Niobium Pentachloride Catalysed Ring Opening of Epoxides

Abstract

Epoxide ring opening is a frequently required transformation in Organic Synthesis. In this paper we describe the application of NbCl₅ for this purpose using three different substrates. Chlorohydrins, 1,2-diols, products containing solvent residues as well as rearrangement products are obtained, depending on both the substrate structure and reaction conditions. Rationalizations to account for some of the results are suggested.

Introduction

Opening of epoxide rings has been a frequently required transformation in our studies on the synthesis of natural products [1]. Boron trifluoride etherate is the most commonly used Lewis acid for this purpose, but we decided to investigate the possibility of obtaining different products by using niobium pentachloride (NbCl₅), which is a stronger Lewis acid and is currently attaining increasing popularity as a reagent [2].

In preliminary experiments we found that the reaction of epoxides with NbCl₅ is very fast and follows different paths and mechanisms depending on the substrate structure and the reaction medium. We have now chosen three model substrates, compounds 1 – 3 (Figure 1), containing suitable structural variations, with the aim of comparing the results and, as much as possible, rationalize the observed structure-result correlations and extend them to further more detailed studies. Thus, compounds 1 and 3 have a fully substituted carbon in the epoxide ring that could lead to a carbocation formation; compound 1 has, additionally, an efficient migrating group. Compound 2 cannot form a tertiary carbocation.

Figure 1. Substrates chosen for treatment with NbCl₅.

Figure 1. Substrates chosen for treatment with NbCl₅.

Results and Discussion

We first treated each of the three substrates with NbCl₅ dissolved in anhydrous ethyl acetate, at room temperature under N₂ atmosphere. Ethyl acetate dissolves NbCl₅ completely, giving a yellow solution. When the substrate, dissolved in ethyl acetate, was added to this solution, an instantaneous reaction took place, giving a solution that was quenched after 1.0 min with a 10 % aqueous citric acid solution [3].

The results are summarized in Scheme 1. Isolated crude product (mixture) yields are given, together with an indication of the product taken as a base for the calculation, as well as the ratio, determined by GLC, between the products of each reaction. All products were isolated and characterized by ¹H-NMR, ¹³C-NMR, IR and mass spectra.

Scheme 1. Results obtained on treatment of epoxides with NbCl₅ using ethyl acetate as solvent.

Scheme 1. Results obtained on treatment of epoxides with NbCl₅ using ethyl acetate as solvent.

The formation of compounds 4 and 5 from substrate 1, as well as the formation of compound 8 from 3 clearly show a preferential breakage of the epoxide bond to the more substituted carbon atom. The resulting tertiary carbocation from compound 1 (either incipient or fully formed) has undergone acyl group migration to form 4, or H⁺ elimination to give 5 [4]. The cation from compound 3, having neither groups with high migratory aptitude nor a carbonyl group to favor the formation of a conjugated double bond through H⁺ elimination, followed a different path, namely addition of Cl^- which, considering the stereochemistry of the product, apparently came directly from the catalyst residue still bonded to the epoxide oxygen. At this stage we cannot suggest a mechanism to account for the formation of compound 9, which apparently was formed during the hydrolysis process. Its definite stereochemistry suggests the participation of the catalyst in the introduction of the OH group in the tertiary carbon atom.

Substrate 2 has no fully substituted carbon atom to favor the ring opening and thus reacted in a different way. The trans stereochemistry of both products suggests that the epoxide, activated by the Lewis acid, reacted through S_N2 mechanisms, the attack by Cl^- to form 6 or by the solvent to form 7 possibly playing an important role in the opening of the epoxide ring.

Changing the solvent for the reactions of compounds 2 and 3 resulted in meaningful differences (Scheme 2). Compound 2, with NbCl₅ dissolved in acetonitrile, gave only the chlorohydrin 6; the comparatively low yield (29 %) apparently being due to loss of product in the aqueous phase during extraction. More productive extraction methods are currently being investigated.

Scheme 2. Results obtained on treatment of epoxides with NbCl₅ using acetonitrile as solvent.

Scheme 2. Results obtained on treatment of epoxides with NbCl₅ using acetonitrile as solvent.

Compound 3, under the same conditions, also gave a lower yield of crude product, but the structures of products 8 and 10 are consistent with the interpretation of ring opening by breaking the bond to the more substituted carbon atom, followed by incorporation of a nucleophile from the same face of the oxygen atom and the catalyst, thus suggesting that the nucleophile came from the catalyst.

The relative stereochemistry of compounds 6, 7 and 9 was assigned based on their ¹H- and ¹³C-NMR spectra, and careful comparison with literature data [5]. For compounds 8 and 10, however, no literature data was available. Molecular mechanics studies [6] show that NOE experiments could lead to ambiguous results, but the coupling constants between the carbinolic hydrogen and the two neighboring hydrogens would be very different for the two possible isomers (cis or trans) of each compound (see Table 1).

Table 1. Calculated and experimental vicinal coupling constantsfor compounds 8 and 10.

Compound	Compound 8		Compound 10
Coupling constant	J1 (Hz)	J2 (Hz)	J1 (Hz)	J2 (Hz)
Calculated for trans isomer:	2.7	4.3	3.0	6.4
Calculated for cis isomer:	10.6	4.4	11.0	4.4
Experimental values:	9.5	4.3	10.7	4.8

Table 1. Calculated and experimental vicinal coupling constantsfor compounds 8 and 10.

Compound	Compound 8		Compound 10
Coupling constant	J1 (Hz)	J2 (Hz)	J1 (Hz)	J2 (Hz)
Calculated for trans isomer:	2.7	4.3	3.0	6.4
Calculated for cis isomer:	10.6	4.4	11.0	4.4
Experimental values:	9.5	4.3	10.7	4.8

The good agreement between the experimental J values and the calculated values for the cis isomer of both compounds, as well as the absence of values around 10 Hz in the calculated J values for the trans isomers, clearly establish that compounds 8 and 10 should have a cis relative stereochemistry as depicted in Scheme 1 and Scheme 2.

Experimental

General

NMR spectra were measured using a Bruker DPX-300 instrument (300 MHz for ¹H-NMR and 75 MHz for the ¹³C-NMR); deuteriochloroform was used as solvent and tetramethylsilane as an internal standard. GC-MS spectra were obtained by EI ionization at 70 eV on a HP-5988-A spectrometer. IR spectra were measured with a Perkin-Elmer 1600 FT spectrometer. TLC was performed on plates precoated with silica gel 60 F₂₅₄ plates (0.25 mm thick, Merck), and for column chromatography, silica gel 60 (70-230 mesh, Merck) was used. Analytical gas chromatography (GLC) separations were performed on a Varian GC 3400 instrument with a fused silica capillary column (30 m length x 0.25 mm i.d.) coated with DB 1701 (phase thickness 0.25 μm) operating at temperatures in the range 50-200ºC.

Preparation of substrates

(±) – 4,4,6-Trimethyl-7-oxabicyclo[4.1.0]heptan-2-one (1).

Prepared as described in the literature [7]; ¹H-NMR: δ 2.93 (dd, 1H, J₁=1.1, J₂=0.9 Hz); 2.51 (dd, 1H, J₁=13.4, J₂=0.9 Hz); 1.97 (dt, 1H, J₁=15.0, J₂=J₃=0.9 Hz); 1.69 (ddd, 1H, J₁=13.4, J₂=2.2, J₃=1.1 Hz); 1.59 (dd, 1H, J₁=15.0, J₂=2.2 Hz); 1.30 (s, 3H); 0.9 (s, 3H); 0.8 (s, 3H); ¹³C-NMR: δ 207.68 (C=O); 64.05 (C); 61.17 (CH); 47.77 (CH₂); 42.51 (CH₂); 35.92 (C); 30.61 (CH₃); 27.61 (CH₃); 23.82 (CH₃); IR (film)ν_max 2952; 2930; 2870; 1717; 1465; 1153 cm^-1; MS m/z (rel. intensity)(%): 154 [M⁺] (4); 139 (24); 111 (10); 97 (24); 83 (100); 69 (60); 55 (48); 41 (35).

7-Oxa-bicyclo[4.1.0]heptane (2) [8].

To an ice-cooled solution of cyclohexene (1.48 g, 18.1 mmol) in methylene chloride (28 mL) was added a solution of m-chloroperbenzoic acid (5.0 g of 85 % MCPBA, 29 mmol) in methylene chloride (70 mL) from a dropping funnel (50 min). The ice bath was removed and the stirring was continued for 1.5 h at room temperature. The resulting mixture was treated with 10 % aqueous sodium sulfite solution (140 mL) and stirred for 1 h to remove excess of peracid. The organic phase was separated and the aqueous phase was extracted with methylene chloride. The combined organic extracts were washed in succession with 5 % sodium bicarbonate and saturated brine and then dried over anhydrous magnesium sulfate. The solvent was removed under vacuum and the residue was distilled (Kugelrohr) at 25 ºC (0.5 mmHg). Yield 1.15 g (65%); ¹H-NMR: δ 3.11 (br. s, 1H); 3.10 (br. s, 1H); 1.94 (m, 2H); 1.81 (m, 2H); 1.43 (m, 2H); 1.23 (m, 2H); ¹³C NMR (75 MHz, CDCl₃) δ 52.09 (CH); 24.53 (CH₂); 19.52 (CH₂); IR (film) ν_max 2977; 2932; 2890; 1439; 1261; 1112; 1072 cm^-1; MS m/z (rel. intensity) (%) 97 [(M-1)⁺] (14); 83 (100); 70 (22); 69 (30); 55 (41); 54 (44); 41 (65); 39 (50).

(±)-1-Methyl-7-oxa-bicyclo[4.1.0]heptane (3).

The same procedure as described for compound 2 was used, starting from 1-methylcyclohexene (1.74 g, 18.1 mmol). Yield 1.50 g (74%); ¹H-NMR: δ 2.86 (t, 1H); 1.69 (m, 4H); 1.21 (s, 3H); 1.12 (m, 4H); ¹³C NMR: δ 59.53 (CH); 57.49 (C); 29.92 (CH₂); 24.80 (CH₂); 23.99 (CH₂); 20.08 (CH₂); 19.70 (CH₂); IR (film) ν_max 2977; 2961; 2931; 1434; 1214; 1114; 1070 cm^-1; MS m/z (rel. intensity) (%) 112 [M⁺] (3); 111 (5); 97 (88); 69 (29); 67 (22); 55 (49); 43 (100); 41(46).

General Procedure for the Reactions of Epoxides with NbCl₅.

A solution of the epoxide 1, 2 or 3 (1.0 mmol) in 1.0 mL of anhydrous solvent (ethyl acetate or acetonitrile) was added to a solution of niobium pentachloride (0.135 g, 0.5 mmol) in 1.0 mL of anhydrous solvent (ethyl acetate or acetonitrile), maintained at room temperature under nitrogen atmosphere. After 1.0 min, the reaction mixture was quenched with a 10% aqueous citric acid solution (2.0 mL). The mixture was diluted with water and ethyl acetate, the organic layer was separated and washed with 5% aqueous sodium bicarbonate, saturated brine, and dried over anhydrous magnesium sulfate. The solvent was removed under vacuum and the products were purified by column chromatography through silica gel using a mixture of hexane and ethyl acetate (8:2) as eluent.

Characterization of the products

(±)-1,4,4-Trimethyl-2-oxo-cyclopentanecarbaldehyde (4):¹H-NMR δ 9.48 (s, 1H); 2.59 (dd, 1H, J₁=13.7, J₂=1 Hz); 2.27 (d, 1H, J₁=17.2 Hz); 2.17 (dd, 1H, J₁=17.2, J₂=1 Hz); 1.61 (d, 1H, J₁=13.7 Hz); 1.3 (s, 3H); 1.17 (s, 3H); 1.03 (s, 3H); ¹³C-NMR δ 215.28 (C=O); 198.68 (C=O); 63.06 (C); 53.14 (CH₂); 44.22 (CH₂); 33.80 (C); 29.63 (CH₃); 28.92 (CH₃); 21.22 (CH₃); IR (film) ν_max 2956; 2868; 2712; 1730; 1714; 1453; 1370; 1149cm^-1; MS m/z (rel. intensity) (%) 139 [(M-15)⁺] (21); 126 (13); 111 (23); 97 (22); 83 (100); 69 (47); 55 (40); 41 (49).

3,5,5-Trimethyl-cyclohexane-1,2-dione (5): ¹H-NMR: δ 6.02 (br. s, 1H); 2.35 (s, 2H); 2.25 (s, 2H); 1.88 (s, 3H); 1.06 (s, 6H); ¹³C-NMR: δ 194.08 (C=O); 143.03 (C); 127.87 (C); 49.31 (CH₂); 44.66 (CH₂); 33.45 (CH₃); 28.36 (CH₃); 17.09 (CH₃); IR (film) ν_max 3402; 2947; 2876; 1666; 1640; 1405; 1173 cm^-1; MS m/z (rel. intensity) (%) 154 [M⁺] (56); 126 (9); 111 (24); 98 (42); 70 (100); 55 (51); 41 (52); 27 (21).

(±)-trans-2-Chloro-cyclohexanol (6): ¹H-NMR: δ 3.70 (ddd, 1H, J₁=11.4, J₂=9.2, J₃=4.4 Hz), 3.49 (m, 1H), 2.58 (br. s, 1H), 2.20-2.10 (m, 1H), 2.10-2.00 (m, 1H), 1.75-1.50 (m, 3H), 1.35-1.15 (m, 3H); ¹³C- NMR: δ 75.36 (CH), 67.55 (CH), 35.16 (CH₂), 33.11 (CH₂), 25.67 (CH₂), 23.98 (CH₂); IR (film) ν_max 3411, 2916, 1451, 1083, 961, 737 cm^-1; MS m/z (rel. intensity) (%) 134 [M⁺] (36), 99 (61); 98 (100); 81 (93); 79 (34); 70 (39) 57 (22); 41 (30).

(±)-trans-2-Hydroxy-cyclohexyl acetate (7): ¹H-NMR: δ 4.57 (ddd, 1H, J₁=10.2, J₂=9.0, J₃=4.7 Hz), 3.54 (ddd, 1H, J₁=10.4, J₂=8.8, J₃=4.5 Hz), 2.78 (br. s, 1H), 2.08 (s, 3H), 2.02 (m, 2H), 1.71 (m, 2H), 1.31 (m, 4H); ¹³C-NMR: δ 171.27 (C=O), 77.92 (CH), 72.32 (CH), 32.89 (CH₂), 29.80 (CH₂), 23.68 (CH₂), 23.60 (CH₂), 21.16 (CH₃); IR (film) ν_max 3441, 2947, 2857, 1730, 1451, 1375, 1248 cm^-1; MS m/z (rel. intensity) (%) 158 [(M-60)⁺] (63); 83 (24); 79 (29); 70 (53); 57 (22); 43 (100); 41 (31); 28 (12).

(±)-cis-2-Chloro-2-methyl-cyclohexanol (8): ¹H-NMR: δ 3.77 (dd, 1H, J₁=9.5 and J₂=4.3); 2.57 (s, 1H); 2.11 (m, 1H); 1.91 (m, 2H); 1.67 (m, 2H); 1.57 (s, 3H); 1.40 (m, 3H); ¹³C-NMR: δ 77.10 (CH); 76.35 (C); 40.68 (CH₂); 30.14 (CH₂); 23.26 (CH₂); 23.18 (CH₂); 22.80 (CH₃); IR (film) ν_max 3456; 2932; 2872; 1451; 1375; 1248; 1084; 1041 cm^-1; MS m/z (rel. intensity) (%) 112 [(M-36)⁺] (5); 84 (3); 68 (10); 56 (13); 55 (16); 41 (21); 32 (25); 28 (100).

(±)-cis-1-Methyl-cyclohexane-1,2-diol (9): ¹H-NMR: δ 3.49 (m, 1H); 1.86 (m, 1H); 1.74 (m, 3H); 1.62 (m, 1H); 1.33 (m, 3H); 1.20 (s, 3H); ¹³C-NMR: δ 77.26 (CH); 73.88 (C); 38.64 (CH₂); 31.02 (CH₂); 24.00 (CH₂); 23.28 (CH₂); 19.69 (CH₃); IR (film) ν_max 3381; 2947; 2857; 1451; 1079; 1050 cm^-1; MS m/z (rel. intensity) (%) 84 [(M-46)⁺] (44); 51 (28); 49 (88); 47 (18); 37 (6); 35 (16); 32 (28); 28 (100).

(±)-cis-N-(2-Hydroxy-1-methyl-cyclohexyl)acetamide (10): ¹H-NMR: δ 5.49 (br. s, 1H); 3.65 (dd, 1H, J₁= 10.7 and J₂=4.7 Hz); 2.00 (s, 3H); 1.4 (s, 3H); 1.91-1.50 (m, 4H); 1.35 (s, 3H); 1.40-1.25 (m, 4H); ¹³C- NMR:δ 171.43 (C=O); 75.95 (CH); 59.82 (C); 38.70 (CH₂); 31.04 (CH₂); 24.29 (CH₃); 23.85 (CH₂); 21.55 (CH₂); 16.47 (CH₃); IR (film) ν_max 3312; 2917; 1650; 1370; 1633; 1077; 734 cm^-1; MS m/z (rel. intensity) (%) 128 [(M-43)⁺] (7); 112 (78); 86 (26); 84 (13); 70 (86); 60 (47); 43 (100); 42 (68).