Experimental and theoretical study on the reactivity of the R-CN/H2O2 system in the epoxidation of unfunctionalized olefins

Article

Experimental and Theoretical Study on the Reactivity of the R-CN/H₂O₂ System in the Epoxidation of Unfunctionalized Olefins

Maria Luiza A. von Holleben*, Paolo R. Livotto, Cristina M. Schuch^#

Instituto de Química, Universidade Federal do Rio Grande do Sul, Av. Bento Gonçalves, 9500, Campus do Vale, 91501-970, Porto Alegre - RS, Brazil

A reatividade da acetonitrila, tricloroacetonitrila, benzonitrila, m-clorobenzonitrila, 3-cianopiridina, 1-naftonitrila e 9-antracenonitrila na geração in situ do respectivo ácido peroxicarboximídico foi estudada por cálculos de OM, utilizando o método AM1, e experimentalmente na epoxidação do cicloexeno e R-(+)-limoneno. Os resultados experimentais mostraram que a reatividade do sistema Cl₃CCN/H₂O₂ foi semelhante ao MCPBA na epoxidação de olefinas não-funcionalizadas, quando o solvente utilizado foi uma mistura bifásica CH₂Cl₂/H₂O.

A study on the reactivity of peroxycarboximidic acids in situ generated from acetonitrile, trichloroacetonitrile, benzonitrile, m-chlorobenzonitrile, 3-cyanopyridine, 1-naphthonitrile and 9-anthracenenitrile was performed by semiempirical AM1 method and experimentally in the epoxidation of cyclohexene and R-(+)-limonene. Experimental results showed that the reactivity of Cl₃CCN/H₂O₂ was quite similar to MCPBA in the epoxidation of unfunctionalized olefins when a biphasic mixture CH₂Cl₂/H₂O was employed as solvent.

Keywords: epoxidation, alkenes, peroxycarboximidic acid

Introduction

Many reagents have been utilized to epoxidize unfunctionalized olefins, among them peracids are the most widely used. The structurally related peroxycarboximidic acids¹ have been much less explored in spite of non-acidic conditions and good yields obtained with peroxybenzimidic acid (Payne epoxidation)^2a-d. The acid (I) is a highly reactive species which has not yet been isolated. It reacts with an alkene leading to epoxide and amide (Scheme 1). In the absence of the alkene, it reacts with a second mole of hydrogen peroxide affording oxygen, water and an amide.

The activation of the nitrile by hydrogen peroxide³ to form a peroxycarboximidic acid (I) occurs in methanol as solvent and basic medium. An interesting variation of Payne epoxidation was introduced by Bach and coworkers⁴. These authors utilized trichloroacetonitrile instead of benzonitrile and a biphasic mixture of dichloromethane/water (1:1) as solvent in the epoxidation of some olefins. This work was preceded by a theoretical study of the structural and electronic features of peroxyformic and peroxyformimidic acids using ab initio level calculations⁵. The results suggested that the reactivity of both species would be quite similar and that the presence of an electron withdrawing substituent should enhance the reactivity of the peroxycarboximidic acid on the epoxidation. To our knowledge, no other study has been put forward to investigate the effect of the nitrile substituent in the epoxidation of unfunctionalized olefins.

In this context, we studied the theoretical and experimental reactivity of the acetonitrile (1a), trichloroacetonitrile (1b), benzonitrile (1c), 3-chlorobenzonitrile (1d), 3-cyanopyridine (1e), 1-naphthonitrile (1f) and 9-anthracenenitrile (1g) in the epoxidation of cyclohexene (2a) and R-(+)-limonene (2b). Our goal was to develop a reagent based on hydrogen peroxide that could compete with MCPBA in cost and reactivity.

In the experimental study we employed some different solvents as methanol, a biphasic mixture of dichloromethane/water (1:1) and water as depicted in the Scheme 2.

A theoretical study was carried out at the AM1^6a level implemented by MOPAC program^6b in order to investigate the conformational features and intrinsic properties of each peroxycarboximidic acid.

Experimental

General

Unless otherwise noted, materials were obtained from commercial suppliers and used without further purification. R-(+)-limonene was distilled prior to use. Flash column chromatography was carried out using 230-400 mesh silica gel. ¹H NMR spectra were obtained with a Varian VXR-200 spectrometer. Proton-decoupled ¹³C-spectra were obtained at 50MHz with the same instrument. GC analyses were carried out using a capilar column HP-1 (50m x 0.2mm x 0.11 mm). Theoretical calculations were performed at Cray Y-MP2E/232 computer employing the AM1 method implemented by MOPAC 7.0 program. The energy minimization was realized with respect to all geometrical parameters from an appropriate region of the Potential Energy Surface (PES).

Typical procedure for the epoxidation of 2a or 2b with nitriles (1a, 1c-g)/H₂O₂:

In a typical run, olefin (2a or 2b, 10 mmol), nitrile (1a, 1c-g, 10 mmol) and KHCO₃ (2 mmol) were dissolved in methanol (20 cm³). The mixture was treated with hydrogen peroxide (30% wt. solution in water, 10 mmol) dropwise. The stirring was continued for 24 hours at room temperature. The mixture was diluted with water (5 cm³) and extracted three times with dichloromethane (10 cm³). The organic layer was washed with 20% aq. solution of NaHSO₃ and dried over MgSO₄. The yield of epoxide was determined by GC analysis based on internal standard (n-decane). The respective amide was isolated by filtration, washed with hexane and dried in vacuo. Purification by column chromatography on silica gel (10% EtOAc/hexane, v/v) afforded epoxides 3a or 3b-c (yields are indicated in Table 1). Compounds 3a, 3b and 3c were isolated as chromatographically pure materials and exhibited acceptable ¹H NMR, ¹³C NMR, IR and MS spectral data, which were identical with those reported in the literature⁷.

Typical procedure for the epoxidation of 2a or 2b with Cl₃CCN/H₂O₂:

To a solution of olefin (2a or 2b, 10 mmol) and Cl₃CCN (1b, 10 mmol) in CH₂Cl₂ (10 cm³) was added 10 cm³ of 0.2 mol. dm^-3 aqueous solution of KHCO_3. The biphasic mixture was cooled to 0^oC and treated with hydrogen peroxide (30% wt. solution in water, 10 mmol) dropwise under vigorous stirring. The mixture was vigorously stirred for 24 hours at room temperature. The organic layer was separated, the aqueous layer was three times extracted with dichloromethane (10 cm³), and dried over MgSO₄. The solvent was removed under vacuum and trichloro-acetamide was isolated by filtration. Purification by column chromatography on silica gel (10% EtOAc/hexane, v/v) afforded epoxides 3a or 3b-c in 76% and 72% yield, respectively. In the aqueous condition, 10 cm³ of dichloromethane was replaced by 10 cm³ of water.

Results and Discussion

The epoxidation of cyclohexene (2a) and R-(+)-limonene (2b) with nitriles (1a-g), mediated by H₂O₂ and KHCO₃ was carried out. The results described in Table 1 were obtained after 24 hours at room temperature (standard procedure).

Epoxides were obtained in 10-65% yields with nitriles 1a, 1c-f (entries 1-2, 6-13, Table 1) using methanol as solvent. Trichloroacetonitrile (1b) did not furnish epoxides or amide in methanol. In this case, methyl trichloroaceti-midate was isolated as result of the nucleophilic attack of the solvent at the nitrile carbon. The trichloroacetimidic acid was formed only when a biphasic mixture of dichloromethane/water (1:1) (entries 3 and 4) or water (entry 5) were used as solvent. Interestingly, nitriles 1a, 1c-g did not furnish epoxides or amides using the biphasic mixture of solvent. An amide/epoxide ratio superior to 1.0 was observed in some cases (entries 5, 8-10, and 12-13) and suggested that the peroxycarboximidic acid reacted with hydrogen peroxide in the fast step⁸. It is important to consider that the reduction of reactive intermediate can be promoted by the olefin or the hydrogen peroxide, as represented in the Scheme 1.

The epoxidizing system Cl₃CCN/H₂O₂ showed high chemoselectivity in the epoxidation of the more nucleophilic double bond of R-(+)-limonene (entries 4-5) in spite of the unselective behavior of the other systems RCN/H₂O₂ studied⁹. A 13:1 ratio of 3b:3c (Scheme 2) was obtained in CH₂Cl₂/H₂O (entry 4) and 9:1 in water (entry 5), which were quite similar with results obtained in the epoxidation of R-(+)-limonene with MCPBA¹⁰. Besides, when the reaction was carried out in water (entry 5) the product 3b' (trans) was obtained in 82% d.e. (ratio trans:cis of 10:1, determined by ¹H NMR). In this case the diastereoselectivity was attributed to a deleterious effect of the aqueous 0.1mol. dm^-3 solution of KHCO₃ on the cis stereoisomer which resulted an enriched mixture of trans isomer in only 12% yield. Jones and coworkers¹¹ verified the same effect when an aqueous 1mol. dm^-3 solution of NaHSO₃ was employed. They noted that the oxirane ring opening in the cis isomer was faster than in the trans isomer.

The experimental reactivity order observed in the olefin epoxidation with nitriles 1a, c-g using a standard procedure and methanol as solvent, was 1c > 1a > 1d > 1e > 1f > 1g. This experimental behavior is not in accordance with LUMO energies of nitriles 1a-g calculated by AM1 method¹². Besides, it is important to consider that Cl₃CCN (1b) was not included in the above order because good results only were attained when methanol was replaced by CH₂Cl₂/H₂O.

In order to attempt rationalize these results, a conformational analysis and intrinsic properties calculations of the conformers of each peroxycarboximidic acids (5a-g, Figure 1) were performed¹³. The two most stable conformers of the eight possible⁵ were Z (C=N geometry), s-cis (C-O bond), antiperiplanar and E (C=N geometry), s-cis (C-O bond), synperiplanar as represented in Figure 1.

We calculated the relative energies and dipole moment values of the corresponding E and Z conformers of the peroxycarboximidic acids 5a-g. The results are summarized in Table 2.

The calculations showed that Z is the conformer of minimum energy with an internal hydrogen bond between H imidic and hydroxyl group. The E conformer is an analog of the peracids with an anti imine structure containing an internal hydrogen bond between nitrogen atom and H hydroperoxide. Both conformers showed a minimum conformation when H hydroperoxide was out of plane formed by the functional group (Figure 1).

The butterfly mechanism is usually accept for Payne epoxidation¹⁴. In this context, it was postulated that only the E conformer has an adequate geometry to react with an olefin⁵. However, we assumed that Z conformer could be able to react with hydrogen peroxide in the fast step.

We observed an unusual selectivity of trichloroaceti-midic acid in the epoxidation of the more substituted double bond of R-(+)-limonene, which was comparable with that observed by peracids. Since the formation of the peroxycarboximidc acid is usually accepted as the rate-determining step we believed that, in this particular case, the strong electron-withdrawing effect of the trichloromethyl group could be affecting the kinetics of the reaction. Unfortunately, it was impossible to compare the reactivity of all RCN/H₂O₂ systems studied in the same solvent. Further additional kinetics studies would be needed to understand the different reactivity showed by these systems.

Conclusion

In conclusion, the choice of a nitrile for olefin epoxidation mediated by hydrogen peroxide could not be carried out through a simple analogy between the peroxycarboximidic acid, in situ generated, and the structurally related peracid.

Moreover, we concluded that the reactivity of benzonitrile, 3-chlorobenzonitrile and acetonitrile were quite similar in the epoxidation of unfunctionalized olefins, mediated by hydrogen peroxide and KHCO₃ in methanol as solvent. Trichloroacetonitrile showed the best results in this study. In the epoxidation of R-(+)-limonene, the Cl₃CCN/H₂O₂ system presented a chemoselective behavior similar to MCPBA. Besides, the utilization of a biphasic mixture of solvent makes the purification of oxiranes easier and avoids acidic conditions, which makes the Cl₃CCN/H₂O₂ system an excellent alternative method to MCPBA epoxidation.

Acknowledgments

The authors thank FAPERGS and CNPQ for financial support and studentship and CESUP/UFRGS (FINEP) for computer time.

References

1. Wiberg, K. B. J. Am. Chem. Soc. 1953, 75, 3961.

2. (a) Payne, G. B.; Williams, P. H. J. Org. Chem. 1961, 26, 651. (b) Payne, G. B.; Deming, P.H.; Williams, P. H. J. Org. Chem. 1961, 26, 659. (c) Payne, G. B. Tetrahedron 1962, 18, 763; (d) Bachmann,C.; Gesson, J. P.; Renoux, B.; Tranoy, I. Tetrahedron Lett. 1998, 39, 379.

3. von Holleben, M. L. A.; Schuch, C. M. Quím. Nova, 1997, 20, 58.

4. Arias, L. A.; Adkins, S.; Nagel,C. J.; Bach, R. D. J. Org. Chem. 1983, 48, 888.

5. Lang, T. J.; Wolber, G. J.; Bach, R. D. J. Am. Chem. Soc. 1981, 103, 3275.

6. a) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. J. Am. Chem. Soc. 1985, 107, 3902. b) Stewart, J. J. P. MOPAC 7.0, Quantum Chemistry Program Exchange, Program 452.

7. Carman, R. M.; DeVoss, J. J.; Greenfield, K. L. Aust. J. Chem. 1986, 39, 441.

8. On the mechanism of the conversion of nitriles to amides by hydrogen peroxide in the absence of olefin, see: McIsaac, J. J. E.; Ball, R. E.; Behrman, E. J. J. Org. Chem. 1961, 36, 3048.

9. von Holleben, M.L.A.; Schuch, C.M.; Livotto, P.R. VII Brazilian Meeting on Organic Synthesis, Rio de Janeiro, Brazil, 1996, 135, PS109.

10. Carlson, R. G.; Behn, N. S.; Cowles, C. J. Org. Chem. 1971, 36, 3832.

11. Santos,A. G.; Castro,F. L.; Jones Jr, J. Synthetic Commun. 1996, 26, 2651.

12. We calculated the LUMO energies of nitriles 1a-g by AM1 method. However, we cannot correlate the theoretical reactivity order, based in the LUMO energies of nitriles with the experimental reactivity order observed. (a) von Holleben, M. L. A.; Schuch, C. M.; Livotto, P.R. 5^th European Symposium on Organic Reactivity (ESOR V), Santiago de Compostela, Spain, 1995, 95. (b) Schuch, C. M. Dissertação de Mestrado, Universidade Federal do Rio Grande do Sul, Brazil, 1996.

13. Geometrical Parameters for Peroxycarboximidic acids 5a-5g: Z Conformers: (O-O): 1.291 Å (5a, 5f); 1.290 Å (5b, 5c, 5d, 5e); 1.292 Å (5g). (O-H): 0.986 Å (5b, 5e, 5f); 0.985 Å (5a, 5c, 5d, 5g). (C-O): 1.434 Å (5a); 1.440 Å (5c); 1.438 Å (5d); 1.437 Å (5e, 5f, 5g); 1.433 Å (5b). (C=N): 1.279 Å (5a, 5f); 1.278 Å (5b); 1.282 Å (5c, 5e); 1.280 Å (5d, 5g). (N-H): 0.998 Å (5a); 1.000 Å (5b, 5e, 5g); 0.999 Å (5c, 5d, 5f). (C-R): 1.499 Å (5a); 1.525 Å (5b); 1.480 Å (5c); 1.479 Å (5d); 1.474 Å (5e); 1.482 Å (5f); 1.484 Å (5g). (O-C-R): 106.9^o (5a); 108.5^o (5b); 108.2^o (5c); 108.4^o (5d); 108.3^o (5e); 107.1^o (5f); 107.6^o (5g). (O-C=N): 127.2^o (5a); 127.3^o (5b); 126.6^o (5c); 126.5^o (5d); 126.8^o (5e); 127.0 (5f); 127.1^o (5g). (C=N-H) 116.1^o (5a, 5b); 115.8^o (5c, 5e, 5f, 5g). (O-O-C): 114.2^o (5a, 5c, 5f); 113.6^o (5b); 114.1^o (5d, 5g). (O-O-H): 107.2^o (5a, 5b, 5e, 5f); 107.3^o (5c); 107.4^o (5d); 107.1^o (5g). E Conformers: (O-O): 1.286 Å (5a, 5d, 5f, 5g); 1.284 Å (5b); 1.290 Å (5c); 1.285 Å (5e). (O-H): 0.985 Å (5a, 5f, 5g); 0.986 Å (5c, 5d); 0.987 Å (5b, 5e). (C-O): 1.420 Å (5a); 1.422 Å (5b, 5d, 5f, 5g); 1.430 Å (5c); 1.423 Å (5e). (C=N): 1.284 Å (5a, 5f, 5g); 1.277 Å (5b); 1.286 Å (5c); 1.285 Å (5d, 5e). (N-H): 0.996 Å (5a, 5e); 0.997 Å (5b, 5c, 5d, 5f, 5g). (C-R): 1.499 Å (5a); 1.530 Å (5b); 1.480 Å (5c, 5d); 1.475 Å (5e); 1.483 Å (5f); 1.485 Å (5g). (O-C-R): 107.4^o (5a); 108.3^o (5b); 108.8^o (5c, 5e); 108.5^o (5d); 108.1^o (5g); 108.7^o (5g). (O-C=N): 121.4^o (5a,5g); 121.3^o (5b); 121.0^o (5c, 5e); 121.1^o (5d); 121.5^o (5f). (C=N-H) 115.2^o (5a); 117.6^o (5b); 115.3^o (5c, 5f); 115.4^o (5d, 5g); 115.5^o (5e). (O-O-C): 117.5^o (5a); 116.4^o (5b); 117.1^o (5c, 5f); 117.0^o (5d, 5g); 117.2^o (5e). (O-O-H): 108.9^o (5a, 5d, 5f, 5g); 108.8^o (5b); 109.0^o (5c); 109.1^o (5e).

14. In the olefin epoxidation, the electrophilic mechanism occurs in the cyclic transition state (Butterfly) or in the opened transition state. For discussion see: Rebek, J. Heterocycles 1981, 15, 517.

Received: September 30, 1999

Published on the web: October 13, 2000

Publication Dates

• Publication in this collection
  19 Apr 2001
• Date of issue
  2001

History

• Received
  30 Sept 1999