Novel Behavior of Thiiranium Radical Cation Intermediates. Reactions of Dimethyl Disulfide with Alkenes in the Presence of Pd(OAc)₂

Abstract

Reaction of dimethyl disulfide (1) with cyclohexene (2a) in AcOH in the presence of Pd(OAc)₂ yields trans-1-acetoxy-2-methylcyclohexane (3a). The equivalent reactions with hex-1-ene (2b) and 2-methylpent-1-ene (2c) or 1-methylcyclohex-1-ene (2d) preferentially give anti-Markovnikov and Markovnikov adducts 4 and 3, respectively, by acetoxymethylthiolation of the alkene. The Markovnikov regioselectivity 3b/4b for the reaction with 2b is higher than that for the reaction using AgOAc instead of Pd(OAc)_2, which proceeds via a thiiranium ion. Addition of a polar solvent (MeCN or MeNO₂) to the reactions with 2b or 2c using Pd(OAc)₂ abnormally decreases the Markovnikov regioselectivity. The total yield of 3 and 4 increases with an increased concentration of AcOH. Compounds 3 and 4 are also formed and the reactions in MeCN or MeNO₂ not containing AcOH. A solution of Pd(OAc)₂ in 1 exhibits λ_max 380 nm (log ε 3.6) assigned to the absorption of a relatively stable sulfonium salt. These indicate that the reactions using Pd(OAc)₂ proceed by S_N2 ring-opening of a new type of thiiranium radical cations paired with ⁻OAc via the sulfonium salts. The insensitivity of the 3/4 ratios to the reaction time at 25-60°C in the reactions with 2c-d shows the ring-opening to be controlled kinetically, but the increased ratio with reaction time at 116°C in the reaction with 2b suggests that the ring-opening is thermodynamically governed. The reaction product with 2d also undergoes a skeletal rearrangement to a thietanium radical cation to give 1-acetoxymethyl-2-methylthiocyclohexane.

Introduction

Thiiranium ion intermediates have been generated by reactions of disulfides with alkenes in the presence of oxidants such as Pb(OAc)₄ [1,2,3,4], Mn(OAc)₃ [2,3,4], Fe(OAc)₃ [2,3,4] or Cu(OAc)₂ [5], by reactions of β-thiohalides with AgY [6] and by reactions of thiiranes with RY [6]. The intermediates which have soft counterions such as BF₄⁻, PF₆⁻, SbF₆⁻, SbCl₅⁻, FSO₃⁻ and CF₃CO₂⁻ behave like carbenium ions [1] of a weakly bridged strong electrophile, undergoing a rapid ring-opening by the outer nucleophile at low temperature (−20-0^oC) to give the observed predominant Markovnikov regioselectivity; in some cases, the skeletal rearrangement or stereoconversion of the thiiranium ions has been observed. We may infer that the reactions of RSCl with alkenes would generate not a σ-sulfurane but a strongly bridged thiiranium ion pairing tightly with hard Cl⁻ [1], and thus the ring-opening by the pairing Cl⁻ leads to the anti-Markovnikov regioselectivity without undergoing the stereoconversion and skeletal rearrangement.

In this work, reactions of dimethyl disulfide with 2-alkyl and 2,2-dialkylalkenes in AcOH in the presence of Pd(OAc)₂ lead to the anti-Markovnikov and Markovnikov regioselectivity products, respectively, for the addition to the alkenes. Interestingly, the Markovnikov regioselectivity is lower than that observed for thiiranium ion intermediates, and is decreased by addition of polar solvents such as MeCN or MeNO₂ or the nonpolar solvent n-hexane. This novel behavior can be reasonably explained by considering a S_N2 ring-opening mechanism of a new type of thiiranium radical cation intermediates paired with ⁻OAc. Furthermore, the reaction with 1-methylcyclohex-1-ene yields a product via an anionotropic rearrangement of the thiiranium radical cation intermediate to a thietanium radical cation intermediate.

Results and Discussion

Reactions with Cyclohexene (2a)

Reaction of dimethyl disulfide (1) with cyclohexene (2a) in AcOH in the presence of Pd(OAc)₂ produced only trans-1-acetoxy-2-methylthiocyclohexane (3a) at 25 or 116^oC (

Table 1. Reactions in the presence of Pd(OAc)₂ (0.68 mmol) of dimethyl disulfide (1) (2.0 mmol) with 4.8 mmol of cyclohexene (2a) and hex-1-ene (2b) in AcOH (10.0 cm³) containing Ac₂O (1.0 cm³).

Alkene	T/oC	t/h	Yielda(%)		Ratio
			3	4	3/4
2a	25	24	69b	-	-
2a	116	24	68b	-	-
2b	0	2.0	5.6	15	0.37
2b	0	4.0	8.8	24	0.37
2b	25	2.0	14	32	0.44
2b	25	24	16	37	0.43
2b	25	96	16	36	0.44
2b	60	1.0	13	29	0.45
2b	60	2.0	20	40	0.50
2b	116	1.0	17	23	0.74
2b	116	2.0	36	34	1.1
2b	116	24	76	17	4.5

^aThe yields are calculated to be 100% when two moles of product is formed from one mole of Pd(OAc)₂ used, and were obtained by GLC analysis; ¹H NMR analysis gave nearly the same results. ^bThe trans-stereochemistry is assigned by the fact that the compound given by hydrolysis of 3a is identical with that formed by S_N2 reaction of cyclohexene oxide with MeSH in benzene.

). The structure of 3a was assigned to the trans-form on accord of the fact that the mass spectra of 3a were identical with those of the product formed by the reaction of cyclohexene oxide with MeSH in benzene (Scheme 1). The trans-configuration shows that 3a arises from an S_N2 ring-opening of a three-membered ring intermediate, and that the intermediate is still stable at 116^oC without transforming to the corresponding carbocation at such a high temperature.

Reactions with Hex-1-ene (2b)

The reaction with hex-1-ene (2b) at 0-60^oC gave preferentially the anti-Markovnikov adduct 4b, and the 3b/4b ratio was observed to increase with temperature (Table 1). The ratio at 0-60^oC was almost insensitive to the reaction time whereas the ratio at 116^oC was enhanced in the prolonged reaction (Table 1). The insensitivity of the ratio to the reaction time means that the reaction is controlled kinetically at 0-60^oC, but thermodynamically governed at 116^oC. In fact, when a mixture of 3b/4b = 0.34 was treated in AcOH, the ratio 3b/4b at 25^oC did not change for 24 h , but the ratio at 116^oC increased to 1.4 or 6.6 after 2 h or 24 h, respectively. The preferential formation of anti-Markovnikov adduct at 0-60^oC suggests that the kinetic S_N2 ring-opening of the final three-membered ring intermediate is principally affected by steric factors (Scheme 2).

Formation of Thiiranium Radical Cations II and II’ from Sulfonium Salts I and I’

From the above trans-addition, the final intermediate for the formation of the products should be logically three membered ring intermediates. As an oxidant Pd(OAc)₂ can act in two ways. The first action is electron transfer: One electron transfer from 1 to Pd(OAc)₂ forms a disulfide radical cation [7], and the radical cation and/or a sulfenium ion [8] arising from the radical cation should generate a thiiranium ion as the final intermediate by attack on the alkene. As seen in Table 1 and

Table 4. Effects of oxidants in the reactions of dimethyl disulfide (1) (2.0 mmol) with hex-1-ene (2b) (4.8 mmol) in the presence of oxidant in AcOH (10 cm³) containing Ac₂O (1.0 cm³) at 25°C for 24 h.

Oxidant	Oxidant/mmol	Yielda(%)		Ratio
		3b	4b	3b/4b
Pd(OAc)2+Cu(OAc)2	b	7.7	17	0.45
Cu(OAc)2	0.68	0	0	-
H2SO4	0.68	trace	trace	-
PdSO4	0.68	trace	trace	-
AgOAc	1.36	5.2	19	0.27
Mn(OAc)3	2.4	0	0	-
Pb(OAc)4	2.4	0	0	-

^aThe yields are calculated to be 100% when two moles of product are formed from one mole Pd(OAc)₂ or Cu(OAc)₂ and from two moles AgOAc. ^bPd(OAc)₂ (0.60 mmol) and Cu(OAc)₂ (3.6 mmol) were used, and the yields are obtained on the basis of Pd(OAc)₂.

, the reaction using Pd(OAc)₂ shows higher Markovnikov regioselectivity (3b/4b = 0.44) at 25°C than that observed at 25°C (0.27) in the reaction via a thiiranium ion [9] using AgOAc instead of Pd(OAc)₂. Furthermore, the Markovnikov regioselectivity 3/4 for the reaction using Pd(OAc)₂ is unusually decreased on dilution with a polar solvent (MeCN or MeNO₂) while that using AgOAc is increased on dilution with MeNO₂ (

Table 3. Effects of solvents on the reaction of dimethyl disulfide (1) (2.0 mmol) with hex-1-ene (2b) (4.8 mmol) or 2-methylpent-1-ene (2c) (4.8 mmol) in the presence of Pd(OAc)₂ (0.68 mmol) at 25°C for 24 h.

Alkene	Solvent (cm3)					Yielda(%)		Ratio
	AcOH	CH2Cl2	n-C6H14	MeCN	MeNO2	3	4	3/4
2b	10	0	0	0	0	16	37	0.43
2b	9.0	1.0	0	0	0	12	29	0.41
2b	5.0	5.0	0	0	0	10	26	0.38
2b	1.0	9.0	0	0	0	6.5	19	0.34
2b	5.0	0	5.0	0	0	5.3	14	0.38
2b	0	0	0	10	0	4.1	17	0.24
2b	0	0	0	0	10	3.2	17	0.19
2b	5.0	0	0	0	5.0	13	42	0.31
2bb	10	0	0	0	0	5.2	19	0.27
2bb	5.0	0	0	0	5.0	5.4	16	0.34
2c	10	0	0	0	0	34	17	2.0
2c	9.0	1.0	0	0	0	31	18	1.7
2c	5.0	5.0	0	0	0	24	18	1.3
2c	1.0	9.0	0	0	0	16	19	0.84
2c	9.0	0	1.0	0	0	32	20	1.6
2c	5.0	0	5.0	0	0	15	12	1.3
2c	1.0	0	9.0	0	0	17	13	1.3
2c	5.0	0	0	0	5.0	47	27	1.7

^aThe yields are calculated to be 100% when two moles of product is formed from one mole of Pd(OAc)₂ and from two moles of AgOAc, and were obtained by GLC analysis; Ac₂O (1.0 cm³) is added to these reactions, but Ac₂O is not used in the reactions with only MeCN or MeNO₂. ^bAgOAc (1.36 mmol) instead of Pd(OAc)₂ was used in this reaction.

). These data show that the reaction using Pd(OAc)₂ does not proceed via a thiiranium ion, and thus this action for Pd(OAc)₂ can be ruled out. Another action is coordination. Pd(OAc)₂ coordinates to 1 to give sulfonium ions I and I’ (Scheme 2). The sulfonium ions undergo a homolytic S-S scission by spontaneous or induced (by AcOH) decomposition to give MeS● and thiiranium radical cations II and II’ which bond with the Pd atom (Scheme 2). If a heterolytic S-S scission of I and I’ occurs, then thiiranium ion intermediates should be formed by reaction with sufenium fragment. Thus, the heterolytic scission can be excluded. Therefore, the latter action predominates and II and II’ as the final intermediates should be formed by the homolytic S-S scission of I and I’ (Scheme 2). The following evidence supports the fact that the products are formed via thiiranium radical cations II and II’ from sulfonium ions I and I’.

A solution containing Pd(OAc)₂ in AcOH also containing Ac₂O (9.1% v/v) shows λ_max 397 nm (log ε 2.3) at 25^oC while a solution made by dissolving Pd(OAc)₂ in 1 at 25^oC has λ_max 380 nm (log ε 3.6). The species with λ_max 380 nm shows the half life-time ca. 30 h in 1, so that the species could not be the unstable disulfide radical cation [7] but rather a sulfonium salt I [MeS-S⁺(Me)PdOAc] ⁻ OAc. When different oxidants AgOAc and AlCl₃ are dissolved in 1, we can observe the same λ_max 375 nm with a very short lifetime. Thus, the species with λ_max 375 nm might be the disulfide radical cation, which would be independent of oxidative metals [7]. Considering that the intermediate using AlCl₃ can bring about an aromatic methylthiolation [7] whereas that using AgOAc is not subject to the aromatic methylthiolation but can react with alkene to form a thiiranium ion [9], we can regard the following: a very small amount of the disulfide radical cation and a more large amount of a sulfonium ion [MeS-S⁺(Me)Ag] ⁻OAc (this has no visible absorption due to a σ S-Ag bond) are generated as the intermediates when AgOAc is dissolved in 1. Dissolving AlCl₃ in 1, the disulfide radical cation is only formed as the intermediate.

During the reaction of 1 with 2b using Pd(OAc)₂, we obtained a small amount of a dark reddish powdery precipitate having υ_max (KBr)/cm^-1 2900 and 2980 (C-H), 1577 and 1410 (−CO₂⁻), 1050, 1020, 950, 930 and 652 (C-S). This precipitate might be a sulfonium salt I’ [MeS-S⁺(Me)Pd] ⁻OAc; this salt does not have ⁻OAc attached to Pd since the carbonyl of Pd(OAc)₂ appears at ca. 1610 cm^-1. Considering that one mole of product is formed from one mole of Pd(OAc)₂, then we observe that the total yield of 3 and 4 is more than 100%. Thus, the generation of the two species I and I’ from one mole of Pd(OAc)₂ is reasonable.

The novel behavior for the reaction using Pd(OAc)₂ instead of AgOAc (i.e. the high 3b/4b ratio using Pd(OAc)₂ relative to AgOAc and the decreased 3b/4b ratio upon addition of polar solvent) as described above is probably due to the higher positive charge on the C₂-carbon of II than that of the corresponding thiiranium ion since II contains a positive Pd atom; the effect of polar solvent is explained below. The possibility that the final intermediate is a thiiranium dication formed by attack of thiiranium ion on Pd(OAc)₂ can be excluded on account of charge-repulsion between the positive species on the attack.

Reactions with 2-Methylpent-1-ene (2c)

The reaction with 2-methylpent-1-ene 2c gave the Markovnikov regioselectivity exhibiting 3c/4c of ca. 1.8 at 0-60°C, and the ratio was almost insensitive to reaction time (

Table 2. Reactions of dimethyl disulfide 1 (2.0 mmol) with 2-methylpent-1-ene 2c and 1-methylcylohex-1-ene 2d of 4.8 mmol in the presence of Pd(OAc)₂ (0.68 mmol) in AcOH (10 cm³) containing Ac₂O (1.0 cm³).

Alkene	T/oC	t/h	Yielda(%)			Ratio
			3	4	5	3/4
2c	0	2.0	15	8.5	0	1.8
2c	0	4.0	20	11	0	1.8
2c	25	2.0	18	10	0	1.8
2c	25	24	34	17	0	2.0
2c	60	2.0	24	13	0	1.8
2c	60	4.0	47	26	0	1.8
2d	0	2.0	15	4.1	3.2	3.7
2d	0	4.0	26	7.4	2.8	3.5
2d	25	2.0	30	10	5.3	3.0
2d	25	24	39	14	3.5	2.8
2d	60	2.0	38	14	3.5	2.7
2d	60	4.0	44	20	4.1	2.2

^aThe yields are calculated as shown in Table 1. The products 3d or 4d correspond to trans-1-acetoxy-1- or -2-methyl-2-methylthiocyclohexane, respectively.

). The time-insensitivity reflects the product stability, and means that the ring-opening of the intermediates IIc and II’c is controlled kinetically. The Markovnikov regioselectivity indicates that the ring-opening of IIc and II’c is mostly governed electronically. The reaction with 2c at 116^oC is described below.

Reactions with 1-Methylcyclohex-1-ene (2d)

The reaction with 1-methylcyclohex-1-ene (2d) at 25-60°C gives predominantly Markovnikov trans-adduct (3d/4d = 3.7-2.2) (Table 2). This can be also interpreted by the idea that the S_N2 ring-opening of IId and II’d occurs under preferential electronic control. A slightly lower ratio in the prolonged reaction (Table 2) is presumably due to the thermodynamic instability of 3d compared to 4d.

We obtained 1-acetoxymethyl-2-methylthiocyclohexane (5d) in the reaction with 2d, but did not obtain the corresponding 5c in the reaction with 2c (Table 2). According to the molecular model, the S-atom of IId and II’d is closer to the 2-methyl carbon than that of IIc and II’c, so IId and II’d may rearrange to the thietanium radical cations IIId and III’d by a concerted reaction involving C₂-S bond breaking and hydride migration of the 2-methyl group, leading to 5d (Scheme 3). A similar anionotropic rearrangement from thiiranium ion to thietanium ion has been proposed for the concerted reaction involving C₂-S breaking and methyl migration at tertiary carbon [10,11,12,13,14].

The reaction with 2c or 2d at 116^oC gave many unidentified products according to GLC analysis. A stable tertiary carbocation is probably formed by bridge-cleavage of IIc and II’c or IId and II’d at a high temperature such as 116^oC, hence this result. But IIa and II’a or IIb and II’b, which cannot generate a tertiary carbocation by the cleavage are still bridged at 116^oC, thus yielding 3a or 3b and 4b by the S_N2 ring-opening.

Effects of Solvent and Concentration of AcOH

The total yield of 3 and 4 in the reaction with 2b or 2c in AcOH–CH₂Cl₂ increases with an increase in the concentration of AcOH, and 3b and 4b are also produced in the reaction in MeCN or MeNO₂ not containing AcOH (Table 3). These indicate that 3 and 4 are formed by the S_N2 attack on II and II’ of both AcOH and AcO^– pairing with II and II’ (Scheme 2).

In the reaction with 2b or 2c, the Markovnikov regioselectivity 3/4 was interestingly decreased upon dilution with CH₂Cl₂ or the polar solvents MeCN or MeNO₂, and it was also depressed on dilution with nonpolar n-hexane (Table 3). By contrast, in the reaction of 1 with 2b using AgOAc instead of Pd (OAc)₂, the Markovnikov regioselectivity for the ring-opening of thiiranium ion was increased on dilution with MeNO₂ (Table 3).

The decreased Markovnikov regioselectivity observed on dilution with the polar solvents suggests that the solvation with the polar solvent occurs strongly at the highly charged C-2 of II and II’ rather than at the low charged C-3, suppressing the S_N2 reactivity at C-2 relative to the C-3. The depressed 3b/4b ratio upon dilution with nonpolar n-hexane allows us to infer that this dilution transforms the solvent separated ion-pair of II and II’ into a tight ion-pair in which the counteranion interacts to a greater extent with the more positive C-2 of II and II’ than the less positive C-3 to decrease the S_N2 reactivity more at the C-2 than the C-3, leading to the observed results.

Effects of Oxidants

As described above, the oxidative attack of Pd(OAc)₂ on 1 may form the sulfonium salts I and I’, giving thiiranium radical cations II and II’ and MeS● (Scheme 2); MeS● dimerizes to give 1 because of an unfavorable reaction with 2 [15]. An addition of Cu(OAc)₂ to the reaction with 2b using Pd(OAc)₂ resulted in rather suppressed yields of 3b and 4b (Table 1 and Table 4). The use of Cu(OAc)₂, Mn(OAc)₃ or Pb(OAc)₄ instead of Pd(OAc)₂ in the reaction with 2b did not yield the products 3b and 4b (Table 4); Cu(OAc)₂ forms a complex with the alkene, and thus it would not operate as oxidizing agent. Mn(OAc)₃ is not effective since it oxidizes AcOH [16]. Pb(OAc)₄ is too strong an oxidant to be effective, giving a lot of unidentified products. PdSO₄ is difficult to dissolve in AcOH, so that it could scarcely be effective. The use of H₂SO₄ gave many unidentified products due to the strong acidity, leading to only a trace of the target products. Since AgOAc is complexed with alkene, it would be not so effective for the reaction.

Experimental

General 

IR spectra were obtained on a JASCO Cooperation A-302 spectrometer. ¹H NMR spectra were taken with a Nippondenshi PMX-60SI instrument (J values are given in Hz). GLC-MS were recorded with a Shimadzu QP-5000 spectrometer linked to a Shimadzu GC-17A; the column (15 m × 0.25 mm, coated with a 0.25 μm film of DB-1) is made by J & W Scientific. GLC was performed with a Shimadzu GC-8A using a glass column (1m × 3 mm) packed with 10% SE-30 on 60-80 mesh Chromosorb W (AW-DMCS). Two runs agreed within 3% error for the yields of the products, which were determined by replicate GLC analyses. The isolation or separation (as a mixture) of the products was done by thin layer chromatography (using Merck Silicagel 60PF254). AcOH, CH₂Cl₂, MeCN, MeNO₂, n-hexane and ethyl acetate were purified by standard methods before use. The following compounds were reagent grade (Nacalai Tesque), and used without further purification: Ac₂O, H₂SO₄, Pd(OAc)₂, Cu(OAc)₂, PdSO₄, AgOAc, Mn(OAc)₃, Pb(OAc)₄, dimethyl disulfide (1), cyclohexene (2a), hex-1-ene (2b), 2-methylpent-1-ene (2c) and 1-methyl-cyclohex-1-ene (2d).

Experimental procedures 

Reactions of dimethyl disulfide (1) with cyclohexene (2a), hex-1-ene (2b), 2-methylpent-1-ene (2c) and 1-methylcyclohex-1-ene (2d) in AcOH containing Ac₂O in the presence of Pd(OAc)₂ 

A mixture of 1 and alkene (2a, 2b, 2c or 2d) was added to a mixture of AcOH and Ac₂O in the presence of Pd(OAc)₂ with stirring (using a magnetic stirrer) under the conditions described in Table 1 and Table 2. After the reaction, any precipitate formed was filtered off, and aqueous Na₂CO₃ was added to the filtrate until the solution reached pH > 7. The organic layer was extracted with ethyl acetate. The yields of products 3a-d, 4b-d and 5d were determined by GLC analysis, and the results are summarized in Table 1 and Table 2. Product 3a was isolated by thin layer chromatography, and the structure was confirmed as follows.

Trans-1-acetoxy-2-methylthiocyclohexane (3a) 

Liquid, υ_max (neat)/cm^-1 2930, 2860, 1734, 1450, 1375, 1235 and 1048; δ_H(CDCl₃) 1.08-1.09 (8H, m, CH₂), 2.06 (3H, s, Me), 2.09 (3H, s, Me), 2.33-2.80 (1H, m, SCH) and 4.85-5.08 (1H, m, OCH); GLC-MS, m/z 188 (M⁺), 128, 113, 81, 80, 79, 77, 61 and 53. The GC-MS mass spectra for the compound which is formed by hydrolysis of 3a in aq. NaOH were identical with those for the product [i.e. trans-1-hydroxy-2-methylthiocyclohexane; m/z 146 (M⁺), 98, 87, 81, 80, 79, 70, 61, 57, and 55] given by a reaction of cyclohexene oxide with MeSH in benzene. This indicates that product 3a has a trans-configuration.

The products 3b and 4b were not isolated but rather obtained as a mixture by thin layer chromatography. This mixture showed the following spectral data: υ_max (neat) /cm^-1 2930, 2870, 1744, 1460, 1380, 1240, 1040 and 818; δ_H (CDCl₃) 0.85-1.88 (9H, m, MeCH₂CH₂CH₂), 2.08 (3H, s, SMe or OCOMe), 2.12 (3H, s, OCOMe or SMe), 2.48-2.78 (br, SCH₂ of 3b), 2.78-2.3.08 (br, SCH of 4b), 4.18 (ABq, J=6 Hz, OCH₂ of 4b) and 4.88-5.23 (m, OCH of 3b). The products 3b and 4b have the following mass spectra by GLC-MS: m/z for 3b 130 (M⁺–AcOH), 115 [M⁺− (AcOH+Me)], 87 (n-BuCH⁺OH)), 82 (radical cation of n-BuC≡CH), 67, 61 (MeSCH₂⁺) and 55 (C₄H₇⁺); m/z for 4b 130, 115, 88, 87, 82, 67, 61 and 55. The fact that the relative intensity of m/z 130 and 87 to m/z 55 is higher in 3b than in 4b also suggests these structures. The observation that the GLC peak for 3b has a shorter retention time than that for 4b supports the fact that the AcO group for 3b is more shielded by non polar groups than the AcO group for 4b. The yields of 3b and 4b were obtained using the assumption that these compounds have the same GLC response factors. A mixture (33 mg) containing 3b and 4b (3b/4b = 0.34) was treated in AcOH (10 cm³) containing Ac₂O (1.0 cm³) at 25°C for 24 h and at 116°C for 2 h or 24 h. Product 3c was isolated by thin layer chromatography, and the structure was confirmed by the following spectra.

2-Acetoxy-2-methyl-1-methylthiopentane (3c) 

Liquid, υ_max (neat) /cm^-1 2920, 2870, 1732, 1435, 1370, 1240, 1198, 1140, 1030 and 950; δ_H (CDCl₃) 0.77-1.88 (7H, m, MeCH₂CH₂), 1.47 (3H, s, Me), 1. 96 (3H, s, OCOMe), 2.11 (3H, s, SMe) and 2.98 (2H, s, SCH₂); m/z 130 (M⁺–AcOH), 101 [M⁺−(AcOH+Et)], 87, 82, 67, 61, 59, 55 and 53.

Compound 4c was obtained as a mixture with 3c, and the proposed structure was supported by the GLC-MS results; the GLC peak for 4c has a longer retention time than that for 3c because the AcO group of 4c is less shielded by non polar groups than the AcO group of 3c, and the mass spectra for 4c showed the following peaks: m/z 130, 117 (M⁺−CH₂OAc), 101, 83, 82, 67, 61, 59, 55 and 53). The peak m/z 117 is not observed for 3c, and the appearance of this peak is in good agreement with the proposed anti-Markovnikov structure of 4c. The yield of 4c was determined using the assumption that 4c has the same GLC response factor as 3c.

The products 3d, 4d and 5d were also obtained as a mixture, and the mixture indicated the following spectra: υ_max (neat) /cm^-1 2920, 2860, 1728, 1440, 1368, 1242, 1173, 1147, 1112, 1040 and 1025; δ_H (CDCl₃) 4.75-4.05 (m, OCH of 4d), 4.05 (s, CH₂OCOMe of 5d), 2.45-2.75 (m, SCH of 3d and 5d), 2.16 (3H, s, SMe), 2.00 (3H, s, OCOMe), 1.56 (s, Me) and 1.05-1.65 (m, CH of 5d and CH₂). The GLC-MS for these products exhibited the following separate mass spectra: m/z for 3d 202 (M⁺), 142 (M⁺−AcOH), 113, 95 [M⁺−(AcOH+SMe)], 94, 93, 79, 77, 67, 55 and 53; m/z for 4d 202(M⁺), 142, 113, 95, 94, 93, 79, 77, 67, 55 and 53; m/z for 5d 202 (M⁺), 163, 142, 129 (M⁺−CH₂OAc), 95, 94, 81, 79, 77, 67, 61, 55 and 53. The comparatively strong peak for m/z 129 was observed for 5d but for neither 3d nor 4d, and the ¹H NMR spectra for the mixture of 3d, 4d and 5d showed a peak at 4.05 ppm (singlet, CH₂OCOMe). These results lend strong support to the structure of 5d. Considering 2a to be trans-compound, 3d and 4d should also have trans-configuration. The low retention time of 3d compared to 4d and the high relative-intensity for m/z 142 of 3d relative to 4d support the proposed structures for 3d and 4d. The configuration for 5d is not unambiguous, but it seems reasonable to assume it to be in a trans-form in which the SMe and CH₂OAc groups are located at the stable equatorial positions. The yields of 3d, 4d and 5d were also determined with the assumption that these compounds have identical GLC response factors.

Effects of solvents and concentration of AcOH on the reaction of dimethyl disulfide (1) with hex-1-ene (2b) or 2-methylpent-1-ene (2c) 

The reaction of dimethyl disulfide (1) with hex-1-ene (2b) or 2-methylpent-1-ene (2c) was performed similarly using various solvents under the conditions indicated in Table 3. Analysis and characterization of the products were carried out as described above. The results are summarized in Table 3.

Effects of oxidants in the reaction of dimethyl disulfide (1) with hex-1-ene (2b) 

The reaction of dimethyl disulfide (1) with hex-1-ene (2b) was also conducted with a variety of oxidants under the conditions shown in Table 4. The yields of 3b and 4b were determined by the same procedure as that described above, and the results are summarized in Table 4.