Introduction

Carbocations1,2 are positively charged, short-lived intermediates involved in SN1 reactions. Their high electrophilicity enables efficient transformation reactions that directly alkylate non-anionic, low-reactive nucleophiles. In typical SN1 reactions, a leaving group (LG) present in a suitable precursor (R–LG), such as alcohols (LG = OH)3,4,5,6 and alkyl halides (LG = X)7,8 is activated through interaction with an activator (A+) such as Brønsted or Lewis acids (Fig. 1a)9. LG activation by one-electron oxidation10 or photo-irradiation11 has also been reported. The generated carbocation reacts immediately with a co-existing nucleophile to afford the corresponding alkylation product. However, the scope of this reaction methodology is limited owing to competing interactions between the nucleophile and A+ that deactivate/decompose one or both. Only compatible combinations of the nucleophile and A+ allow the desired alkylation to proceed. In particular, unstable carbocations tend to require a strongly acidic activator with high-temperature heating for their generation, which leads to the decomposition of nucleophiles bearing sensitive functionalities. This limitation can be reasonably resolved if the SN1 reactions are conducted stepwise, completing the carbocation generation in another reaction vessel prior to alkylation. However, carbocations generally cannot be accumulated in solution due to their rapid decomposition (Fig. 1b). Whereas exceptionally stable carbocations such as the triphenylmethyl cation12 can be isolated, unstable carbocations, such as primary benzyl cations that lack the conjugation from heteroatom lone pairs, require the use of superacidic media for accumulation at low temperature13,14, which is not suitable for synthetic applications. Thus, the inherent instability of carbocations is a longstanding problem for the synthetic methodology of the SN1 reactions.

Fig. 1: Direct alkylation of low-reactive nucleophiles.
figure 1

a Alkylation reaction via an SN1 mechanism involving carbocation generation in the presence of a nucleophile. LG leaving group, Nu nucleophile, A+ activator. b Stepwise SN1 reaction process limited by the inherent instability of carbocations. c Conceptual energy diagram illustrating the coordinative stabilization of carbocation intermediates with a ligand to form carbocationoids. L ligand. d Benzylation of nucleophiles through a two-step procedure of carbocationoid formation followed by alkylation.

Full size image

Yoshida developed the cation pool method for the accumulation of electrochemically generated carbocations15,16 as a methodologically different approach to the SN1 reactions involving LG activation. This method has been applied to relatively stable carbocations (e.g., heteroatom-stabilized carbocations and secondary benzyl cations) that can be generated by the selective oxidation of precursors.

Stabilization by complex formation allows for the accumulation of highly reactive intermediates without compromising their characteristic reactivity, as exemplified by carbenoids17, used as “preserved carbenes” for synthetic purposes. Applying this idea to the carbocation intermediates in the SN1 reactions leads to “preserved carbocations” formed by coordinative stabilization with an external ligand (Fig. 1c). If these accumulable complexes retain carbocation-like alkylating ability, they can likewise be termed carbocationoids. We herein report a distinct synthetic methodology to conduct SN1 reactions in an alternative two-step procedure using carbocationoids, which is demonstrated by the benzylation of neutral nucleophiles (Fig. 1d). In step 1, these benzylic carbocationoids are formed from the corresponding benzyl alcohols in a nucleophile-free solution. In step 2, the alkylation of nucleophiles proceeds at ambient temperature, even under mildly basic conditions, because the addition of acidic LG activators is not required.

Results and Discussion

Design and preparation

The carbocationoids employed in this study were designed based on our previously developed acid-catalyzed O-benzylating reagent 1 (Fig. 2a)18. Upon the protonation of reagent 1 with trifluoromethanesulfonic acid (TfOH) in 1,4-dioxane, the resulting 1-H+ released benzyl cation species (benzyl trifluoromethanesulfonate)19 and triazinedione 2 with a half-life of 47 min at 25 °C. Using 2 as the ligand for the benzyl cation, the structural modification of 1-H+ led to the benzylic carbocationoids 3a and 3b (Fig. 2b), derived from triazinedione ligands 4a and 4b, respectively, with the following features: (1) N,N’-Dimethyl groups were introduced to the triazinedione skeleton20 to prevent deactivation by N-deprotonation. (2) Ligand 4a possesses an O-neopentyl group that is resistant to dealkylation in place of the O-Me group in 1-H+ that is susceptible to demethylation21. (3) A 4-t-butyl group has been introduced to the benzyl groups of 3a and 3b for experimental convenience, as it increases their solubility in organic solvents and reduces the volatility of the alkylation products of nucleophiles. (4) The benzyl group of 3b bears electron-donating 2,6-dimethyl groups that facilitate C–O bond cleavage at the benzylic position. This electronic effect can be compensated for by the presence of the morpholino group in 4b, whose strong electron donating ability enhances the coordinating ability of the ligand.

Fig. 2: Design and preparation of carbocationoids.
figure 2

a Acid-catalyzed O-benzylation reaction of reagent 1 involving 1-H+ that generates benzyl cation species. b Formation of carbocationoids 3 from ligands 4 and benzyl alcohols 5 by trifluoromethanesulfonic anhydride (Tf2O)-mediated dehydrative condensation.

Full size image

CH2Cl2 solutions of carbocationoids 3 were prepared by dehydrative condensation of the corresponding ligands 4 (1 equiv.) and benzyl alcohols 5 (1.1 equiv.) using trifluoromethanesulfonic anhydride (Tf2O, 1 equiv.) in the presence of a sterically hindered tert-amine base, 1,2,2,6,6-pentamethylpiperidine (pempidine, 1.2 equiv.). These materials were combined at −78 °C and then warmed to 0 °C to complete the carbocationoid formation.

Preservability and alkylating ability

To evaluate the stability and reactivity of these carbocationoids, we conducted the O-alkylation of (10-acetoxy)decanol (6, 1 equiv.) in the presence of pempidine using 3a and 3b (prepared from 2.2 and 2.0 equiv. of 4a and 4b, respectively) by changing the preservation time at 0 °C prior to use (Table 1). 1,4-Dioxane was used as a co-solvent, as our previous studies indicated that ethereal solvents have a favorable effect on the product yields of O-alkylation reactions involving benzylic carbocation species22,23,24. In entry 1, the O-alkylation of 6 using 3a (preserved for 1 h) proceeded at room temperature to afford ether 7a in 89% nuclear magnetic resonance (NMR) yield and 88% isolated yield. Significantly, the yield remained at 85% even with an extended preservation time of 20 h (entry 2). By contrast, a control experiment conducted without the use of ligand 4a did not form 7a (entry 3). These results indicate the critical role of 4a and the sufficient stability of 3a in the absence of suitable nucleophiles. The reaction shown in entry 4 used 3b (preserved for 1 h) as the electrophile to afford 7b in 89% yield (or 87% yield using N-isobutylmorpholine instead of pempidine). Furthermore, the satisfactory stability of 3b was confirmed by extending the preservation time to 20 h (89% NMR yield and 87% isolated yield, entry 5). Entry 6 was an attempt to synthesize 7b without the use of ligand 4b. In this case, benzyl alcohol 5b was treated directly with Tf2O and pempidine at −78 °C for 1 h before the addition of 6. As the in situ-formed carbocation species reacted rapidly with 5b even at the low temperature, symmetric ether 8b was the major product (95% yield based on 5b), with only a trace quantity of 7b detected (<3% based on 6). Thus, the coordinative stabilization of the ligand 4b to form 3b was crucial for the successful alkylation of 6.

Table 1 O-Alkylation of alcohol using carbocationoids 3
Full size table

Characterization by NMR spectroscopic analysis

The 1H and 13C{1H} NMR spectra (together with the 2D NMR data) of carbocationoids 3 in CDCl3 supported their structures (Supplementary Tables 1 and 2). The 1H NMR spectral comparisons between carbocationoids 3, benzyl alcohol 5, and ligands 4 are shown in Supplementary Fig. 1. The benzylic proton signal of 3b (5.60 ppm) is downfield shifted than that of benzyl alcohol 5b (4.72 ppm), while it is largely upfield shifted than that of the corresponding carbocation (8.67 ppm, measured in SbF5–SO2)14, reflecting the coordinatively stabilized carbocation character of 3b. Upon the addition of nucleophile 6 and pempidine at room temperature, the 1H NMR signals of the carbocationoids 3 decreased and disappeared with time (~4 h), accompanied by increase of the corresponding ethers 7 and ligands 4, respectively (Supplementary Fig. 2). The yields of 3a and 3b freshly prepared in CDCl3 were 69% and 62% (based on 4a and 4b), respectively (Supplementary Fig. 3). Symmetric ethers 8 were also formed in a mole percentage of 15–24% relative to the quantity of ligands 4 used. The majority (~70%) of the initially-formed carbocationoids 3 were still present after 20 h at 0 °C, indicating their preservability at this temperature.

Reaction generality

Carbocationoids 3 were utilized for the alkylation of various nucleophiles (9–23, Fig. 3a). Benzhydrol (9), tert-alcohol 10, and 1-adamantanol (11) are suitable electrophiles for SN1 reactions because they produce the corresponding carbocations in the presence of acidic activators25,26,27. Nevertheless, these acid-labile alcohols 9–11 were successfully converted to the ether products 24–26 in 57–75% yield when treated with carbocationoid 3a under mildly basic conditions. Furthermore, the acid-sensitive triphenylmethyl ether in alcohol 12 remained intact during the alkylation reaction with 3a, resulting in 93% yield of the ether product 27. In contrast, such a triphenylmethyl ether functionality was completely decomposed under the acidic reaction conditions reported for the SN1-type alkylation using 9 as the electrophile26,27 (see Supplementary Fig. 4 for details). Acid-catalyzed SN1 reactions using an alcohol (1 equiv.) as the carbocation precursor typically require an excess amount (2 equiv. or more) of a nucleophile to suppress the competing side reaction that forms the symmetric ether derived from the precursor alcohol. Consequently, completing multiple alkylation of polyol nucleophiles using this method is difficult. In contrast, the dialkylation of diol 13 was achieved with 3a, resulting in 81% yield of the product 28. The C-alkylation of silyl ketene acetal 14 and silyl enol ethers 15–17 with 3a afforded the corresponding carbonyl compounds 29–31 in 53–80% yield. The smooth reactions of the neutral nucleophiles 14–17 underscore the carbocation-like alkylating ability of 3a. Notably, amide 18 underwent C-alkylation at the α-position of the carbonyl group to afford 32 in 77% yield (reaction mechanism predicted in Supplementary Fig. 5). This reaction process may be initiated by the O-alkylation of the weakly nucleophilic carbonyl oxygen atom within the amide, similar to the O-alkylation of amides using benzyl cation equivalents that we have previously reported in the study of an amide cleavage reaction28. When the alkylation reactions using 3b were conducted for alcohols 9–11, silyl ketene acetal 14, and silyl enol ethers 15 and 19, the corresponding O- and C-alkylated products 33–38 were obtained in 60–94% yield. Furthermore, allylsilane and allylstannane compounds 20–23 reacted with 3b to afford the alkylated products 39–41 in 64–92% yield. Again, the efficient alkylation of these neutral nucleophiles highlights the carbocationic reactivity of 3b.

Fig. 3: Alkylation reactions using carbocationoids.
figure 3

a Reactions using carbocationoids 3a and 3b. Isolated yields are presented unless otherwise noted. * indicates a yield calculated by 1H NMR spectroscopy using an internal standard. The symbols indicate the reaction conditions: † 3a [from 4a (3.3 equiv.)] and N-isobutylmorpholine (3.1 equiv.); ‡ 3a [from 4a (2.2 equiv.)] and pempidine (2.1 equiv.); § 3a [from 4a (6.6 equiv.)] and N-isobutylmorpholine (6.3 equiv.); || 3a [from 4a (3.3 equiv.)] and pempidine (3.1 equiv.); 3a [from 4a (4.4 equiv.)] and pempidine (4.2 equiv.); # 3b [from 4b (3.0 equiv.)] and N-isobutylmorpholine (3.5 equiv.); ✩ 3b [from 4b (2.0 equiv.)] and pempidine (2.3 equiv.); ** 3b [from 4b (3.0 equiv.)] and N-isobutylmorpholine (1.5 equiv.). b Secondary alkylation reaction using carbocationoid 3c.

Full size image

Our two-step alkylation procedure successfully achieved the secondary O-alkylation of nucleophile 6 at room temperature (Fig. 3b). The treatment of nucleophile 6 with a solution of carbocationoid 3c, which was prepared from secondary alcohol 5c and ligand 4b, afforded the ether product 42 in 76% yield.

Comparison to other alkylation methods

To investigate the reactivity of neutral nucleophiles toward alkyl halides, the reaction of alcohol 6 with benzyl iodide 43a or 43b in the presence of pempidine was monitored by 1H NMR spectroscopy in CDCl3 at 25 °C (Fig. 4a). The observations revealed that no product ether 7a or 7b was detected after 24 h (Supplementary Fig. 6), whereas 6 and 43a or 43b remained unreacted. Likewise, no formation of 30 or 37 was observed when silyl enol ether 15 was used as the nucleophile (Fig. 4b and Supplementary Fig. 7). Therefore, under the reaction conditions employed, nucleophiles 6 and 12 lacked reactivity with alkyl iodides, which are common reagents for alkylation via an SN2 mechanism. These results support the carbocationic reactivity of 3a and 3b observed in Fig. 3a.

Fig. 4: Comparison of alkylation methods.
figure 4

a Attempts at the alkylation of alcohol 6 using benzyl iodides 43a or 43b. b Attempts at the alkylation of silyl enol ether 15 using benzyl iodides 43a or 43b. c Selective formation of asymmetric ether 8ad through a stepwise procedure of carbocationoid formation followed by alkylation. d Unselective ether formation reactions promoted by an acidic catalyst system. * indicates a ratio or a yield calculated by 1H NMR spectroscopy. † indicates an isolated yield. e The N-alkylation of sterically hindered secondary amine 44 using carbocationoid 3b. f Attempts at the acid-catalyzed N-alkylation of 44 using trichloroacetimidate 46 and trifluoromethanesulfonic acid.

Full size image

Next, we conducted the selective preparation of asymmetric ether 8ad from benzylic alcohols 5a and 5d (Fig. 4c) to emphasize the synthetic advantages of our stepwise alkylation methodology utilizing carbocationoids. When 5a (1.5 equiv.) was used as the precursor of carbocationoid 3a and 5d (1 equiv.) as the nucleophile, the product ratio of 5a-derived ether 8a, asymmetric ether 8ad, and 5d-derived ether 8d was 5:95: < 1. The isolated yield of 8ad was 82% based on 5d. Exchanging the roles of the benzylic alcohols (5d used as the precursor of carbocationoid 3d and 5a as the nucleophile) produced similar results. The product ratio of 8a:8ad:8d was <1:92:8, and the yield of 8ad was 80% based on 5a. In contrast, when the ether formation was conducted under SN1-type reaction conditions using a boronic acid–oxalic acid catalyst system29 (Fig. 4d), the product ratio of 8a:8ad:8d was 29:50:21 [with 5a (1.5 equiv.) and 5d (1 equiv.); 25% yield of 8ad based on 5d] or 44:44:12 [with 5a (1 equiv.) and 5d (1.5 equiv.); 32% yield of 8ad based on 5a]. These results indicate that the formation of carbocationoids 3a or 3d was crucial for the selective formation of 8ad.

Finally, we carried out the N-alkylation of a hindered secondary amine, 2,2,6,6-tetramethylpiperidine (44), to demonstrate the effectiveness of our alkylation methodology in contrast to acid-catalyzed alkylation using a trichloroacetimidate reagent. The N-alkylation of 44 using 3b afforded the desired product 45 in 64% yield, despite the significant steric hindrance around the reaction site (Fig. 4e). We subsequently attempted the acid-catalyzed alkylation of 44 to 45 using trichloroacetimidate 46 (Fig. 4f). Benzylic trichloroacetimidates are expected to release carbocation species via an SN1 mechanism upon activation of the LG by protonation30,31. However, the use of TfOH (0.2 equiv.) as an acid catalyst for 46 did not yield any product. This can be attributed to the neutralization of TfOH by amine 44. Increasing the amount of TfOH to 2 equiv. led to the decomposition of 46. However, the desired product 45 could not be obtained because amine 44 was completely protonated by the excess TfOH, and thus, lost its nucleophilicity. These results clearly delineate the limitations of the conventional acid-catalyzed alkylation, where the LG requires activation in the presence of a nucleophile.

Conclusions

We demonstrated that the methodological limitations of SN1 reactions can be overcome by conceptually extracting their highly reactive intermediates in the form of carbocationoids that can be accumulated and preserved in a nucleophile-free solution. The methodology presented in this study has the potential to be applied to other carbocations by modifying the coordinating ability of the ligand, and has the potential to provide new opportunities in carbocation chemistry.

Methods

General procedure for the preparation of carbocationoid 3a in CH2Cl2 (GP-1)

Tf2O (37.0 μL, 0.22 mmol, 1.0 equiv.) was added dropwise to a suspension of ligand 4a (50.0 mg, 0.22 mmol, 1.0 equiv.), benzyl alcohol 5a (40.6 μL, 0.24 mmol, 1.1 equiv.), pempidine (47.0 μL, 0.26 mmol, 1.2 eqiuv.), and powdered molecular sieves 4 A (33.3 mg, a dehydrating agent to remove residual moisture in reaction mixtures) in CH2Cl2 (1.33 mL) at –78 °C. After 10 min, the reaction mixture was warmed to 0 °C and stirred for a period of the indicated preservation time. The supernatant was used for the alkylation reactions of nucleophiles.

Synthesis of (10-Acetoxy)decyl 4-(tert-butyl)benzyl ether (7a)

A solution of carbocationoid 3a in CH2Cl2 [prepared from ligand 4a (50.0 mg, 0.22 mmol) following GP-1, the preservation time of 1 h] was added to a suspension of nucleophile 632 (21.6 mg, 0.10 mmol), pempidine (37.9 μL, 0.21 mmol), and powdered molecular sieves 4 A (41.7 mg) in 1,4-dioxane (0.33 mL) at room temperature. After 19 h, the reaction mixture was passed through a silica pad (EtOAc as an eluent). The eluent was concentrated under reduced pressure. The residue was purified by silica gel column chromatography (hexane/EtOAc = 9:1) and preparative thin layer chromatography (hexane/EtOAc = 19:1) to afford a clear colorless oil (32.0 mg, 88%). 1H NMR (600 MHz, CDCl3): δ 7.39–7.35 (m, 2H), 7.29–7.25 (m, 2H), 4.47 (s, 2H), 4.05 (t, J = 6.9 Hz, 2H), 3.46 (t, J = 6.7 Hz, 2H), 2.04 (s, 3H), 1.65–1.57 (m, 4H), 1.40–1.22 (m, 21H); 13C{1H} NMR (150 MHz, CDCl3): δ 171.4, 150.5, 135.8, 127.6, 125.4, 72.8, 70.6, 64.8, 34.6, 31.5, 29.9, 29.62, 29.58, 29.4, 28.7, 26.3, 26.0, 21.2; HRMS (DART): calcd for C23H39O3 [M + H]+: 363.2899; found: 363.2905.

General procedure for the preparation of carbocationoid 3b in CH2Cl2 (GP-2)

Tf2O (65.4 µL, 0.40 mmol, 1.0 equiv.) was added dropwise to a suspension of ligand 4b (90.5 mg, 0.40 mmol, 1.0 equiv.) and powdered molecular sieves 4 A (53.2 mg) in CH2Cl2 (1.20 mL) at –78 °C. After 30 min, a solution of pempidine (86.6 µL, 0.48 mmol, 1.2 equiv.) and benzyl alcohol 5b (84.6 mg, 0.44 mmol, 1.1 equiv.) in CH2Cl2 (1.46 mL) was added dropwise at –78 °C. After 5 min, the reaction mixture was warmed to 0 °C and stirred for a period of the indicated preservation time. The supernatant was used for the alkylation reactions of nucleophiles.

Synthesis of (10-Acetoxy)decyl (4-tert-butyl-2,6-dimethyl)benzyl ether (7b)

A solution of carbocationoid 3b in CH2Cl2 [prepared from ligand 4b (90.5 mg, 0.40 mmol) following GP-2, the preservation time of 20 h] was added to a suspension of nucleophile 632 (43.2 mg, 0.20 mmol), pempidine (83.0 μL, 0.46 mmol), and powdered molecular sieves 4 A (66.6 mg) in 1,4-dioxane (0.67 mL) at room temperature. After 22 h, the reaction mixture was treated with H2O (0.1 mL). After 10 min, the mixture was passed through a silica pad (EtOAc as an eluent). The eluent was concentrated under reduced pressure. The residue was purified by silica gel column chromatography (hexane/EtOAc = 9:1) and preparative thin layer chromatography (hexane/EtOAc = 9:1) to afford a clear colorless oil (68.1 mg, 87%). 1H NMR (600 MHz, CDCl3): δ 7.03 (s, 2H), 4.47 (s, 2H), 4.05 (t, J = 6.8 Hz, 2H), 3.49 (t, J = 6.5 Hz, 2H), 2.39 (s, 6H), 2.05 (s, 3H), 1.65–1.55 (m, 4H), 1.40–1.22 (m, 21H); 13C{1H} NMR (150 MHz, CDCl3): δ 171.4, 150.7, 137.5, 131.8, 125.4, 70.8, 67.1, 64.8, 34.4, 31.4, 30.0, 29.65, 29.59, 29.57, 29.4, 28.7, 26.4, 26.0, 21.2, 20.0; HRMS (DART): calcd for C25H43O3 [M + H]+: 391.3212; found: 391.3204.

General information, experimental procedure and characterization data

For general information, see Supplementary Methods (page S11). For experimental procedure and characterization data, see Supplementary Methods (pages S12–S27).

NMR spectra

For 1H, 13C{1H}, and 2D NMR spectra, see Supplementary Data 1.