Phase-vanishing reactions utilize a perfluorinated solvent as a liquid membrane to separate a substrate and a reagent. Since their introduction less than ten years ago, phase-vanishing reactions have become a valuable alternative to reactions that require a slow addition of a reagent. A variety of experimental designs allow reactions to be carried out under anhydrous conditions, under photolytic conditions, under solvent-free conditions, with a gas as a reagent, and under reflux.
One result of our quest for new synthetic methods is the emergence of a new field of fluorous chemistry.1 Phase-vanishing (PV) reactions have their origin in fluorous triphasic reactions, introduced by Curran et al. in 2001.2 In 2002, Curran et al. reported the first fluorous PV reaction.3 Unlike the triphasic fluorous reaction, the PV method does not utilize any fluorous reactants. Since the initial report, new fluorous solvents and reagents as well as new experimental designs have been introduced. Some of the phenomena that make PV reactions possible include the following: 1) fluorous media are both hydrophobic and lipophobic, hence immiscible with most organic and inorganic media, 2) most reagents have a low solubility in the fluorous phase, which allows them to be transported through it, 3) fluorous media generally have a higher density than most organic media, 4) some reagents have a higher density than fluorous media, and 5) reaction products have little or no solubility in fluorous media.
Thus, when a fluorous phase is mixed with a reagent of higher density, they separate into two phases: the bottom is the reagent the top is fluorous phase (Figure 1 a). When a substrate-containing organic phase is added, one obtains a triphasic PV system consisting of the reagent phase at the bottom, the fluorous phase in the middle, and the organic phase at the top (Figure 1 b). The fluorous-phase screen prevents the mixing of the otherwise miscible reagent and substrate phases. In the course of the reaction the reagent phase at the bottom diffuses through the fluorous phase and vanishes (Figure 1 c). The reaction occurs at the interface of the fluorous and organic phases. Thus, the fluorous phase acts as a bridging solvent that allows slow delivery of the reagent. The reaction proceeds at a moderate rate, which prevents a vigorous and often violent reaction if the two reactants were mixed directly. An alternative extractive mechanism operates in cases in which the reagent is completely insoluble in the fluorous phase (see below).
Curran et al. established that fluorous media are superior to other phase screens, such as acetonitrile, in which organic reactants and products are soluble, and water, through which with diffusion of a reagent is too slow.3 The most recent developments include a quadraphasic/tetraphasic4 PV reaction,5 a photochemical PV reaction,6 a stacked PV reaction,7 a PV reaction on neat substrates,6, 8–10 and a PV reaction with polytetrafluoroethylene (PTFE, Teflon) as a phase screen.11, 12 Two reviews summarize developments up to 2008.13 A study of miscibilites and solubilities of fluorous/organic phases has been reported by Curran et al.14
Value of PV Reactions
Phase-vanishing reactions are a new and alternative synthetic technology. A typical synthesis involves the preparation, isolation, and purification of a product as separate steps. Any method that improves or combines these steps is welcome.
The PV method presents several advantages over conventional methods for the controlled delivery of reactants. With the PV method, exothermic reactions normally carried out at low temperatures (typically −78 to 0 °C) can both be conducted at ambient temperature and without an expensive delivery system, such as a syringe drive. The PV method also allows for reactions that otherwise would be too vigorous without a solvent to be carried out on neat reagents.6, 8–10 These features simplify the setup and workup of a reaction, reduce costs, and offer greater versatility than conventional methods. Sometimes the PV method may be the best or only method to accomplish certain transformations, such as the anti addition of bromine to acyclic (Z)-alkenes7, 12 or the preparation of β-bromo-γ-butyrolactone.11
While this method has some advantages over conventional methods, it is not without its drawbacks. One of the most common fluorous phase screens, fluorinert liquid FC-72 (a mixture of perfluorohexanes), has a high global warming potential (GWP ≈10 000) and an atmospheric lifetime of 3200 years.15 Although, in a PV reaction, FC-72 is typically used in a closed system and can be recycled. Since it has a low boiling point, one should be careful to minimize losses. In addition, some solvents, reagents, and reaction products exhibit some solubility in it,9, 10, 14, 16 which limits the recycling ability. Improvements have been realized through the use of heavier fluorous liquids, such as Galden HT-135,5 a polymeric perfluoroether, and PTFE tape.11, 12
Triphasic PV reactions: A typical PV reaction setup is shown in Figure 1.3 The reaction rate can be controlled by the amount of fluorous-phase screen and the rate of stirring: a greater depth of the phase screen results in a slower reaction, while an increased rate of stirring results in a faster reaction. If a very slow rate is desired, the reaction can be performed without any stirring. Very fast stirring is not recommended as it may directly mix the top and bottom phases.17
Starting with the initial report by Curran and co-workers, brominations have been among the most extensively studied reactions (Scheme 1).3, 5–7, 9–12 They are simple to perform and easy to monitor. The products are known and easy to characterize. In a solvent-free modification of that design (Figure 2), the top phase is a neat substrate. The reaction worked well even when the reaction product was a solid.
Extractive PV reactions: Verkade and Jana designed a PV reaction in which a solvent denser than the fluorous phase was utilized as the bottom phase.18 This experimental setup is an alternative to the U-tube method in which the reactants are less dense than the fluorous phase (see below). Furthermore, it allows the use of reagents that are completely insoluble in a fluorous media and cannot be transported through it. 1,2-Dibromoethane was used as a solvent of highest density to dissolve reagents, such as dimethyl sulfate or m-CPBA, and FC-72 was used as the phase screen. The substrate, dissolved in dichloromethane, was the top layer. It is interesting that the top layer vanished in the course of the reaction (Figure 3).
Curran and Werner have shown that the reaction proceeds by an extractive PV mechanism as opposed to a diffusion-controlled delivery of the reagent.16 In addition, the conditions under which either the top or the bottom phase can be the vanishing phase were identified. Extractive PV reactions allow slow addition of an entire phase rather than just a reagent to another phase without the need for an addition funnel or a syringe drive.
Tetraphasic PV reactions: Ryu et al. devised a tetraphasic reaction design in which water was added as a fourth “acid scavenger phase” to carry out the bromination of ketones (Figure 4).4, 5 Galden HT-135, a polymeric perfluoroether, was utilized as a phase screen. It is superior to low-molecular-weight fluorous solvents (e.g., FC-72) because it is less expensive and less volatile. In addition, diffusion of bromine through it proceeded at a faster rate. Galden HT-135 has good phase-screen properties. However, as it is a commonly used refrigerant, there is currently a world-wide shortage of it. The authors used chloroform as the organic solvent and either bromine or PBr3 as the bromination reagent. Interestingly, water-sensitive PBr3 gave good results even though aqueous K2CO3 was the top phase. Following the report by Ryu et al., other authors have reported various tetraphasic PV reactions.6, 7, 11
While the initial report focused on trapping HBr because it was an undesired noxious by-product, tetraphasic reaction conditions gave very good results in some bromolactonization reactions compared with traditional experimental conditions (Scheme 2).10
Stacked reactor: The “stacked reactor” is another variation on the tetraphasic PV reaction design. Weiss et al. described the use of a new ionic liquid, tridecylmethylphosphonium tribromide (1P10Br3), in the bromination of alkenes.7 In a PV variation of the procedure, perfluorohexane (FC-72) was used to separate the bottom bromine layer from the 1P10Br3 ionic-liquid layer (Scheme 3). The least dense hexadecane/substrate layer was on top. The reagent 1P10Br3 was produced in a reaction between tridecylmethylphosphonium bromide (1P10Br) and bromine. It was regenerated in situ with bromine that diffused from the bottom layer. Thus, a PV reaction is utilized to generate the bromination reagent. The authors were able to carry out several sequential reactions without a workup. A drawback of this design is its rather long reaction times (5–20 d).
The authors used a test tube with a side arm so that bromine could be replenished in the course of the reaction (Figure 5 a). An alternative is to employ a syringe equipped with a Teflon needle that reaches the bottom phase (Figure 5 b).
Anhydrous PV reactions: Ryu’s group has reported several anhydrous reactions carried out under PV conditions.19 Since the fluorous phase dissolves little or no water, some reactions that require anhydrous conditions can be carried out without special precautions. Thus, a reaction that utilizes PBr3, a moisture-sensitive reagent, has been successfully carried out in the presence of water as the fourth phase (see above).5 A recent report includes a successful Grignard reaction (Figure 6).20
PV reactions on reactants lighter than the fluorous phase: The U-tube was used in the original fluorous triphasic reactions (Figure 7 a).2 Nakamura et al. introduced a triphasic U-tube method for reactants lighter than the fluorous phase (Figure 7 b).21 Thionyl chloride and phosphorus trichloride were used as reagents in the chlorination of several alcohols.
Nakamura et al. also designed a multislit apparatus with an open bottom (Figure 8) that allowed for several reactions to be conducted simultaneously. A similar parallel synthesis design, with the reagent as the bottom phase, has been used by Ryu et al. in the PV Friedel–Crafts acylation of aromatic compounds.22 No mixing of the different reaction products has been reported. However, since some organic compounds are slightly soluble in fluorous solvents, one should be on alert for such a possibility.
An alternative to the U-tube design is the use of a tube, or tubes in case of a parallel reaction design, immersed in a vial or a test tube (Figure 9). One reactant is placed inside the tube and the other one outside. By changing the depth of the tube in the fluorous phase, one can adjust the distance through which a reagent has to diffuse and thereby control the rate of the reaction.
PV reactions on reactants heavier than the fluorous phase: A PV reaction involving two reactants heavier that the fluorous phase screen can be carried out in an inverted U-tube.23 However, an inverted U-tube may be difficult to fill and stoppered inverted arms may develop leaks. An alternative is a reaction in two vials (Figure 10). We carried out a reaction between magnesium metal and bromine to prepare anhydrous MgBr2. Bromine was added to the smaller vial and overlaid with fluorous phase (FC-72). It was placed inside a larger vial, magnesium was added next to it and the vial was filled up with FC-72.
Solvent-free PV reactions: PV reactions were successfully carried out on neat substrates.6, 8–10 Slow diffusion through the fluorous-phase screen provided the slow delivery of the reagent, which allowed the reaction to be carried out under solvent-free conditions. Aromatization, isomerization, and halogenation of neat substrates under PV conditions gave the expected products in good to excellent yields (Scheme 4).9
The solvent-free PV method is also effective in halolactonization reactions. Traditionally, a halolactonization utilizes Br2,26 I227 or ICl28 in a mixture of an aqueous solvent and an organic cosolvent in the presence of a base, such as sodium bicarbonate.24, 25 PV halolactionization of neat substrates avoids use of solvents, other than the fluorous-phase screen, and basic reaction conditions.10 Both γ- and δ-alkenoic acids and the corresponding methyl esters were suitable substrates. The reaction worked well on solid and liquid substrates, and the products were obtained in good to excellent yields (Scheme 5). Bromine and iodine monochloride were suitable electrophiles. Iodine gave poor results. In the halolactonization of 4-pentenoic acid, reaction in dichloromethane gave iodolactone contaminated with chlorolactone. A PV reaction gave essentially pure iodolactone (only ≈2 % of chlorolactone) (Scheme 6).
A further development of the PV method was the integration of a single-phase reaction, in which a reaction occurred in the organic phase, with a PV reaction into a tandem single-phase–PV (SP-PV) reaction.9 Thus, two reactants, placed in the top phase, in a fast reaction, afforded the intermediate, which in a subsequent PV reaction reacted with the reagent from the bottom phase to give the final product (Figure 11).
A Diels–Alder reaction was performed as the single-phase reaction and the reaction of the resulting Diels–Alder adduct with a halogen as the PV reaction. Diels–Alder reactions under solvent-free conditions (SFC) were suitable for the SP-PV method as long as the resulting Diels–Alder adduct was a liquid. A halogen reagent was placed at the bottom of the reaction vessel, a phase screen (FC-72) was added and Diels–Alder reactants were placed on the top of the phase screen. Thus, a relatively fast, SFC Diels–Alder reaction occurred first in the top phase and was followed by a slower PV reaction between the resulting Diels–Alder adduct and the halogen reagent. Even though dienes are highly reactive towards halogens, a sufficient depth of the FC-72 phase screen prevented the reaction between the halogen and either the starting diene or the dienophile.
Photochemical PV reactions: Iskra et al. reported the first photochemical PV reaction.6 The reaction mixture was irradiated with a 40 W incandescent light bulb to accomplish a photolytic bromination of substituted alkylbenzenes (Figure 12). The results were compared with those obtained by conducting experiments under conventional (non-PV) conditions. Besides investigating the reaction in a variety of solvents, the authors successfully carried out the reaction on a neat substrate. Addition of a fourth aqueous phase as an acid trap improved conversion and selectivity. The authors also made an ingenious stirring bar for PV reactions, which we found easy to reproduce. To stir both the reagent (bottom) and the substrate (top) phases without mixing the three phases, they employed two stirring bars. An ordinary stirring bar was used to stir the bottom phase, and a specially designed bar was used to stir the top phase. It consisted of a hollow glass bubble with an iron wire inside. The stirring bar was light enough to float on the top of the fluorous phase and it stirred simultaneously with the stirring bar in the bottom phase. An alternative, less elegant, solution would be to use a mechanical stirrer.
A similar reaction design was utilized by Ryu and co-workers to carry out an addition reaction in situ (Scheme 7). Hydrogen bromide was generated in a photolytic reaction between bromine and isooctane and, in an apparent free-radical addition to 1-dodecene, afforded 1-bromododecane in a high yield.29 A PV photolytic bromination accompanied by HBr addition to an alkene has been adapted into a laboratory exercise for undergraduate organic chemistry students.30
PTFE as a phase screen: A recent development in PV reaction design is the use of PTFE as a phase screen.11, 12 Ryu et al. showed advantages of using heavier, polymeric, fluorous liquid (Galden HT-135) over lighter ones such as FC-72.5 Earlier, Gladysz and Dinh described the use of PTFE as a catalyst support.31 In a PV-PTFE reaction setup, the substrate in a solvent was placed in a reaction vessel, such as a vial, a test tube, or a flask.11 The reagent was placed in a delivery vessel, such as a glass tube, which was sealed on both ends with PTFE tape. This tube was then inserted into the reaction vessel so that both reactants were in contact with the PTFE phase screen (Figure 13). In the course of the reaction the reagent diffuses through the PTFE tape into the substrate phase.
PTFE tape is inexpensive, easy to use, and the PTFE-sealed tube may be reused. In addition, there is no limitation concerning the density of a phase because the heavier phase can be in the top layer. Thus, a reagent such as bromine, with a density higher than that of the phase screen, can be the top phase, which made a reaction under reflux possible (Figure 14). A slow addition of bromine by means of PTFE phase screen enabled the bromination to proceed smoothly under the vigorous conditions necessary for a simultaneous esterification reaction. A variety of other experimental designs are possible, including sequential, tandem, and various tetraphasic reactions (Figure 15).11 In one variation of a tetraphasic reaction setup, a test tube was filled with a fluorous phase (FC-72), which was overlaid with a phenol solution in water. In an ordinary triphasic PV bromination of phenol the product was obtained in a high yield, but it was impure. In a tetraphasic reaction, a bromine-filled PTFE tube was immersed into the aqueous phase (Figure 15 a). Pure 2,4,6-tribromophenol formed a light porous solid, which floated on FC-72, while heavy liquid impurities settled on the bottom (Figure 15 b).
While the PTFE-PV method generally gave good results, it was not suitable in some situations. PTFE did not present a barrier for TiCl4, since it easily passed through it. Iodine monochloride as a reagent gave poor results. Finally, precipitated product sometimes formed an impermeable barrier on the PTFE tape (e.g., formation of dibromostilbene). In an attempt to address some of the shortcomings, solvent-free PTFE-PV reactions were developed.12
PV reaction with a gas/vapor as the vanishing phase: Iskra et al.utilized a U-tube in a reaction in which chlorine gas was the vanishing phase (Figure 16 a).8 Chlorine, under a slight pressure, was applied to one arm of the U-tube and the other arm was utilized as a reaction phase. In the course of the reaction, chlorine dissolved in the fluorous phase and was then extracted by the reaction phase in which it has higher solubility. Studies of solubility of halogens32 and other gases33 in fluorous media have been reported by Costa Gomes et al. This procedure gave good results both on neat liquid alkenes and in solution.
An alternative reaction design was applied in a reaction of dry HBr with alkenes (Figure 16 b).9 The reaction was carried out under solvent-free conditions. The fluorous phase allowed gas to be delivered without the delivery tube becoming clogged by the formation of solid reaction product, which was the case when gas was introduced directly into the organic phase. Interestingly, bubbling gaseous HBr through neat cis-stilbene, overlaid on FC-72, resulted in isomerization into trans-stilbene without any addition of HBr. Apparently, isomerization was a very fast process and solid trans-stilbene was unreactive.9 For addition of HBr to occur, an alkene had to be dissolved in a solvent such as acetic acid.9
In a solvent-free PV reaction with PTFE tape as the phase screen, a neat substrate was placed in a flask and a PTFE sealed tube filled with a reagent was inserted into the vapor phase above it (Figure 17 a). When a substrate is placed in the reaction vessel, the reagent rapidly diffuses out of the tube as its vapors are consumed.12 The procedure generates little or no waste, while providing the reaction products in high yield and purity.
In the case of the bromination of cis-stilbene, bromine vapors were initially rapidly consumed (Figure 17 a and b). For an anti-addition to cis-stilbene, it is essential that the reaction is conducted in a small flask (10 mL for a 2 mmol scale reaction) and that the delivery tube is close to the surface of the substrate. Larger reaction flasks and more distant delivery tubes resulted in a slower reaction and increased syn-addition. The end of the reaction was indicated by the presence of an excess of bromine vapors (Figure 17 c). At that point, the bromine delivery tube was replaced with a reaction workup tube filled with aqueous thiosulfate (Figure 17 d).
With this experimental setup, reactions involving iodine monochloride worked well (Scheme 8). There were no problems associated with a solid product coating the PTFE tape and the corresponding iodolactones were obtained in good yields.
Very recently, Ley et al. reported the use of PTFE-2400, a copolymer of PTFE and perfluorodimethyldioxolane, as a semipermeable membrane to introduce ozone into a reaction mixture.34
Summary and Outlook
A variety of reaction designs have been developed over a relatively short period of time since the introduction of fluorous triphasic reactions in 2001 and fluorous PV reactions in 2002. The variety of the reported transformations indicates that this is a potentially general and inexpensive method for the slow addition of a reagent or a phase. While some of the current PV processes may be limited by the cost of the fluorous solvent (FC-72, FC-77), availability (Galden HT-135), reaction rate, or reaction scale, alternative designs are continually being introduced. As many of the procedures have been reported only as preliminary communications, they should not be considered to be optimal. Although there are some drawbacks, the procedures are relatively green, or at least greener, than conventional methods. Furthermore, in some instances, a PV method may be the best way to carry out a reaction.
While the focus of this article has been on the practical and technical aspects of PV reactions, one should consider the mechanistic implications of utilizing fluorous phases. For instance, a recent report by Gladysz and da Costa on fluorous-phase-transfer activation of catalysts illustrates the mechanistic importance of fluorous phases,35 which may provide opportunities for the development of PV reactions that include reagents (fluorous or not) activated by the fluorous phase. This may result in blurring the boundaries between PV and fluorous reactions. Additional research of physicochemical aspects of PV reactions is likely to provide valuable insights into their nature. Finally, utilization of PV reactions in the synthesis of natural products or in the design of novel reactions will be a great advance in this area. The authors believe that this may be the next frontier in the development of PV methods.