Room Temperature Ionic Liquids in Asymmetric Hetero-Ene Type Reactions: Improving Organocatalyst Performance at Lower Temperatures

Room temperature ionic liquids (RTILs) have been widely used as (co)solvents in several catalytic processes modifying, in most of the cases, the catalyst activity and/or the selectivity for the studied reactions. However, there are just a few examples of their use in hydrogen bonding organocatalysis. In this paper, we show the positive effect of a set of imidazole-based ionic liquids ([bmim]BF4 and [hmim]PF6) in the enantioselective addition of formaldehyde tert-butylhydrazone to prochiral α-keto esters catalyzed by a sugar-based chiral thiourea. Reactions performed in the presence of low percentages of RTILs led to an increase of the catalyst activity, thereby making possible to work at lower temperatures. Thus, the chiral tert-butyl azomethyl tertiary alcohols could be obtained with moderate to good conversions and higher enantioselectivities for most of the studied substrates when using up to 30 vol% of [hmim]PF6 as a cosolvent in processes performed in toluene.


Introduction
Room temperature ionic liquids (RTILs) are a class of organic salts that have found application in a number of contexts [1][2][3][4][5]. Among them, RTILs are currently used as solvents for different processes due to their unique properties, such as a high polarity, a poor (although tunable) coordination between anion and cation, and the fact that their structure can be easily modified. Moreover, it has been widely proved that ionic liquids may affect the outcome of different reactions, exerting beneficial effects in product recovery, reaction rate enhancement, and improved stereoselectivity [6][7][8]. For these reasons, the role of RTILs has been extensively studied as solvents or cosolvents in metal-catalyzed [9], biocatalyzed [10][11][12][13], and organocatalyzed reactions [14,15]. Regarding the latter, most of the examples reported so far refer to the use of ionic liquids in covalent organocatalysis. On the other hand, there are very few examples of their use in hydrogen-bonding catalysis using (thio)ureas or squaramides [16,17]. Very recently, a set of thiourea-based ionic liquids were used as catalysts for the cycloaddition of CO 2 to epoxides. These imidazolium-based thioureas showed an improved activity due to the coexistence of the imidazole and the thiourea group as active moieties in the molecule [18,19].

Results and Discussion
Initial experiments were performed to analyze the effect of small amounts (10 vol%) of different RTILs in toluene in the addition of hydrazone 1 to ethyl 2-oxo-4-phenylbutyrate (2a) catalyzed by II (10 mol%), as shown in Table 1. This amount of RTIL was deliberately chosen for the first set of experiments to ensure a 1:1 ratio of catalyst/RTIL, assuming that at least one molecule of RTIL interacts with one molecule of catalyst. In the absence of RTIL, the azomethyl alcohol (S)-3a was obtained with 68% conversion, measured by 1 H-NMR and 64% ee after 16 h (entry 1). In order to compare the effect of the RTIL on the Scheme 1. (Thio)urea-organocatalyzed preparation of optically active azomethyl alcohols (S)-3a-h employing catalysts II or III and preparation of compound (R)-3h using catalyst I in toluene containing different room temperature ionic liquids.

Results and Discussion
Initial experiments were performed to analyze the effect of small amounts (10 vol%) of different RTILs in toluene in the addition of hydrazone 1 to ethyl 2-oxo-4-phenylbutyrate (2a) catalyzed by II (10 mol%), as shown in Table 1. This amount of RTIL was deliberately chosen for the first set of experiments to ensure a 1:1 ratio of catalyst/RTIL, assuming that at least one molecule of RTIL interacts with one molecule of catalyst. In the absence of RTIL, the azomethyl alcohol (S)-3a was obtained with 68% conversion, measured by 1 H-NMR and 64% ee after 16 h (entry 1). In order to compare the effect of the RTIL on the catalyst activity independent of the reaction time, we also determined the catalyst turnover frequency (TOF, h −1 ), which reached a value of 0.43 h −1 in this reference experiment. When employing different imidazole-based ionic liquids (entries 2-6), the reaction was accelerated. Thus, similar conversions were achieved in shorter reaction times (7 h). TOF values under these reaction conditions were 2 to 3 times higher than in pure toluene. The use of [bmim]BF 4 , [bmim]PF 6 , and [hmim]PF 6 led to highest conversions of the α-keto ester, while the optical purities were the same as those achieved in toluene for [bmim]BF 4 (entry 3, 63% ee) and [hmim]PF 6 (entry 5, 65% ee). Imidazole-based ionic liquids containing NTf 2 and especially MeSO 4 as anions had a negative effect in the system stereoselectivity. This phenomenon can be ascribed to interferences in the proposed recognition mode, as compared to BF 4 and PF 6 --containing imidazolium-based RTILs. The reaction was also performed with 10 vol% of the pyridine-based ionic liquid [bmp]BF 4 (entry 7). After 7 h, a 45% conversion was achieved, showing a worse performance than the imidazole-based ionic liquids, but still higher TOF values than the reference reaction were recorded (TOF 0.64 vs. 0.43 h −1 , respectively). Molar fraction (X RTIL ) of ionic liquids were calculated for a 10 vol% volumetric ratio. As can be observed in Table 1, X RTIL for each cosolvent depended on the structure of the cation and the anion molecular weight and varied from 3.5 to 5.9%. Unfortunately, no clear trend between the molar fraction percent and the TOF could be established. In view of the higher activity found at −15 • C using a 10 vol% of certain imidazolebased ionic liquids, the temperature was lowered to −30 • C in the presence of [bmim]BF 4 or [hmim]PF 6 , aiming to reach higher enantiomeric excesses of the azomethyl alcohol (S)-3a (entries 9 and 10) at lower temperatures. After 22 h, (S)-product was obtained in both cases with conversions of 83% and 89%, respectively, and optical purities around 70% ee, while the reaction in pure toluene occurred with a 29% conversion and the same optical purity (entry 8). At this temperature, the TOF for this reaction when using [hmim]PF 6 was 3 times higher than in absence of RTIL. Temperature was further lowered to −45 • C. Under these conditions, the organocatalyzed process in pure toluene led to a 21% conversion of (S)-3a after 24 h, reaching a slightly better 71% ee (entry 11), while the addition of the RTILs increased the conversions up to 70% in the same reaction times (entries 12 and 14), keeping the enantiomeric excess of the (S)-alcohol around 75%. This means a 3-times increase in TOF values. The reaction was also tested employing as a catalyst urea III, an analogue to thiourea II (entries 13 and 15). For both ionic liquids, conversions and selectivities in the reactions catalyzed by III were slightly lower when compared with II. The organocatalyzed reaction in toluene alone and containing 10 vol% [hmim]PF 6 was also carried at -60 • C (entries 16 and 17), but at this temperature the final product (S)-3a was obtained with no improvement in optical purity as at -45 • C. The reaction was slowed down to a significant extent in presence of 10 vol% [hmim]PF 6 (34% conversion after 72 h, entry 17). However, the reactivity virtually doubled in the presence of the ionic liquid (TOF 0.03 vs. 0.06 h −1 , in absence and presence of RTIL, respectively).
There is not a clear explanation about the influence of RTILs in organocatalyzed reactions, despite the fact that many examples have been described in which these compounds have performed a beneficial effect in catalytic systems. In our case, a plausible explanation of the observed positive effect on catalyst performance is that the RTIL increased the acidity of the (thio)urea moiety by a hydrogen bonding interaction, as proposed in Figure 1 [30,[37][38][39]. In this way, the ionic liquid (IL) strengthened the interaction between the catalyst and the α-keto ester. It must be pointed out that in toluene, the IL ion pair is in tight proximity, so the ionic liquid should be considered as a single zwitterionic specie rather than a pair of free counter ions. On the other hand, the poor Lewis basic character (or coordination ability) of the BF 4 − or PF 6 − counteranions enhances the relatively high acidity of the imidazolium cation and its availability to interact with the catalyst [40,41]. Moreover, such interactions will also have a beneficial effect by avoiding catalyst self-aggregation and, consequently, increasing the solubility of the catalyst [33] at lower temperatures. This assumption is in line with the results obtained with catalysts II and III, showing that the thiourea moiety, being essentially more acidic than the urea analogue, provided a higher catalytic activity. Finally, this rationale is also in agreement with the fact that the IL did not significantly interfere/modify the chiral recognition mode, since the enantiomeric excesses were quite similar to those obtained in the absence of RTIL (for instance, see entries [11][12][13][14][15], experiments conducted at the same temperature (-45 • C)). purity (entry 8). At this temperature, the TOF for this reaction when using [hmim]PF6 was 3 times higher than in absence of RTIL. Temperature was further lowered to −45 °C. Under these conditions, the organocatalyzed process in pure toluene led to a 21% conversion of (S)-3a after 24 h, reaching a slightly better 71% ee (entry 11), while the addition of the RTILs increased the conversions up to 70% in the same reaction times (entries 12 and 14), keeping the enantiomeric excess of the (S)-alcohol around 75%. This means a 3-times increase in TOF values. The reaction was also tested employing as a catalyst urea III, an analogue to thiourea II (entries 13 and 15). For both ionic liquids, conversions and selectivities in the reactions catalyzed by III were slightly lower when compared with II. The organocatalyzed reaction in toluene alone and containing 10 vol% [hmim]PF6 was also carried at -60 °C (entries 16 and 17), but at this temperature the final product (S)-3a was obtained with no improvement in optical purity as at -45 °C. The reaction was slowed down to a significant extent in presence of 10 vol% [hmim]PF6 (34% conversion after 72 h, entry 17). However, the reactivity virtually doubled in the presence of the ionic liquid (TOF 0.03 vs. 0.06 h −1 , in absence and presence of RTIL, respectively).
There is not a clear explanation about the influence of RTILs in organocatalyzed reactions, despite the fact that many examples have been described in which these compounds have performed a beneficial effect in catalytic systems. In our case, a plausible explanation of the observed positive effect on catalyst performance is that the RTIL increased the acidity of the (thio)urea moiety by a hydrogen bonding interaction, as proposed in Figure 1 [30,[37][38][39]. In this way, the ionic liquid (IL) strengthened the interaction between the catalyst and the α-keto ester. It must be pointed out that in toluene, the IL ion pair is in tight proximity, so the ionic liquid should be considered as a single zwitterionic specie rather than a pair of free counter ions. On the other hand, the poor Lewis basic character (or coordination ability) of the BF4 − or PF6 − counteranions enhances the relatively high acidity of the imidazolium cation and its availability to interact with the catalyst [40,41]. Moreover, such interactions will also have a beneficial effect by avoiding catalyst self-aggregation and, consequently, increasing the solubility of the catalyst [33] at lower temperatures. This assumption is in line with the results obtained with catalysts II and III, showing that the thiourea moiety, being essentially more acidic than the urea analogue, provided a higher catalytic activity. Finally, this rationale is also in agreement with the fact that the IL did not significantly interfere/modify the chiral recognition mode, since the enantiomeric excesses were quite similar to those obtained in the absence of RTIL (for instance, see entries [11][12][13][14][15], experiments conducted at the same temperature (-45 °C)). The effect of some reaction parameters on the enantiomeric excess of the products (addition of 1 to 2a catalyzed by II in mixtures toluene/[hmim]PF6) were analyzed. Thus, the synthesis of (S)-3a was carried out using different volumetric proportions of the ionic The effect of some reaction parameters on the enantiomeric excess of the products (addition of 1 to 2a catalyzed by II in mixtures toluene/[hmim]PF 6 ) were analyzed. Thus, the synthesis of (S)-3a was carried out using different volumetric proportions of the ionic liquid, as shown in Figure 2. The reactions could be performed with good results in terms of activity and selectivity using up to 30 vol% of [hmim]PF 6 , while higher RTIL proportions led to an important drop in both the conversion toward the (S)-azomethyl alcohol and its enantiomeric purity. In pure RTIL, the reaction still proceeded, but neither conversion nor selectivity were satisfactory. Different groups have studied the dependence between the RTIL proportion and the reaction outcome, being observed herein a similar trend to what was previously reported by Harper et al. [42] with a more pronounced effect of [hmim]PF 6 between 10-30 vol%. The effect of substrate concentration was also analyzed in the preparation of (S)-3a while keeping constant the amount of catalyst. Reactions were typically performed at keto ester concentrations of 0.6 M. When increasing the substrate concentration up to 1.0 M, a 31% conversion was obtained after 24 h (0.22 h −1 TOF), whereas there was a slight loss in the enantioselectivity, as the product 3a was obtained with 63% ee. The use of 0.3 M of 2a also afforded (S)-3a with a lower TOF (0.14 h −1 ) with respect to that observed at 0.6 M. A 65% conversion was measured after 24 h with the same optical purity. Reactions were also performed at lower substrate concentrations (0.1 M and 0.05 M), which obtained lower TOF values (0.12 and 0.11 h −1 , respectively), whereas the optical purities were slightly higher than at 0.6 M concentration, as (S)-3a was recovered with 78% ee for both concentrations. Overall, an α-keto ester concentration of 0.6 M seemed to be the best for this organocatalyzed process.
tions led to an important drop in both the conversion toward the (S)-azomethyl alcohol and its enantiomeric purity. In pure RTIL, the reaction still proceeded, but neither conversion nor selectivity were satisfactory. Different groups have studied the dependence between the RTIL proportion and the reaction outcome, being observed herein a similar trend to what was previously reported by Harper et al. [42] with a more pronounced effect of [hmim]PF6 between 10-30 vol%. The effect of substrate concentration was also analyzed in the preparation of (S)-3a while keeping constant the amount of catalyst. Reactions were typically performed at keto ester concentrations of 0.6 M. When increasing the substrate concentration up to 1.0 M, a 31% conversion was obtained after 24 h (0.22 h −1 TOF), whereas there was a slight loss in the enantioselectivity, as the product 3a was obtained with 63% ee. The use of 0.3M of 2a also afforded (S)-3a with a lower TOF (0.14 h −1 ) with respect to that observed at 0.6 M. A 65% conversion was measured after 24 h with the same optical purity. Reactions were also performed at lower substrate concentrations (0.1 M and 0.05 M), which obtained lower TOF values (0.12 and 0.11 h −1 , respectively), whereas the optical purities were slightly higher than at 0.6 M concentration, as (S)-3a was recovered with 78% ee for both concentrations. Overall, an α-keto ester concentration of 0.6 M seemed to be the best for this organocatalyzed process. Once selecting both [bmim]BF4 and [hmim]PF6 as the best cosolvents/additives for the organocatalyzed additions, we examined the synthesis of the isopropyl derivative (S)-3b (Table 1, entries [18][19][20][21]. This compound was obtained with 98% conversion and 58% ee when employing toluene at −15 °C [29]. When the reaction was carried out at −45 °C in the same reaction medium (entry 18), (S)-3b was recovered with 29% conversion and 62% ee. The presence of 10 vol% of either of these two selected RTILs at −45 °C allowed obtaining the final product with a slightly higher optical purity (66-67% ee). The use of [hmim]PF6 led to a higher conversion (75% after 24 h, entry 20), and this RTIL was employed for the other substrates 2c-g. Addition of 1 to 2b was also catalyzed by urea III. In the presence of 10 vol% [hmim]PF6, (S)-3b was obtained with 61% conversion and 52% ee, as shown in entry 22.
The best conditions found for the enantioselective preparation of both (S)-3a and (S)-3b, using [hmim]PF6 at 10 vol% in toluene (4.6% molar fraction of this RTIL) and −45 °C, were extended to other substrates (Scheme 2). Thus, the addition of hydrazone 1 to the n- Once selecting both [bmim]BF 4 and [hmim]PF 6 as the best cosolvents/additives for the organocatalyzed additions, we examined the synthesis of the isopropyl derivative (S)-3b (Table 1, entries [18][19][20][21]. This compound was obtained with 98% conversion and 58% ee when employing toluene at −15 • C [29]. When the reaction was carried out at −45 • C in the same reaction medium (entry 18), (S)-3b was recovered with 29% conversion and 62% ee. The presence of 10 vol% of either of these two selected RTILs at −45 • C allowed obtaining the final product with a slightly higher optical purity (66-67% ee). The use of [hmim]PF 6 led to a higher conversion (75% after 24 h, entry 20), and this RTIL was employed for the other substrates 2c-g. Addition of 1 to 2b was also catalyzed by urea III. In the presence of 10 vol% [hmim]PF 6 , (S)-3b was obtained with 61% conversion and 52% ee, as shown in entry 22.
The best conditions found for the enantioselective preparation of both (S)-3a and (S)-3b, using [hmim]PF 6 at 10 vol% in toluene (4.6% molar fraction of this RTIL) and −45 • C, were extended to other substrates (Scheme 2). Thus, the addition of hydrazone 1 to the n-propyl keto ester 2c afforded the azomethyl alcohol (S)-3c with 61% ee (0.25 h −1 ) in toluene/RTIL, which represents a 16% increase in the optical purity value with respect to the process in toluene at −15 • C. The methyl derivative (S)-3d was obtained with a 71% ee and a 68% conversion (0.28 h −1 ) after 24 h in the reaction catalyzed by II at -45 • C in the presence of 10 vol% [hmim]PF 6 . The presence of a longer alkyl chain (n-hexyl) in the α-keto ester (2e) allowed obtaining the final azomethyl alcohol with 75% conversion after 24 h at -45 • C and 77% ee (0.25 h −1 ), which represents a 14% increase in the optical purity regarding the reaction in toluene. A bulky alkyl substrate as 2f was also tested. The reaction in toluene led to a low conversion even at 0 • C. When the reaction was performed at -30 • C using [hmim]PF 6 , (S)-3f was obtained with a very low conversion after 30 h (13%, 0.04 h −1 ), while for this compound the presence of the ionic liquid did not improve the enantiomeric excess. This low productivity may be due to steric clash between the t Bu group of the substrate and the organocatalyst, thus establishing a loose interaction and precluding a proper activation. Noteworthy, the i Pr-derivative 2b (75% conversion, 67% ee (0.31 h −1 )) and the t Bu-containing 2f (13% conversion, 66% ee (0.04 h −1 )) only differed in a methyl group. This exemplifies that both substrates have the same substrate-catalyst binding mode (almost identical enantioselectivity) but a different extent of activation. The novel reaction medium was extended to the benzylic derivative 2g, for which a 63% ee was observed when employing 10 vol% [hmim]PF 6 , while a 70% conversion was achieved after 24 h (0.29 h −1 ). process in toluene at −15 °C. The methyl derivative (S)-3d was obtained with a 71% ee and a 68% conversion (0.28 h −1 ) after 24 h in the reaction catalyzed by II at -45 °C in the presence of 10 vol% [hmim]PF6. The presence of a longer alkyl chain (n-hexyl) in the α-keto ester (2e) allowed obtaining the final azomethyl alcohol with 75% conversion after 24 h at -45 °C and 77% ee (0.25 h −1 ), which represents a 14% increase in the optical purity regarding the reaction in toluene. A bulky alkyl substrate as 2f was also tested. The reaction in toluene led to a low conversion even at 0 °C. When the reaction was performed at -30 °C using [hmim]PF6, (S)-3f was obtained with a very low conversion after 30 h (13%, 0.04 h −1 ), while for this compound the presence of the ionic liquid did not improve the enantiomeric excess. This low productivity may be due to steric clash between the t Bu group of the substrate and the organocatalyst, thus establishing a loose interaction and precluding a proper activation. Noteworthy, the i Pr-derivative 2b (75% conversion, 67% ee (0.31 h −1 )) and the t Bu-containing 2f (13% conversion, 66% ee (0.04 h −1 )) only differed in a methyl group. This exemplifies that both substrates have the same substrate-catalyst binding mode (almost identical enantioselectivity) but a different extent of activation. The novel reaction medium was extended to the benzylic derivative 2g, for which a 63% ee was observed when employing 10 vol% [hmim]PF6, while a 70% conversion was achieved after 24 h (0.29 h −1 ).

Scheme 2.
Substrate scope for the enantioselective addition of hydrazone 1 to different aliphatic αketoesters catalyzed by II in toluene containing a 10 vol% of [hmim]PF6 (4.6% X [hmim]PF6). Reactions were performed at −45 °C for 30 h, except for 2a (24 h). Reported values are the mean of two replicates, with the uncertainties reported as half of the range. In order to spread the benefits of the optimized reaction medium to other hydrogenbonding catalysts, we analyzed the addition of hydrazone 1 to ethyl benzoylformate 2h catalyzed by (R)-BINAM-derived bis-urea I (Scheme 3). The reaction performed in toluene at −15 • C led to a 53% conversion of the azomethyl alcohol (R)-ethyl 3-(tert-butyldiazenyl)-2-hydroxy-2-phenylpropanoate (3h) after 24 h (0.22 h −1 ), while the addition of the IL increased the conversion up to 71% (0.33 h −1 ). For both reactions, the optical purity of the final product was the same (89% ee).
In order to spread the benefits of the optimized reaction medium to other hydrogenbonding catalysts, we analyzed the addition of hydrazone 1 to ethyl benzoylformate 2h catalyzed by (R)-BINAM-derived bis-urea I (Scheme 3). The reaction performed in toluene at −15 °C led to a 53% conversion of the azomethyl alcohol (R)-ethyl 3-(tert-butyldiazenyl)-2-hydroxy-2-phenylpropanoate (3h) after 24 h (0.22 h −1 ), while the addition of the IL increased the conversion up to 71% (0.33 h −1 ). For both reactions, the optical purity of the final product was the same (89% ee).

General Materials and Methods
NMR spectra were recorded by 1 H NMR (300 or 400 MHz) and 13 C NMR (75 or 100 MHz) with the solvent peak used as the internal reference (7.26 and 77.0 ppm for 1 H and 13 C, respectively) on a Bruker AC-300-DPX. Column chromatography was performed on silica gel (Merck Kieselgel 60, Merck, Kenilworth, United States). Analytical TLC was performed on aluminum backed plates (1.5 × 5 cm) precoated (0.25 mm) with silica gel (Merck, Silica Gel 60 F254, Merck, Kenilworth, United States). The compounds were visualized by exposure to UV light or by dipping the plates into solutions of KMnO4 or vanillin, followed by heating.

Typical Procedure for the Hetero-Ene Reaction Catalyzed by (Thio)Ureas I-III
Formaldehyde tert-butyl hydrazone 1 (134 μL, 1.2 mmol) was added to a solution of α-keto ester 2a-h (0.6 mmol) and catalyst I-III (0.06 mmol) in the corresponding toluene/RTIL mixture (total volume 0.6 mL) at the selected temperatures. Reactions were stirred for the established times (TLC monitoring). The solvent was removed under reduced pressure and the conversions were determined by 1 H-NMR in CDCl3. For some selected reactions, flash chromatography purifications were performed in different toluene/EtOAc mixtures to measure the isolated yield of the corresponding optically active azomethyl alcohols (S)-3a-g and (R)-3h (see Supplementary Materials). Physical and spectroscopical properties of the azomethyl compounds 3a-h matched those previously reported [29,30]. The optical purity of compounds (S)-3a, (S)-3c, (S)-3g, and (R)-3h was directly determined by HPLC on chiral stationary phases, while (S)-3b and (S)-3d-f were previously converted into their corresponding azoxymethyl alcohols by treatment with magnesium monoperoxyphthalate (MMPP) in methanol (for additional details see Supplementary Materials, Table S1).

General Materials and Methods
NMR spectra were recorded by 1 H NMR (300 or 400 MHz) and 13 C NMR (75 or 100 MHz) with the solvent peak used as the internal reference (7.26 and 77.0 ppm for 1 H and 13 C, respectively) on a Bruker AC-300-DPX. Column chromatography was performed on silica gel (Merck Kieselgel 60, Merck, Kenilworth, United States). Analytical TLC was performed on aluminum backed plates (1.5 × 5 cm) precoated (0.25 mm) with silica gel (Merck, Silica Gel 60 F 254 , Merck, Kenilworth, United States). The compounds were visualized by exposure to UV light or by dipping the plates into solutions of KMnO 4 or vanillin, followed by heating.

Typical Procedure for the Hetero-Ene Reaction Catalyzed by (Thio)Ureas I-III
Formaldehyde tert-butyl hydrazone 1 (134 µL, 1.2 mmol) was added to a solution of αketo ester 2a-h (0.6 mmol) and catalyst I-III (0.06 mmol) in the corresponding toluene/RTIL mixture (total volume 0.6 mL) at the selected temperatures. Reactions were stirred for the established times (TLC monitoring). The solvent was removed under reduced pressure and the conversions were determined by 1 H-NMR in CDCl 3 . For some selected reactions, flash chromatography purifications were performed in different toluene/EtOAc mixtures to measure the isolated yield of the corresponding optically active azomethyl alcohols (S)-3a-g and (R)-3h (see Supplementary Materials). Physical and spectroscopical properties of the azomethyl compounds 3a-h matched those previously reported [29,30]. The optical purity of compounds (S)-3a, (S)-3c, (S)-3g, and (R)-3h was directly determined by HPLC on chiral stationary phases, while (S)-3b and (S)-3d-f were previously converted into their corresponding azoxymethyl alcohols by treatment with magnesium monoperoxyphthalate (MMPP) in methanol (for additional details see Supplementary Materials, Table S1).

Conclusions
The addition of certain imidazole-based ionic liquids ([bmim]BF 4 and [hmim]PF 6 ) in the thiourea-organocatalyzed addition of formaldehyde tert-butyl hydrazine (1) to different aliphatic α-keto esters (2a-h) had a positive effect in the catalyst activity. Thus, reactions can be conducted at lower temperatures, which allow obtaining the corresponding chiral azomethyl alcohols with moderate to good conversions and optical purities higher than those obtained in the absence of ionic liquids for most of the substrates tested, making it possible to attain enantiomeric excesses around 70-77% and up to a 3-fold TOF increase. To the best of our knowledge, this is one of the few examples of the positive effect of engineering the reaction of medium properties by adding ionic liquids in hydrogen bond organocatalysis. RTILs may find application in improving the performance of highly selective but "lazy" catalysts or enabling catalysis in otherwise nonproductive extreme temperature regimes.
Supplementary Materials: The following are available online, Experimental procedures and Table S1: HPLC conditions for the determination of the enantiomeric excess of azomethyl alcohols 3a-h.