found for the two ionic liquids. In the case of [C 12 mim]-[( i C 8 ) 2 PO 2 ], considerable leaching of the palladium catalyst probably in its NHC complex forminto the top heptane layer was observed, resulting in contamination of the crude product. Consequently, yields decrease drastically in the first three runs due to both catalyst and ionic liquid losses. In contrast, catalyst leaching was not an issue with microemulsions composed of [C 12 dmim][( i C 8 ) 2 PO 2 ]. As shown in Table 3, the same batch of [C 12 dmim][( i C 8 ) 2 PO 2 ] and catalyst were run consecutively for five times with no significant loss in performance. Although traces of ionic liquid were found in the product after the third run, this did not affect the catalytic performance, and a yield >90% could be isolated from [C 12 dmim][( i C 8 ) 2 PO 2 ].

■ CONCLUSIONS
The design of novel surface-active ionic liquids that are able to form stable microemulsions systems has been achieved, and their application as novel reaction media for transition metal catalysis has been demonstrated. High reactivity was observed even at low catalyst loadings, while the temperature-dependent phase behavior allowed simple product separation and successful catalyst recycling through thermomorphic catalysis. While more investigations of the microemulsion nanostructure are clearly required, and currently under way, a fundamental difference was found between carbene-and noncarbene forming ionic liquids, suggesting that the ionic liquid can play a dual role as surfactant and ligand. The exploitation of these tailor-made microemulsion systems for a number of transition metal catalyzed reactions, and also their application in multiple separation problems, are currently ongoing in our laboratories.

■ EXPERIMENTAL SECTION
Commercially available reagents and solvents were used, as received without further purification, unless otherwise specified. All ionic liquids were dried for at least 24−48 h at room temperature or 50°C and 0.01 mbar before use. 1 H and 13 C NMR spectra were recorded on a 400 MHz spectrometer using the solvent peak as reference, and heteronuclear single quantum coherence experiments (HSQC) were used to confirm the peak assignments. Infrared spectra were recorded on a FT IR spectrometer equipped with a single reflection ATR unit. TEM studies were performed on a transmission electron microscope operating at 160 kV. The samples for TEM studies were prepared by precipitation

The Journal of Organic Chemistry
Article of the nanoparticles in ethanol, centrifuged (13 000 rpm), and redispersed in ethanol. A drop of the dispersion (3 μL) was added on a 300-mesh carbon-coated copper grids and the solvent was subsequently evaporated in air.
HPLC analysis was performed on a HPLC unit equipped with a PDA detector under reverse-phase conditions. A reversed phase C18 column (250 × 4.6, 5 μm) was used with MeOH:H 2 O (70:30; 0.1% trifluoroethanoic acid) as solvent and a flow of 0.8 cm 3 min −1 ; detection was at 210 nm, at 30°C column oven temperature and ethyl benzoate as internal standard. Calibration curves were prepared in the range from 2.0 to 0.01 mg cm −3 for starting materials and products.
Ternary phase diagrams were investigated at 25°C for {[C 12 mim]-[( i C 8 ) 2 PO 2 ]:water:heptane}, where [C 12 mim] + is 1-dodecyl-3-methylimidazolium, and ( i C 8 ) 2 P(O)OH is bis(2,4,4-trimethylpentyl)phosphinic acid. The isotropic phase was determined by visual inspection titration of the water:heptane mixture (at different ratios) with [C 12 mim][( i C 8 ) 2 PO 2 ] at 25°C. After adding each drop, the mixture was stirred and left to equilibrate for appropriate times until clear phases were obtained. Conductivity measurements (equilibrated at 25°C) were carried out in order to study the microstructural transformations from a droplet microemulsion to a bicontinuous phase for one phase area at different ratios of [C 12 mim][( i C 8 ) 2 PO 2 ]:heptane (50−100% of ionic liquid). After each water addition, but before each final measurement, the mixture was mixed until stable conductivity measurements were achieved.
General Procedure for the Suzuki Reaction. A screw-capped vial equipped with stirring bar was charged with ionic liquid [C 12 mim]-[( i C 8 ) 2 PO 2 ] (0.1 g), PdCl 2 (0.1 mol%), heptane (0.6 g), and water (1 g). K 2 CO 3 (0.025 g, 0.18 mmol), aryl halide (0.1 mmol), and phenylboronic acid (0.018 g, 0.15 mmol) were added and the three-phase reaction mixture was stirred in an oil bath at 80°C for 1 h. After the reaction was complete, the vial was placed in a water bath and allowed to cool to 60°C, resulting in the reformation of a three-phase system. At this temperature, the heptane phase was separated and left to cool to room temperature. The product crystallized directly from the heptane phase at room temperature and was isolated by filtration.
Recycling Strategy for the Thermomorphic Suzuki Reaction. A screw-cap vial was charged with ionic liquid (1 g), catalyst (0.1 mol% PdCl 2 ), heptane (6 g), and water (10 g). To this were added K 2 CO 3 (0.25 g, 1.8 mmol), 4-bromoacetophenone (0.20 g, 1 mmol), and phenylboronic acid (0.18 g, 1.5 mmol). The mixture was magnetically stirred in an oil bath at 80°C for 1 h, during which a shift of phase boundaries to a two-phase system was observed. The reaction vial was placed in a water bath set to 60°C, which resulted in the formation of three phases within a minute. At this temperature, the upper heptane phase was separated to isolate the product directly. The catalyst is located in the ionic liquid-rich middle phase, which (after removing the lower water phase) remains inside the vial for the next run. Fresh heptane, water, base, and starting materials were added to the vial. The isolated yields for five runs with the same batch of microemulsion middle phase were determined.