Ionic liquid solvents: the importance of microscopic interactions in predicting organic reaction outcomes

Introduction

Ionic liquids are salts that are typically composed of a bulky, charge diffuse, organic cation and an anion that can be either organic or inorganic [1], [2]. The charge diffuse nature of the constituent ions of these salts results in them having melting points much lower than typical inorganic salts [1], [2]. Ionic liquids have generated interest as solvents due to a number of attractive properties, such as low vapour pressure and flammability [3], [4], [5], [6], and the ability to dissolve a wide range of solutes [7], [8], [9]. In addition, it is estimated that there are >10¹⁴ anion and cation combinations [2] that can be used (before mixtures containing multiple ions are considered [10], [11], [12], [13]), allowing the physical and chemical properties of the resultant ionic liquid to be tuned and different functionality to be introduced to the solvents [3], [5], [14], [15], [16]. Further, it has been widely demonstrated that ionic liquids can affect the outcome of organic processes, with many reactions proceeding more readily and/or more selectively in ionic liquids compared to traditional organic solvents [17], [18], [19], [20].

In order for ionic liquids to be more widely used as solvents and to be more attractive to chemists, a more comprehensive understanding of how and why ionic liquids cause changes to reaction outcome, when compared to molecular solvents, is needed. Our group, along with others [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], has been developing the tools necessary for predicting the effects of using these solvents on the outcomes of typical reactions (for a previous summary in this journal see Ref. [20]). However, the predictive framework still remains somewhat limited.

As background we first summarise what is currently understood about the origin of solvent effects in ionic liquids. Our group has previously studied the effects of ionic liquids on the outcomes of a range of processes including unimolecular [39], [40], [41], [42], bimolecular [41], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52] and aromatic substitutions [51], [53], [54], and cycloaddition processes [55], [56]. The information gained has demonstrated the importance of (i) the proportion of the ionic liquid in the reaction mixture (for a summary, see also Ref. [20]) and (ii) the interactions of components of the solvent with both the starting material(s) and transition state for a process. Importantly, there is the possibility of modifying such interactions by changing the structure of the ionic liquid [45], [46], [47], [48], [49], [51], [54] (with the potential to design novel ionic liquids as a result [51]) and that these interactions may vary with the structure of the reagents [52].

The work described herein summarises recent developments in our understanding of ionic liquid solvent effects, with particular emphasis on understanding the interactions between the ionic liquid components and the species along the reaction coordinate. The importance of understanding the subtle balance between the opposing enthalpic and entropic effects associated with these interactions will be demonstrated. In order to do so, the interactions considered will be broken up into several classes. Firstly, interactions between the ionic liquid cation and the lone pair on the nucleophilic reagent will be considered, followed by discussion of the interactions that exist between the ionic liquid components and either charges or quadrupole moments on the species along the reaction coordinate.

Cation – lone pair interactions

There has been much previous work focussing on the effect of ionic liquid solvents on bimolecular substitution processes [26], [41], [43], [44], [45], [47], [57], [58], [59], [60], [61], with a number of these studies concluding that interaction between the cation of the ionic liquid and the nitrogen lone pair on the nucleophilic reagent was the main interaction affecting reaction outcome [41], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52]. To begin, the main outcomes of this previous work will be briefly presented, followed by discussion of more recent work examining the ionic liquid solvent effects on a related condensation process.

The Menschutkin reaction between benzyl bromide 1 and pyridine 3 (Scheme 1) was previously investigated in the ionic liquid 1-butyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide ([Bmim][N(CF₃SO₂)₂], 6) through a series of kinetic studies [43]. As the mole fraction of [Bmim][N(CF₃SO₂)₂] 6 in the reaction mixture (using acetonitrile as the co-solvent) was increased it was found that there was an increase in the bimolecular rate constant (k₂), with the main changes in k₂ occurring by χ_IL=0.2 and a maximum value approached at higher mole fractions (Fig. 1) [52]. A comparable trend in the rate constant when varying the proportion of [Bmim][N(CF₃SO₂)₂] 6 in the reaction mixture has also been observed for the related reaction between benzyl chloride 2 and pyridine 3 [41].

Scheme 1:

The bimolecular substitution reaction between either benzyl bromide 1 or benzyl chloride 2 and pyridine 3 to give the salts 4 and 5, respectively [41], [52]. Shown also is the ionic liquid [Bmim][N(CF₃SO₂)₂] 6.

Fig. 1:

The changes in the bimolecular rate constant (k₂) for the S_N2 reaction between benzyl bromide 1 and pyridine 3 as the mole fraction of [Bmim][N(CF₃SO₂)₂] 6 in the reaction mixture with acetonitrile was increased [52]. Reproduced from Ref. [52] with permission from the Centre National de la Recherche Scientifique (CNRS) and The Royal Society of Chemistry.

The activation parameters determined for the reaction between species 1 and 3 indicated that when using [Bmim][N(CF₃SO₂)₂] 6 (χ_IL=0.85) there was an increase in both the enthalpy of activation (an ‘enthalpic cost’) and the entropy of activation (an ‘entropic benefit’) when compared to acetonitrile [43]. A similar trend in the activation parameters was observed for the reaction of species 2 and 3 [41]. This work [41], [43], in combination with a series of deconvolution and computational studies [44], showed that the ionic liquid cation was interacting with the nitrogen lone pair on the nucleophile 3, stabilising this reagent and thus increasing the enthalpic barrier when compared to acetonitrile. On moving to the transition state, where this lone pair takes part in the reaction and is therefore no longer free to interact with the cation, there was an increase in the entropy of the system due to release of the ionic liquid cation. The entropic benefit associated with the cation – nucleophile interaction is larger than the enthalpic cost, resulting in the increase in k₂ observed when using [Bmim][N(CF₃SO₂)₂] 6.

Given the importance of the cation – nucleophile interaction identified above, the effect of changing the steric and electronic nature of the ionic liquid cation had on the reaction of benzyl bromide 1 and pyridine 3 was then investigated [45]. It was reasoned that the more charge dense and accessible the cationic centre, the greater its interaction with the lone pair on the nucleophile 3, resulting in an increase in both the enthalpy and entropy of activation for the process. This hypothesis was tested using a number of different ionic liquids, including those shown in Fig. 2, as the cation of these salts cover a range of extents of charge localisation at, and steric congestion about, the cationic centre [45].

Fig. 2:

Ionic liquids 6–10 featuring cations with different extents of charge localisation and steric hindrance about the charge centre.

It was found that the more accessible the charged centre on the cation of the ionic liquid, the larger the rate constant of the reaction of benzyl bromide 1 and pyridine 3 (Scheme 1) [45]. While it could be generally concluded that the rate enhancements are a result of increased interactions with the nucleophile 3, the measured activation parameters were very similar in the different ionic liquids, with many the same within experimental uncertainties [45]. These similarities made it difficult to unambiguously identify the microscopic origin of the rate constant changes seen, and hence it was difficult to validate the above predictions about the importance of the magnitude of the cation – nucleophile 3 interaction.

To further investigate this concept, recent work sought to investigate a similar bimolecular process that would allow this hypothesis to be further examined. The specific reaction examined was the condensation (addition-elimination) reaction between 4-methoxybenzaldehyde 11a and hexan-1-amine 12 (Scheme 2) [46], [49], [50]. As the rate-determining step for this process involves nucleophilic addition of the amine 12 to the aldehyde 11a, this reaction is comparable to the S_N2 reactions of the benzyl halides 1 and 2 with pyridine 3 discussed above [20], [41], [43], [44], [45], [48]; both cases involve a bimolecular rate-determining step, relatively limited charge separation in the transition state and a nitrogen containing nucleophile. Importantly, hexan-1-amine 12 has a more accessible lone pair than pyridine 3; hence it was anticipated that the effect of varying the nature of the cation on the cation – nucleophile interaction would be more marked for this case.

Scheme 2:

The condensation reaction between benzaldehydes 11 and hexan-1-amine 12 to give the imines 13 [46], [49], [50.]

Considering the similarity of these reactions, it was predicted that an ionic liquid solvent would affect reaction outcome in a similar manner for both the S_N2 and condensation processes. As such, initially the changes in reaction outcome when using [Bmim][N(CF₃SO₂)₂] 6 for both the S_N2 and condensation processes will be compared. Interestingly, a similar trend in k₂ as the mole fraction of [Bmim][N(CF₃SO₂)₂] 6 in the reaction mixture with acetonitrile was varied was observed for both the S_N2 reaction of benzyl bromide 1 and pyridine 3 (Fig. 1) and the condensation reaction between 4-methoxybenzaldehyde 11a and the amine 12 (Fig. 3). Interestingly, the rate enhancement observed when using the ionic liquid 6, relative to acetonitrile, was more significant for the reaction of species 11a and 12 [46] than that seen for the reaction between species 1 and 3 [52] (and species 2 and 3 [41]).

Fig. 3:

The changes in the bimolecular rate constant (k₂) for the condensation reaction between 4-methoxybenzaldehyde 11a and hexan-1-amine 12 as the mole fraction of [Bmim][N(CF₃SO₂)₂] 6 in the reaction mixture with acetonitrile was increased [46]. Reproduced from Ref. [46] with permission from The Royal Society of Chemistry.

For the reaction of species 11a and 12 the changes in k₂ observed when using [Bmim][N(CF₃SO₂)₂] 6 (χ_IL=0.85), relative to acetonitrile, were suggested to be due to an interaction between the [Bmim]⁺ cation and the nitrogen lone pair on the nucleophile 12 [46]. This cation – lone pair interaction resulted in an enthalpic cost and entropic benefit, relative to acetonitrile, analogous to that observed for the S_N2 processes described above [41], [43]. This observation suggests that not only are the changes in k₂ as the proportion of the ionic liquid 6 in the reaction mixture is varied similar to that previously reported, but the microscopic origin of these effects are the same. These results are significant as it is the first time that the effect of an ionic liquid on the rate constant of an organic process could be readily predicted by considering the results of previous investigations.

An interesting point to note is that for both the S_N2 [41], [52] and condensation [46] reactions the largest changes in the rate constant occurred at lower ionic liquid concentrations (χ_IL<ca. 0.3), with little change in k₂ on moving to higher mole fractions of [Bmim][N(CF₃SO₂)₂] 6 in the reaction mixture (Figs. 1 and 3). Through a combination of both the kinetic analyses on these processes [41], [46], [52] and a series of small and wide angle X-ray scattering experiments on [Bmim][N(CF₃SO₂)₂] 6/acetonitrile mixtures [62], it has been proposed that this trend arises from the increase in solvent structuring on moving to higher χ_IL. That is, the entropic advantage of removing the coordination of the nucleophile with the cation becomes less significant as the ordering of the solvent itself increases with increasing χ_IL, therefore the entropic effect plateaus at higher χ_IL.

It is also of interest to consider whether the magnitude of the cation – nucleophile 12 interaction could be controlled by changing the constituent ions of the ionic liquid solvent. As described earlier, it was reasoned that by using either a more charge diffuse or a more sterically hindered cation, the magnitude of the cation – nucleophile 12 interaction should be decreased, and that this affect will be more pronounced for hexan-1-amine 12 as the lone pair is more accessible than that on pyridine 2. This was shown to be the case, with a systematic decrease in the activation parameters for the reaction between 4-methoxybenzaldehyde 11a and hexan-1-amine 12 (Scheme 2) as the charge localisation and steric hindrance on the cation was varied (Fig. 4), suggesting that the cation – nucleophile 2 interaction was decreasing as predicted [46].

Fig. 4:

Top: ionic liquids featuring cations with different extents of charge localisation and steric hindrance about the charge centre. Bottom: ionic liquids featuring anions with differing Kamlet–Taft hydrogen bond accepting ability. The arrow indicates a decrease in the magnitude of the cation – nucleophile 12 interaction.

The kinetic analyses on the condensation reaction between species 11a and 12 was then extended to investigate the effect of changing the anion of the ionic liquid on the activation parameters of this process. When using more coordinating anions (that is, those with higher Kamlet-Taft hydrogen bond acceptor ability [63], [64]), there is a stronger cation – anion interaction and hence the cation is less available to interact with the nucleophile 12. This also resulted in a systematic decrease in both the enthalpy and entropy of activation as the cation – nucleophile 12 interaction was reduced (Fig. 4) [49]. The effect of varying the anion of the ionic liquid on the activation parameters of the S_N2 reaction between benzyl bromide 1 and pyridine 3 has also been recently examined, where it was found that there was no clear trend in the activation parameters as the Kamlet–Taft hydrogen bond acceptor ability of the anion was varied [48]. This further reinforces the concept that the condensation reaction species 11a and 12 (Scheme 2) is affected in a more rational manner when changing the ions of the ionic liquid solvent, when compared to the S_N2 reaction of species 1 and 3; this is likely due to the more accessible nucleophilic centre on hexan-1-amine 12, relative to pyridine 3.

Overall, analysis of the effect of a number of ionic liquids on the reaction between species 11a and 12 showed for the first time that by altering the component ions of the ionic liquid solvent, rational changes in the main interaction affecting reaction outcome could be introduced, resulting in predictable changes in the activation parameters of the process. This leads to the potential to design ionic liquids to control reaction outcome [51].

For all of the ionic liquids shown in Fig. 4, the rate constant for the reaction between 4-methoxybenzaldehyde 11a and hexan-1-amine 12 was increased, relative to acetonitrile, when an ionic liquid was present in the reaction mixture. However while the activation parameters varied in a predictable fashion as the ions of the solvent were varied, the changes in the rate constant were less systematic, with no clear trend observed when either the cation [46] or anion [49] of the ionic liquid was varied. This demonstrated that predicting the gross changes in the activation parameters is possible, but understanding the balance between the competing enthalpic and entropic effects (and hence the overall change in the observed rate constant) is much more difficult. This outcome suggested that while the cation – nucleophile 12 interaction appears to be the most significant interaction affecting reaction outcome, other effects also contribute and are causing subtle changes in the activation parameters that affect the delicate balance between the enthalpic and entropic effects that determine changes in the rate constant.

One such example of these additional effects was revealed when examining the effect of the ionic liquids liquid 6, 8 and 9 (Fig. 4) on the reaction of species 11a and 12 (Scheme 2) [46]. Previous work has shown that methylation of the imidazolium ring of the [Bmim]⁺ cation increases the ordering of the ionic liquid, causing the salt 8 to be more ordered than the ‘parent’ ionic liquid 6 [65]. This increase in solvent ordering is proposed to cause greater organisation of the ionic liquids 8 and 9 about the nucleophile 12 than would be expected based on the magnitude of the cation – nucleophile 12 interaction. This results in the entropy of activation becoming less negative in the salts 8 and 9, causing k₂ for the reaction of species 11a and 12 to be higher than expected in the presence of the ionic liquid [46]. Overall, when changing from the ionic liquid 6 to 8 to 9 the energetic effect of the cation – nucleophile 12 interaction was altered by differences in the ordering of each ionic liquid, causing unanticipated changes in k₂ for the ionic liquids 6, 8 and 9.

In summary, recent kinetic analyses on the reaction between 4-methoxybenzaldehyde 11a and hexan-1-amine 12 has demonstrated that it is possible to predict the effect of an ionic liquid solvent on reaction outcome by carefully considering the likely interactions between the solvent and both the starting material(s) and the transition state. For this case, considering previous work [41], [43], [44], [45], it was predicted that interaction between the ionic liquid cation and the nitrogen lone pair on species 12 would be the main interaction affecting reaction outcome, and this was shown to be the case. It was also demonstrated that it is possible to control the magnitude of this interaction through changing the constituent ions of the solvent. However, it was demonstrated that increasing the magnitude of this cation – lone pair interaction does not necessary increase the rate constant of the reaction of species 11a and 12, as it is difficult to predict the balance between the enthalpic and entropic effects, which contribute in opposing ways to the Gibbs energy of activation (and hence the rate constant). It is proposed that there are also other more subtle effects that are contributing to the overall change in activation energy, and understanding these additional contributing factors is necessary when attempting to predict the changes in the rate constant when varying the ions of the ionic liquid solvent.

Ionic liquid – charge interactions

In the above section it was shown that interaction between an ionic liquid cation and a nitrogen lone pair on the nucleophilic reagent can increase the rate constant of a bimolecular process. However for bimolecular processes that involve a charged nucleophile (for example, a chloride anion) a different type of interaction is known to affect the reaction outcome: the electrostatic interaction between the ionic liquid cation and the anionic nucleophile [28], [30], [66], [67]. This strong cation – charged nucleophile interaction results in a significant enthalpic cost when using an ionic liquid solvent (χ_IL>0.7), relative to dichloromethane. A small entropic benefit was also observed, although the entropic advantage of breaking the cation – nucleophile interaction was minor due to ordering of the ionic liquid about the charge separated transition state. Overall, the reaction proceeded slower in ionic liquids than in dichloromethane due to the significant enthalpic cost associated with the cation – charged nucleophile interaction. For this case, the stronger cation – charge interaction resulted in a decreased rate constant, while the weaker cation – lone pair interactions discussed above caused a rate enhancement, relative to molecular solvents.

The importance of interactions between the ionic liquid and charged species along the reaction coordinate has also been demonstrated for unimolecular nucleophilic substitution (S_N1) reactions of neutral species, as these reactions involve considerable charge separation in the transition state [68], [69]. For the S_N1 reaction of the aliphatic substrate (R)-3-chloro-3,7-dimethyloctane 18 (Scheme 3), use of mixtures containing high proportions of [Bmim][N(SO₂CF₃)₂] 6 resulted in a decreased rate constant relative to methanol (Fig. 5) [39], [40]. This decrease in k₁ was shown to arise from the ionic liquid 6 ordering about the relatively charge-separated transition state. While this stabilisation of the incipient charges in the transition state is enthalpically favourable, the entropic cost associated with the increase in ordering about the transition state is more significant [40]. Overall, the outcome is that this ionic liquid – charge interaction caused a decrease in k₁ as the amount of ionic liquid 6 in the reaction mixture with methanol was increased (Fig. 5), as the entropic cost was more significant than the enthalpic benefit.

Scheme 3:

The unimolecular substitution reaction between (R)-3-chloro-3,7-dimethyloctane 18 and methanol, which proceeds through a substantially charge separated transition state leading to the ionic intermediate 19 [39], [40].

Fig. 5:

The changes in the unimolecular rate constant for the reaction between (R)-3-chloro-3,7-dimethyloctane 18 and methanol as the mole fraction of [Bmim][N(CF₃SO₂)₂] 6 in the reaction mixture was increased [39]. Reproduced from Ref. [39] with permission from Elsevier.

Conversely, for S_N1 processes that involve delocalisation of the positive charge in the transition state, such as across neighbouring π systems, there is an increase in the rate of reaction in ionic liquids, relative to polar aprotic solvents [41], [70]. It was speculated that this increase in k₁ arises from a lessened extent of solvation of the transition state (as the charge is delocalised) [41], relative to the aliphatic substrate 18 described above [39], [40]. This decrease in ordering of the ionic liquid about the transition state may result in a less significant entropic cost, and could account for the increase in k₁ observed.

To confirm this hypothesis, recent work has focussed on the reaction between bromodiphenylmethane 21 and 3-chloropyridine 23 (Scheme 4), with a series of kinetic analyses performed in [Bmim][N(SO₂CF₃)₂] 6/acetonitrile mixtures [42]. This reaction proceeds through both a unimolecular pathway involving formation of a benzylic carbocation in the intermediate 22, with the transition state leading to this carbocation involving an extent of charge delocalisation across the π systems [69], [71], along with a parallel bimolecular pathway.

Scheme 4:

The reaction between bromodiphenylmethane 21 and 3-chloropyridine 23 to produce the salt 24 [42].

Interestingly, analysis of the unimolecular mechanism of the reaction between species 21 and 23 demonstrated that there is an initial, significant increase in k₁ when moving from acetonitrile to χ_ILca. 0.2, where there is a rate enhancement of ca. 13-fold relative to acetonitrile (Fig. 6) [42]. There is then a gradual decrease in k₁ with increasing χ_IL, although, importantly, there is always a rate enhancement when [Bmim][N(SO₂CF₃)₂] 6 is present in the reaction mixture, relative to acetonitrile. The difference in the mole fraction dependence of the rate constants shown in Figs. 5 and 6 indicate that the ionic liquid 6 affects the formation of an aliphatic carbocation differently to a benzylic carbocation.

Fig. 6:

The changes in the unimolecular rate constant for the reaction between bromodiphenylmethane 21 and 3-chloropyridne 23 as the mole fraction of [Bmim][N(CF₃SO₂)₂] 6 in the reaction mixture was increased [42]. Reproduced from Ref. [42] with permission from The Royal Society of Chemistry.

The activation parameters determined for the unimolecular reaction between species 21 and 23 indicated that when using a high concentration of the ionic liquid 6 in the reaction mixture (χ_ILca. 0.88) there is a decrease in both the enthalpy and entropy of activation, relative to acetonitrile [42]. This trend is comparable to that determined previously for the S_N1 reaction between (R)-3-chloro-3,7-dimethyloctane 18 and methanol [40], indicating that once again the ionic liquid 6 is stabilising the incipient charges in the transition state. Importantly, for the formation of the benzylic carbocation from species 21 the enthalpic benefit associated with this ion – charged transition state interaction now outweighs the entropic cost, resulting in increased k₁ values across all mole fractions of the ionic liquid 6 in the reaction mixture, relative to acetonitrile (Fig. 3) [42].

Overall, the recent work examining the S_N1 reaction between bromodiphenylmethane 21 and 3-chloropyridne 23 (Scheme 4) demonstrated that the extent of charge localisation in the transition state is very important, as the strength of the ionic liquid – transition state interaction affects the balance between the opposing enthalpic and entropic effects when using an ionic liquid solvent [42]. This is important as it highlights that relatively weak ionic liquid – charge interactions may increase the rate of organic processes (as the enthalpic benefit dominates), while stronger ionic liquid – charge interactions tend to result in a decreased rate constant (as the entropic cost dominates), relative to molecular solvents [28], [30], [39], [40], [66], [67], [72].

An ionic liquid – charge interaction was also found to affect the condensation reaction between benzaldehydes 11 and hexan-1-amine 12 (Scheme 2). In the earlier section examining cation – lone pair interactions discussion focussed on the electrophile 4-methoxybenzaldehyde 11a. Attention is now going to turn to the effect that changing the electronic character of the substituent on electrophile 11 has on the observed ionic liquid solvent effect. This was investigated using the electrophiles 11a–f, where the rate constant of the reaction of these species with hexan-1-amine 12 was determined in a number of mole fractions of [Bmim][N(CF₃SO₂)₂] 6 in acetonitrile [50]. When considering the initial step of this two-step addition-elimination process (nucleophilic attack of an amine on a carbonyl group), the rate constants determined in acetonitrile suggested that there was an increase in the extent of charge development in the transition state on moving from electron donating to electron withdrawing groups on the electrophile 11. In mixtures containing the ionic liquid 6, these changes in the transition state charge separation as the substituent on species 11 was varied resulted in an increase in the electrostatic interactions between [Bmim][N(CF₃SO₂)₂] 6 and the incipient charges in the transition state. The increasing magnitude of this interaction caused some interesting changes in the rate constant of this process as the substituent was varied [50]. This is important as another important interaction that affects reaction outcome, other than the cation – hexan-1-amine 12 interaction, was revealed. Additionally, it was demonstrated that the rate of the reaction between benzaldehydes 11 and hexan-1-amine 12 is much more sensitive to changing the substituents on 11 in the ionic liquid 6 compared to acetonitrile; this is because the differences in transition state charge development result in more marked changes in the ionic liquid – transition state interactions than the changes in the acetonitrile – transition state interactions.

Ionic liquid – quadrupole interactions

In the above sections the importance of interactions between the ionic liquid components and either lone pairs or charged centres on the species along the reaction coordinate have been highlighted. In this final section interactions between an ionic liquid and the electron density in delocalised π systems will be discussed, particularly focussing on ionic liquid – quadrupole interactions.

In early work examining the solubility of aromatic compounds in ionic liquid solvents, it was shown that the high solubility of these species arose from favourable interactions between the ionic liquid components and the quadrupole resulting from the delocalized π system of the aromatic solutes [7], [8], [73]. Previous work examining the effect of ionic liquid solvents on nucleophilic aromatic substitution (S_NAr) processes demonstrated that the ionic liquid – quadrupole interaction can also affect reaction kinetics. Extensive kinetic analyses on a representative S_NAr process were able to show that the ionic liquid was interacting with the delocalized π system on the electrophile [51], [53], [54]. On moving to the transition state, where the aromaticity of the electrophile is disrupted, the ionic liquid – quadrupole interaction was reduced, resulting in an entropic benefit relative to ethanol. This caused an entropically driven rate enhancement when using an ionic liquid solvent [51], [53], [54].

It has also been widely demonstrated that the use of an ionic liquid solvent increases the rate of cycloaddition processes, relative to most organic solvents [25], [27], [29], [55], [56], [74], [75], [76], [77], [78], [79], [80], [81], [82], [83], [84]. In much of this work the changes in reaction outcome when using an ionic liquid solvent were attributed to a combination of hydrogen-bonding and general electrostatic interactions between the ionic liquid and species along the reaction coordinate, as well as the high cohesive energy of ionic liquids. Recent work examining the effect of an ionic liquid solvent on pericyclic rearrangements has highlighted that interactions between the ionic liquid and the delocalized π system that forms in the transition state may also be important to consider when rationalising the ionic liquid solvent effects on pericyclic processes.

The Cope rearrangement of 3-phenyl-1,5-hexadiene 25 (Scheme 5) was found to proceed much faster when using [Bmim][N(SO₂CF₃)₂] 6 (χ_ILca. 0.99) as the solvent, rather than any of the molecular solvents benzene, acetonitrile and ethanol [85]. As this process has no electron rich sites in the starting material 25 and little, if any, charge separation in the transition state [86], [87], [88], the rate enhancement observed could not be attributed to interactions between the ionic liquid and either an electron rich site on the starting material (as demonstrated in interactions with lone pairs above) or stabilisation of incipient charges in the transition state (as demonstrated with interactions with incipient charges above).

Scheme 5:

The Cope rearrangement of 3-phenyl-1,5-hexadiene 25 to give the product 26, which proceeds through a cyclic, aromatic transition state [85].

As this is a pericyclic process, the Cope rearrangement of species 25 involves a negative volume of activation. As such, on forming the cyclic transition state there will be a decrease in the solute – solvent interfacial area and hence a reduction in any unfavourable interactions between the non-polar species 25 and the solvent. As an example, this resulted in an increase in the rate constant when moving from the non-polar solvent benzene to the polar solvent acetonitrile, as there is a more significant solvophobic effect in the polar solvent. When using [Bmim][N(SO₂CF₃)₂] 6 the rate constant was higher than would be expected if only the solvophobic effect associated with the negative volume of activation of this process was important. As the transition state for this rearrangement involves formation of a delocalized π system [89], [90], [91], it was suggested that favourable ion – quadrupole interactions between the ionic liquid 6 and the transition state are also accelerating the rearrangement of species 25. These ion – quadrupole interactions will enhance the solvophobic effect in the ionic liquid 6, as increased solvent – transition state interactions will further favour transition state formation. Overall, this recent work demonstrated that the high k₁ value for the rearrangement of species 25 in the ionic liquid 6 is due to favourable ion – quadrupole interactions with the aromatic system that forms in the transition state [85].

A similar ionic liquid – quadrupole interaction was also proposed to contribute to the changes in k₁ for the Claisen rearrangement of allyl vinyl ether 27 (Scheme 6) [85], as this process also proceeds through an aromatic transition state [92], [93], [94]. When using [Bmim][N(CF₃SO₂)₂] 6 (χ_IL=0.90) it was found that there was once again a rate enhancement, relative to the molecular solvents benzene, ethanol and acetonitrile. The activation parameters for this process suggested that the increased rate constant in the ionic liquid 6 is due to interactions between the ionic liquid and the delocalised π system in the transition state, as well as stabilisation of the incipient charges in the transition state through general coulombic interactions. That is, when using the ionic liquid 6 it is likely that both ion – quadrupole and ion – charge interactions with the transition state affect the rate constant for the rearrangement of species 27. Interestingly, when using ethanol as the solvent it was found that there was a strong hydrogen bonding interaction between ethanol and the ether 27, which inhibited the rearrangement. Such an interaction was not observed in the ionic liquid 6, suggesting that the ionic liquid 6 interacts with species along the reaction coordinate through general electrostatic interactions (more ‘acetonitrile-like’) rather than through hydrogen bonding interactions (less ‘ethanol-like’).

Scheme 6:

The Claisen rearrangement of allyl vinyl ether 27 to give the aldehyde 28; this process proceeds through a cyclic, aromatic transition state that likely features an extent of charge separation [85].

Conclusions

In summary, to allow for predictions of the overall ionic liquid solvent effect to be made, it is essential to have a thorough understanding of both the interactions that exist between the ionic liquid and all species along the reaction coordinate, and the relative strengths of these interactions. Thus far the main interactions that have been identified as contributors to ionic liquid solvent effects are: cation – lone pair, ionic liquid – charge and ionic liquid – quadrupole.

Recent studies have demonstrated that by using the knowledge gained from previous work [41], [43], [44], [45], the main interaction affecting the outcome of a representative condensation process could be accurately predicted [46]. This highlighted that once the likely solvent – reagent and solvent – transition state interactions have been identified and the energetic result of each interaction considered, it is possible to make reasonable predictions about the changes in the activation parameters when using ionic liquid solvents, relative to molecular solvents. It was also shown for the first time that the magnitude of the cation – lone pair interaction could be controlled by varying the constituent ions of the ionic liquid, resulting in predictable changes in the activation parameters of the process [46], [49]. However, assessing the weighting of the opposing enthalpic and entropic effects, and hence predicting the effects of ionic liquids on the rate constant, remains challenging. This work also highlighted that additional effects are likely contributing to the overall ionic liquid solvent effects, adding to the difficulty in predicting the overall change in the rate constant when using different ionic liquids [46], [49].

The recent kinetic analyses on the S_N1 reaction between bromodiphenylmethane 21 and 3-chloropyridine 23 demonstrated that the magnitude of the interaction between the ionic liquid and the incipient charges in the transition state impacts the balance between the opposing enthalpic and entropic effects, and hence the rate constant [42]. In combination with previous studies [28], [30], [39], [40], [62], [63], [71], this work suggests that weaker ionic liquid – charge interactions may increase the rate of organic processes, while strong ionic liquid – charge interactions are more likely to decrease that rate constant, relative to molecular solvents. The magnitude of ionic liquid – charge interactions was also found to affect the condensation reaction between benzaldehydes 11 and hexan-1-amine 12. As the substituent on species 11 was changed the extent of charge separation in the transition state varied, resulting in changes in the magnitude of the ionic liquid – transition state interaction, and hence the activation parameters and rate constant for the process.

Lastly, recent work examining the effect of an ionic liquid solvent on pericyclic rearrangements found that these processes were accelerated by favourable ionic liquid – quadrupole interactions with the aromatic system that forms in the transition state. This highlights that ionic liquid – quadrupole interactions needs to be considered when attempting to predict the effect of an ionic liquid solvent on processes that involve formation of a delocalised π system in the transition state. It was also proposed that for the Claisen rearrangement of allyl vinyl ether 27 the ionic liquid interacts with species along the reaction coordinate through general coulombic interactions rather than hydrogen bonding.