2020, Vol.49, No.7


An efficient method to synthesize allyl sulfides from sulfinate esters and allylsilanes is described. Based on the reactivity of isolated allyl(methoxy)phenyl sulfonium triflate, we have developed a simple one-pot method for the allyl sulfide synthesis by S-allylation of sulfinate esters and subsequent reduction with sodium borohydride. A number of allyl sulfides were prepared by this method leaving various functional groups unreacted.

Organosulfur compounds have played important roles in broad disciplines including not only organic chemistry but also pharmaceutical sciences, materials chemistry, and chemical biology.1,2 Particularly, allyl sulfides have gained attention due to their synthetic versatility.3 A wide range of transformations of allyl sulfides have been reported such as thiol-ene reaction,3g olefin cross-metathesis,3c catalytic allylthiolation of alkynes,3b and insertion reactions with diazo compounds through [2,3]-sigmatropic rearrangement3a,3h catalyzed by rhodium, copper, and so on (Figure 1A). In general, allyl sulfides are synthesized by S-allylation of thiols (Figure 1B). However, accessible allyl sulfides are still limited due to stinking, oxidizable, and highly nucleophilic thiols. Herein, we describe a novel method to prepare allyl sulfides by S-allylation of sulfinate esters and subsequent reduction of the resulting alkoxysulfonium intermediates.

Alkoxysulfonium salts4,5 are well-known intermediates in the Swern oxidation, in which carbonyl compounds are prepared from alcohols with dimethyl sulfoxide, oxalyl chloride, and triethylamine at low temperature.6 Pioneering studies on reactivities of isolated alkoxysulfonium salts were reported by Jonson and Phillips in the 1960s.4a4c For example, O-methylation of sulfoxides with trimethyloxonium tetrafluoroborate provided a limited number of methoxysulfonium salts, which smoothly reacted with sodium borohydride to afford sulfides.4c In 1974, Durst and coworkers reported reduction using sodium cyanoborohydride with 18-crown-6 instead of sodium borohydride.4d Although several sulfides were efficiently synthesized from alkoxysulfonium salts, the harsh conditions preparing alkoxysulfonium salts using the highly electrophilic alkylating reagent have hitherto limited the synthetic utility.

In the course of our studies on organosulfur chemistry,7 the recent success of allyl sulfoxide synthesis from sulfinate esters8 motivated us to revisit the alkoxysulfonium salt chemistry.9 Indeed, we recently found that the interrupted Pummerer-type reaction of sulfinate esters 1 using trifluoromethanesulfonic anhydride (Tf2O)10,11 in the presence of allylsilanes 2 efficiently afforded allyl sulfoxides through alkoxysulfonium intermediates 4 (Figure 1C).8 In this recent study, we succeeded in the isolation of allyl(methoxy)phenylsulfonium triflate (4a) by avoiding aqueous work-up. Since sodium borohydride reduction of 4a furnished allyl phenyl sulfide (5a), we decided to examine the reactivity of 4a to establish the allyl sulfide synthesis from sulfinate esters 1 (Figure 1D).

We first screened a variety of conditions for the reduction of alkoxysulfonium salt 4a (Table 1). As a result, although the reaction in methanol afforded sulfide 5a in high yield (Entry 1),8 changing the solvent to ethanol, tetrahydrofuran (THF), or dichloromethane significantly decreased the yield of sulfide 5a (Entries 2–4). Adding 18-crown-6 or 15-crown-5 slightly improved the efficiency of the reduction in dichloromethane (Entries 5 and 6). According to the reported conditions by Durst and coworkers,4d reduction of 4a with sodium cyanoborohydride in methanol proceeded smoothly (Entry 7). Sulfide 5a was also obtained in moderate yields by the reduction using sodium cyanoborohydride with 18-crown-6 or 15-crown-5 in dichloromethane (Entries 8 and 9). When sulfonium salt 4a was treated with lithium borohydride, sulfide 5a was obtained in low yield (Entry 10). The reaction of 4a with lithium aluminum hydride or diisobutylaluminum hydride also proceeded, although the yields of 5a were low (Entries 11 and 12). In addition, triethylsilane and 1,4-cyclohexadiene were ineffective to reduce sulfonium salt 4a (Entries 13 and 14).

Table 1. Screening of reducing conditions
Table 1. Screening of reducing conditions
Entry Reductant Solvent Yield/%a
1b NaBH4 MeOH 80
2 NaBH4 EtOH 37
3 NaBH4 THF 20
4 NaBH4 CH2Cl2 15
5 NaBH4 + 18-crown-6 CH2Cl2 36
6 NaBH4 + 15-crown-5 CH2Cl2 31
7 NaBH3CN MeOH 78
8 NaBH3CN + 18-crown-6 CH2Cl2 57
9 NaBH3CN + 15-crown-5 CH2Cl2 55
10 LiBH4 MeOH 27
11 LiAlH4 THF 36
12 i-Bu2AlH THF 29
13 Et3SiH THF 0
14 1,4-cyclohexadiene CH2Cl2 0

a 1H NMR yields. bData from ref 8.

To gain insight into the reactivity of sulfonium salt 4a, we performed a theoretical calculation (Figures 2A and 2B). As a result, the LUMO of the optimized structure showed a large coefficient at the sulfur atom, suggesting the electrophilic nature at the sulfonium sulfur. In addition, 1H NMR analysis of sulfonium salt 4a in methanol-d4 showed that most of the sulfonium salt 4a was observed without incorporation of trideuteriomethoxy group after stirring for 10 min, and 4a completely decomposed after 24 h. This result indicated that cleavage of S–O bond resulting to exchange the methoxy group with the solvent did not proceed in the protic solvent and sulfonium salt 4a slowly decomposed in methanol.12

A deuteration experiment using NaBD4 afforded deuterium-free sulfide 5a showing that deprotonation of sulfonium salt 4a did not take place despite the presence of highly acidic allylic proton of 4a (Figure 2C). This result strongly supports the plausible reaction mechanism shown as path A in Figure 2D; substitution reaction at the sulfonium sulfur with hydride ion and subsequent deprotonation lead to sulfide 5a, similar to the pioneering study using methyl(methoxy)phenylsulfonium tetrafluoroborate.4c On the other hand, the mechanism shown as path B in Figure 2D is excludable, since the deprotonation was not observed in the deuteration experiment (Figure 2C).

We then turned our attention to achieve the allyl sulfide synthesis by performing the S-allylation and subsequent reduction in a one-pot manner (Scheme 1). As a result, treatment of sulfinate ester 1a and allyltrimethylsilane (2a) with Tf2O followed by the addition of methanol and sodium borohydride (3.0 equiv) provided sulfide 5a in good yield. Reducing the stoichiometry of sodium borohydride to 1.0 equiv resulted in a decrease of the yield.

A wide variety of allyl sulfides were successfully prepared by the one-pot method (Figure 3). Indeed, allyl sulfide 5b was synthesized from electron-deficient methyl p-chlorophenylsulfinate in excellent yield. Naphthylsulfinate ester also participated in this reaction to afford allyl sulfide 5c. Moreover, the reaction of bulky substrate such as methyl o-bromophenylsulfinate uneventfully proceeded to afford sulfide 5d. The Pummerer-type allylation of sulfinate ester 1a with 2-phenylpropenyl(trimethyl)silane followed by the reduction also smoothly took place to afford sulfide 5e bearing a substituent on the alkene moiety. It is worth noting that allyl sulfides 5f and 5g having chloro and acetoxy groups, respectively, at the allylic position were synthesized in good yields keeping these groups with high leaving-group ability intact. Since allyl sulfides are generally prepared from thiols and allyl halides, these results clearly showed an advantage of the present method for preparing allyl sulfides bearing leaving groups.

In summary, we have developed a facile method to synthesize allyl sulfides from sulfinate esters and allylsilanes. On the basis of a number of examinations using alkoxysulfonium salt 4a, an efficient one-pot procedure by S-allylation and the following reduction was established. Further studies to examine the scope and limitations of the method and application to the synthesis of bioactive organosulfur compounds are now underway.

The authors thank Dr. Yuki Sakata at Tokyo Medical and Dental University for HRMS analyses. This work was supported by JSPS KAKENHI Grant Numbers JP19K05451 (C; S.Y.), JP18H02104 (B; T.H.), JP18H04386 (Middle Molecular Strategy; T.H.), and 19J14128 (JSPS Research Fellow; T.M.); the Naito Foundation (S.Y.); the Japan Agency for Medical Research and Development (AMED) under Grant Number JP19am0101098 (Platform Project for Supporting Drug Discovery and Life Science Research, BINDS); and the Cooperative Research Project of Research Center for Biomedical Engineering.

Supporting Information is available on https://doi.org/10.1246/cl.200285.