The reaction of dimethyldioxirane with 1,3-cyclohexadiene and 1,3-cyclooctadiene: monoepoxidation kinetics and computational modeling

Introduction

Studies have shown dioxiranes to be powerful oxidants [1–4]. Dimethyldioxirane (1), in particular, has been useful in carrying out many synthetic transformations, adding an oxygen to a variety of substrates [1–4]. One of the most studied reactions of 1 is the epoxidation of olefins. Dimethyldioxirane epoxidations have been shown to be sensitive toward steric and substituent effects [1–4]. Kinetic studies have shown the relative reactivities for cis vs. trans epoxidation by 1 of simple alkyl substituted alkenes to be approximately 8:1 or greater [5]. The reverse reactivity trend for the epoxidation of phenyl-substituted alkenes was noted as the major exception [6]. However, all epoxidations by 1 have been shown to be quantitative and stereospecific. Mechanistically, the epoxidations have been postulated to occur via a spiro-transition state [1–7]. Semiempirical (AM1) calculations have been shown to be useful in modeling dimethyldioxirane epoxidation of simple alkenes within a structurally homologous series [8]. Density functional theory (DFT) calculations were used to model the epoxidation of ethylene, propene, and cis- and trans-2-butene by both the parent unsubstituted dioxirane and 1 and predicted a spiro-transition state in agreement with the mechanism postulated based on experimental data [9–12]. Ab initio calculations, employed to model cis and trans epoxidations by 1 also predicted a spiro-transition state and the correct relative reactivity of cis vs. trans simple alkenes [13]. Semiempirical (AM1) calculations failed to predict the relative reactivities of the epoxidation of alkenes by 1 within a series in which there were only minor reactivity differences as well as across two series for which there are small reactivity differences [14]. Basic gas phase DFT calculations proved to be effective for modeling the epoxidation of α,β-unsaturated esters by 1 as well as for the epoxidation of several acyclic conjugated dienes (s-trans systems) by dioxirane, systems for which steric effects are minimized [15, 16]. No studies have been reported for the epoxidation of cyclic conjugated s-cis dienes. We report here the results of a kinetic study with computational modeling of the epoxidation of 1,3-cyclohexadiene (2a) and 1,3-cycloctadiene (2b) by dimethyldioxirane (1).

Results and discussion

The reaction of 1 with excess 1,3-cyclohexadiene (2a) and 1,3-cyclooctadiene (2b) were carried out in dried acetone (Scheme 1). Under these conditions, the monoepoxide was the major product in either case. The monoepoxides were isolated in low yield and characterized by ¹H NMR and GC-MS data. The spectroscopic data were in agreement with the literature values [17].

Scheme 1

Kinetic studies with 2a and 2b were carried out using UV methodology in dried acetone at 23°C. The epoxidations by 1 were shown to be of the first order in dioxirane and in diene; second-order overall. 1,3-Cyclohexadiene was found to be more reactive toward monoepoxidation than 1,3-cyclooctadiene.

Pseudo-first-order conditions with a 10:1 diene/dioxirane ratio were employed to determine the k₂ value for the monoepoxidation of 1,3-cyclooctadiene. Owing to the greater reactivity, second-order conditions were employed to determine the k₂ value for the monoepoxidation of 1,3-cyclohexadiene. Representative plots of the kinetic data are shown in Figures 1 and 2. The k₂ values for the monoepoxidation of 1,3-cyclohexadiene and 1,3-cyclooctadiene are 1.14±0.06 and 0.31±0.03 m^-1 s^-1, respectively. The k₂ value for the monoepoxidation of 1,3-cyclohexadiene is comparable to those for the monoepoxidation of disubstituted acyclic conjugated dienes (1.29–1.42 m^-1 s^-1) by 1 [14]. The k₂ value for the monoepoxidation of 1,3-cyclooctadiene is comparable to that for the epoxidation of simple cis-alkenes by 1 [8].

Figure 1

Representative pseudo-first-order plot for the monoepoxidation of 1,3-cyclooctadiene (2b) by 1 in acetone at 23°C.

Figure 2

Representative second-order plot for the monoepoxidation of 1,3-cyclohexadiene (2a) by 1 in acetone at 23°C.

The monoepoxidations of the cyclic dienes were modeled by applying a basic DFT approach using the B3LYP/6-31G basis set. Regardless of the initial orientation of the reactants, minimized transition-state calculations yielded a spiro-transition state with a slight asynchrony tilt for the monoepoxidation of 1,3-cyclohexadiene and 1,3-cyclooctadiene as shown in Figure 3. The calculated transition-state geometries for these cases are consistent with previous results for the epoxidation of simple alkenes by 1, which also show similar asymmetry and tilts for the spiro-transition states [9–14].

Figure 3

Calculated density functional B3LYP 6-31G* transition-state structures for the monoepoxidation of 1,3-cyclohexadiene (2a) and 1,3-cyclooctadiene (2b) by 1.

The basic DFT approach was employed to quickly predict relative reactivities. These calculations were carried out in the gas phase because solvent corrections have been shown to lower the individual values but not affect the difference between values [18]. The predicted activation energies were determined by taking the difference between the calculated energies of the transition state and those of the ground states of the reactants while controlling for conformational effects. The predicted activation energies for 1,3-cyclohexadiene and 1,3-cyclooctadiene were found to be 13.37 and 14.14 kcal, respectively. The k_rel predicted values were calculated from the electronic activation energy difference for the two cyclic dienes. The ground-state geometry calculations of 1,3-cyclohexadiene and 1,3-cyclooctadiene show that 1,3-cyclohexadiene resembles a relatively flat s-cis system. The ground-state structure of the 1,3-cyclooctadiene shows the two double bonds to have a torsion angle of 51°, reducing conjugation. The predicted relative rates for 2a/2b of 3.9: 1 are consistent with the experimental kinetic data for 2a/2b of 3.7:1. The kinetic and predicted data are shown in Table 1. The HOMO-LUMO energies were taken from the DFT B3LYP 6-31G optimized geometries of 1,3-cyclohexadiene, 1,3-cyclooctadiene, and dimethyldioxirane. The HOMO energies of 1,3-cyclohexadiene and 1,3-cyclooctadiene are -5.59 and -6.05 eV, respectively. The LUMO energy of dimethyldioxirane is -0.47 eV. These data are consistent with the DFT results predicting that the 1,3-cyclohexadiene would be more reactive than the 1,3-cyclooctadiene.

On the basis of the successful modeling of the cyclic dienes, the relative rate constants for cis-/trans-phenyl-substituted alkenes 4a,b were evaluated using this quick, gas-phase calculation method. The transition-state geometries of the epoxidation of cis-1-phenylpropene and trans-1-phenylpropene by 1 were determined (Figure 4). The second-order rate of the epoxidation of cis-1-phenylpropene and trans-1-phenylpropene by 1 (Scheme 2) have been previously reported as 0.18±0.01 and 0.29±0.02, respectively, and noted as an exception of the overall trend of the reactivity of epoxidation by 1 of cis vs. the corresponding trans-alkene [6]. As expected, the calculation predicted that the epoxidation occurs via a concerted, electrophilic process with a spiro-transition state including a slight asynchrony tilt in the oxygen insertion bond formation.

Figure 4

Calculated density functional B3LYP 6-31G* transition-state structures for the epoxidation of cis-1-phenylpropene (4a) and trans-1-phenylpropene (4b) by 1.

Scheme 2

The electronic activation energies for the epoxidation of cis-1-phenylpropene and trans-1-phenylpropene were determined to be 14.00 and 13.80 kcal, respectively. The calculated relative reactivity was calculated to be 1.2:1, trans-1-phenylpropene/cis-1-phenylpropene, whereas the experimental relative reactivity was 1.6:1, thus successfully predicting the greater reactivity of the trans isomer. Table 2 summarizes the electronic and experimental data. The DFT results show that the exception in relative reactivity for 4a and 4b is actually consistent with the steric effect argument, but in this case, the trans isomer has the least steric interactions leading to the apparent reversal.

Conclusions

1,3-Cyclohexadiene (2a) was found to be more reactive than 1,3-cyclooctadiene toward monoepoxidation by 1. Kinetic studies showed that the overall reactivity for the monoepoxidation of the essentially s-cis 1,3-cyclohexadiene is comparable to that of the s-trans acyclic conjugated dienes with similar number of substituents. The results of the study of the less reactive substrate, 1,3-cyclooctadiene (2b), are similar to the second-order rate constants of simple cis-alkenes. It appears that the two double bonds behave as if they are isolated alkenes for this case. Basic DFT calculations (gas phase) are consistent with the concerted (spiro) mechanism for the monoepoxidation of the dienes and predicted the relative reactivity with reasonable agreement with experimental data. Application of this approach to epoxidation of cis-/trans-1-phenylpropene successfully predicted the previously noted unexpected reactivity for this pair of alkenes. The method appears to be able to detect the effect of subtle steric interactions in these closely related systems. The basic DFT method is a reliable quick approach to model dioxirane epoxidation reactions and for predicting relative reactivities as long as the reactants have reasonable structural similarities and conformational factors are controlled without the need for time consuming solvent corrections. The challenge of predicting the relative reactivities for substrates with vast structural differences toward epoxidation by 1 still remains to be addressed.

Experimental

Dimethyldioxirane was prepared by the general procedure developed by Murray, for the reaction of oxone and acetone, with modifications [5, 7]. Stock solutions (≤0.1 m in acetone) were collected by redistillation of the initial azeotrope [1–4]. The dimethyldioxirane concentration in acetone was determined using UV methodology (λ 330 nm and molar absorptivity of 12.9 cm^-1m^-1) [6]. Dimethyldioxirane stock solutions were stored at -20°C over anhydrous Na₂SO₄ until use. Stock solutions were stable for months under these conditions. 1,3-Cyclohexadiene and 1,3-cyclooctadiene were commercially available and were distilled before use. Stock solutions of the dienes in dried acetone were prepared under anhydrous Na₂SO₄ at room temperature and used the same day. For kinetic runs, the temperature of the jacketed sample cell (23.0±0.3°C) was monitored via an YSI probe. HPLC-grade acetone from Aldrich was stored under Na₂SO₄ until use for all kinetic runs. Care needs to be taken to avoid advantageous water since the k₂ values will increase substantially depending on the amount of water present [6].