Kinetic Support for the Generation of a Phenylsulfenium Ion Intermediate

In the reaction of N-methylbenzenesulfenamide (1) with thioanisole (4) in the presence of trifluoroacetic acid (TFA), the initial rate υ0 of total appearance of 2- and 4-(methylthio)phenyl phenyl sulfides (5) is first and zero order with respect to the initial concentrations of 1 and 4, respectively: υ0=kobs[1]0. The pseudo-first order rate constant kobs is evaluated as 5.2×10-4 sec-1 with varying concentrations of 4 (0.72 to 5.0 M) in a mixture of 4 and CH2Cl2 in the presence of TFA at 10 °C. This data supports the notion that a phenylsulfenium ion intermediate 3 interacting with both the counter ion and the unshared electron-pair of amine is generated by heterolytic N-S scission of the protonated sulfenamide 2, leading to the observed formation of 5.


Introduction
Many researchers have made proposals concerning the generation of sulfenium ion intermediates [1]; this intermediate can only be isolated in the case of coordination with two nitrogen atoms [2], yet several studies [3] have failed to find evidence supporting its existence, as it has too short a lifetime to exist as the free ion, being converted to sulfonium [4] or thiiranium ions [5] by reaction with sulfides or alkenes, respectively. Thus, the existence of the sulfenium ion has been the subject of controversy for long time. We have proposed that a sulfenium ion with a very short lifetime can exist by interacting with both a counter ion and the unshared electron pair of an amine (i.e. not as a free ion), considering the effects of the counter ion, the amine, solvent nucleophilicity, etc. on the reactions of N-alkylarenesulfenamides or azides with alkyl aryl sulfides in the presence of trifluoroacetic acid [6].
However, the existence of such an unstable sulfenium ion is still controversial because the product resulting from thiolation of benzene by the sulfenium ion is not detected, the thiolation on toluene rarely occurs [6], and so on. In this paper, we present the first kinetic support for the existence of the ion in the thiolation of a good nucleophilic aromatic compound like thioanisole, and our proposed mechanism [6] for the generation of the sulfenium ion is thus made clear. This work also arouses interest for organic synthesis in view of the observed one-step aromatic thiolation by sulfenium ion.

Results and Discussion
Kinetic study on the reaction of N-methylbenzenesulfenamide (1) with thioanisole (4) in the presence of trifluoroacetic acid (TFA).
The reactions of N-methylbenzenesulfenamide (1) with thioanisole (4) are carried out using both TFA and CH 2 Cl 2 at 10 °C, and give 2-and 4-(methylthio)phenyl phenyl sulfides (5) (the observed para/ortho isomer ratio being ca. 10). The reactions producing 5 are quenched by pouring the reaction mixtures into cold triethylamine (see Experimental section). We have already proposed that mechanistically the reactions of N-alkylbenzenesulfenamides with 4 in the presence of TFA give 2and 4-(alkylthio)phenyl phenyl sulfides by phenylthiolation of a phenylsulfenium ion intermediate [6]. In order to establish kinetically the generation of the sulfenium ion, the reaction-orders for the appearance of 5 with respect to 1 and 4 are evaluated by using the differential method as shown below. As for the evaluation, the initial rate technique dealing with the initial rates υ 0 for appearance of 5 can avoid possible complications due to the products, and leads to the true order with respect to concentration.
The initial rate υ 0 =d [5]/dt at t=0 sec for the formation of 5 can be represented by the equation . The initial rates υ 0 (see Table 1) at such an excess constant concentration of 4 (i.e. 4.3 M) were evaluated for individual initial concentrations of 1. The details of the determination of υ 0 are described in Experimental section. The correlation between the logarithms of υ 0 and of [1] 0 gives a linear plot with a slope of 1.09 (Figure 1). The result shows that the production of 5 obeys first order kinetics for 1. The details for determining the initial rates υ 0 (see Table 2) at the individual initial concentrations of 4 are also shown in the Experimental section. The plot of log (υ 0 /[1] 0 ) against log [4] 0 leads to the slope -0.001 representing the reaction-order of 4 ( Figure 2). Thus, the production of 5 obeys zero order kinetics for 4.
With the variation of 4 from 0.72 M to 4.97 M, the amount of CH 2 Cl 2 was changed from 58%v/v to 8.3%v/v in the presence of a constant quantity (33%v/v) of TFA. As for the effects on the aromatic thiolation of the sulfenium ion, CH 2 Cl 2 is less nucleophilic solvent [6]. The dilution with CH 2 Cl 2 in the reactions of sulfenium ions [6] or nitrenium ions [7] with aromatics scarcely affects the yields when TFA is present in at least 30% v/v. The fact that the almost same k obs is obtained from Figures 1 and 2 using different amounts of CH 2 Cl 2 as described below indicates that the dilution with CH 2 Cl 2 in this reaction system is appropriate.
Equation 1 can be derived by assuming a steady generation of a phenylsulfenium ion 3 as indicated in Scheme 1; k 1 , k -1 and k 2 are the rate constants indicated in Scheme 1.
The concentration of 3 is shown in Equation 2, and the rate for appearance of 5 is shown in Equation 3 .
When k 2 [4] is sufficiently large as compared to k -1 , then Equation 4 can be obtained. As shown in Scheme 1, the sulfenium ion interacts with both counter ion and the unshared electron-pair of the amine, and thus 1 may undergo a protonation, not by free protons, but rather by a TFA molecule to produce 2 [6].
The intercept on the log υ 0 axis in Figure 1 corresponds to log k obs being ca. -3.2 because we can represent as υ 0 =k obs [1] 0 (i.e. log υ 0 =logk obs + log[1] 0 ). As seen in Figure 2 is sufficiently larger than 1.0. Thus, k 1 must be lower than 5.2×10 -4 sec -1 . The low rate constant k 1 well suggests that the reaction yielding 5 proceeds via a very reactive intermediate such as a phenylsulfenium ion. The above-mentioned assumption that k 2 [4] is much larger than k -1 becomes true when the highly reactive intermediate 3 is generated.
We considered the possibility that the thiolation giving 5 occurs via not the sulfenium ion 3 but the protonated sulfenamide 2 as the reactive intermediate. If it were true, the value k obs (equal to 5.2×10 -4 sec -1 ) would be identical with k [TFA] by assuming the steady state for 2 in the similar way to that shown above; k means the rate constant of the formation of 2 from 1 and TFA. But, the rate constant k should not be such a low value as k obs /[TFA] because the proton-transfer [8] to amine occurs at the diffusion-controlled rate. Thus, the above possibility can be ruled out. Actually, the formation of 5 by direct reaction of 2 with 4 has been excluded [6].
In a previous paper [6], we have ruled out the possibility that the phenylthiolation giving 5 occurs via a phenylthiyl radical generated by homolytic N-S scission of the protonated sulfenamide 2. Our earlier study [6] also has shown that the reactions of N-alkylarenesulfenamides with 4 are influenced by the counter ion, the amine, the aryl substituent of the sulfenamide and the solvent nucleophilicity. This excludes the possibility of a neutral intermediate such as PhSOCOCF 3 .
Therefore, the evidence favors the fact that protonated sulfenamide 2 undergoes a spontaneous N-S scission (not induced by 4) to generate the very reactive sulfenium ion 3, leading to the production of 5 (Scheme 1). The sulfenium ion does not exist as a free ion, but can be generated by interacting with not only the counter ion but also the unshared electron-pair of the amine (Scheme 1); the interacting amine is not in the solvent, but formed just by heterolytic S-N scission of the protonated form 2. We have already proposed the similar interaction of nitrenium ions with both the counter ion and unshared electron-pair [9].
From the competition between the phenylthiolation giving 5 and the formation of diphenyl disulfide (6), we have proposed that the protonated sulfenamide 2 undergoes a homolytic S-N scission forming a phenylthiyl radical, producing 6 by its dimerization (Scheme 1) [6]. This type of dimerization is well known [10]. We may consider that the use of the other trapping agents instead of thioanisole 4 provides the different k obs from that in the use of 4 because k 1 depends upon k 2 . In fact, the trapping reaction with anisole or diphenyl sulfide proceeds more rapidly than that with 4, and we could not observe the thiolation product to benzene [11].

Kinetic study on generation of phenylsulfenium ion 3
In the reactions of sulfenamide 1 with thioanisole (4), the reaction-mixture showed the GLC peaks corresponding to those of 5 (the ortho: para isomer ratio for 5 was ca. 1:10) and diphenyl disulfide (6).
Since the total yield of 5 and 6 is less than 100%, some unidentified products must be formed: For example, the reaction of 1 (4.0 mmol) with 4 (12.0 mL) using TFA (8.0 mL) and CH 2 Cl 2 (4.0 mL) at 10 °C for 24 h produced 5 (45% yield) and 6 (47% yield). In spite of the formation of other products aside from 5, the plot of [5] versus reaction time t provides the initial slope υ 0 with satisfactory R 2 during 0-750 sec by using the equation [5]= (-α×10 -β t 2 +γt) (see Tables 1 and 2 below), leading to the good correlation as seen in Figures 1 and 2. This means that the rates of appearance of 5 are not complicated by the presence of the products. At the higher temperature, 5 is formed in higher yields: The reaction of 1 (4.0 mmol) with 4 (3.0 mL) in the presence of TFA (6.0 mL) and CH 2 Cl 2 (1.0 mL) at reflux temperature for 5 h gave 5 (70% yield) and 6 (23% yield). But, in this case, the variation of [5] with time could not be obtained accurately because of the fast reaction at high temperature.
Reaction-order with respect to N-methylbenzenesulfenamide 1.
The reactions of solutions containing varying amounts of 1 (shown below) with 4 (12 mL) were carried out at 10 °C in the presence of both TFA (8.0 mL) and CH 2 Cl 2 (4.0 mL); the reactions were started by addition of 1 dissolved in a small amount of CH 2 Cl 2 to the mixture of 4, TFA and CH 2 Cl 2 . The work-up (i.e. the treatment for stopping the reaction) was involved pouring an aliquot (1.0 mL) of the reaction mixture into Et 3 N (2.0 mL, an excess with respect to the TFA) cooled at −20°C. After the workup, the amount of 5 was determined by GLC analysis using GLC-MS. The reaction-order of 1 for total appearance of 5 is obtained as follows: at the individual initial concentration [1] 0 (83, 170, 250 or 330 mM), the concentrations [5] at times 150, 300, 450 and 750 sec were determined (Table 1). The concentration [5] was plotted against reaction time t, and the initial slope corresponding to the initial rate (υ 0 =d [5]/dt at t=0) for appearance of 5 was evaluated as the γ value by a statistical optimal-method; the optimal values α, β and γ (α or β is a positive integer, and γ is a positive number) as shown in Table 1 are derived by using the equation [5]=(-α×10 -β t 2 +γ t) for the each plot above. R 2 values (i.e. the square of the correlation coefficient) in this derivation were satisfactory (Table 1). Then, the plot of log υ 0 against log [1] 0 gives a straight line with the slope 1.09 (Figure 1).

Reaction-order with respect to thioanisole(4).
Compound 1 (4.0 mmol) was dissolved at 10 °C in a mixture of 4 (16.0 mL) and CH 2 Cl 2 in the presence of TFA (8.0 mL). The work-up of the reaction mixture and the analysis of the products were conducted as indicated above. At individual initial concentrations [4] 0 ranging from 0.72 M to 5.0 M, [5] was plotted against the reaction time (t=150, 300, 450, 600 or 750 sec). The concentration [5] at each time point is shown in Table 2. As described above, the initial rate (υ 0 =d [5]/dt at t=0) was evaluated as the initial slope (being equal to γ) by employing the equation [5]=(-α×10 -β t 2 +γ t). The values υ 0 , α, β and R 2 are shown in Table 2