Intramolecular ene reactions catalyzed by NbCl5, TaCl5 and InCl3

ARTICLE

Intramolecular ene reactions catalyzed by NbCl₅, TaCl₅ and InCl₃

Carlos Kleber Z. Andrade^*; Otilie E. Vercillo; Juliana P. Rodrigues; Denise P. Silveira

Instituto de Química, Universidade de Brasília, CP 4478, 70910-970 Brasília - DF, Brazil

ABSTRACT

Niobium pentachloride, tantalum pentachloride and indium trichloride efficiently catalysed the cyclization of (R)-citronellal to a mixture of isopulegol and neoisopulegol in good yields. A comparative study was carried out which demonstrated that NbCl₅ is the most active Lewis acid whereas InCl₃ is the most selective one. The selectivity of the reactions varied according to the Lewis acid used and to the solvent. The ene reaction of a 1,7-diene was also investigated. In this case, all Lewis acids tested showed excellent selectivity.

Keywords: citronellal, isopulegol, intramolecular ene reactions, Lewis acids, selectivity

RESUMO

Pentacloreto de nióbio, pentacloreto de tântalo e tricloreto de índio catalisaram eficientemente a ciclização do (R)-citronelal a uma mistura de isopulegol e neoisopulegol, em bons rendimentos. Um estudo comparativo foi realizado demonstrando que NbCl₅ é o ácido de Lewis mais ativo enquanto que InCl₃ é o mais seletivo. A seletividade das reações variou de acordo com o ácido de Lewis utilizado e o solvente. NbCl₅ e TaCl₅ mostraram ausência de seletividade enquanto que o InCl₃ apresentou seletividade moderada em favor do isopulegol. A reação ene de um 1,7-dieno também foi investigada. Neste caso, todos os ácidos de Lewis testados apresentaram excelente seletividade.

Introduction

The carbonyl-ene reaction is synthetically very useful since it enables the formation of a C-C bond with the construction of two vicinal chiral centers.¹ One of the most studied intramolecular versions of this reaction is the cyclization of citronellal (1) (Scheme 1).² The main isomers obtained in this reaction are isopulegol (2) and neoisopulegol (3). This is an important reaction since it is one of the steps in the industrial synthesis of menthol.³ In addition, isopulegol (2) has also been applied to the synthesis of natural products such as pheromones.^4,5

The selectivity of this reaction depends on the Lewis acid employed. In this respect, a great number of Lewis acids have been found to promote this cyclization. Good selectivities favoring isopulegol can be achieved with zinc salts (ZnBr₂ and ZnI₂),⁴ although in stoichiometric amounts, and catalytic scandium triflate,⁶ whereas neoisopulegol is slightly favored using SbCl₅⁴ or molybdenum(II) and tungsten(II) catalysts.⁷

We wish to report that citronellal undergoes cyclization in the presence of catalytic amounts of NbCl₅, TaCl₅ and InCl₃.⁸ NbCl₅ has been successfully used in our group as Lewis acid in a variety of reactions⁹ and has gained increasing attention in recent years.¹⁰ We decided to test TaCl₅ for comparison since niobium and tantalum belong to the same family in the periodic table. InCl₃ was chosen in view of its recent growing applications in organic chemistry as a Lewis acid.¹¹

Results and Discussion

The parameters we analyzed in this study were: solvents, Lewis acids, stereoselectivity, reaction time, reaction temperature and stoichiometric relationship of the reagents.

Using a catalytic amount of NbCl₅, the cyclized products were obtained in all solvents except for hexane (Table 1). In ethanol the yield was low due to the formation of several other products. The yields were better in dichloromethane especially when 4Å molecular sieves were used. The selectivities were low and almost did not vary with the solvent. An interesting trend is the inversion of the isopulegol/noeisopulegol ratio when molecular sieves were employed as can be seen in entries 2, 6 and 7. No reaction was observed when citronellal was submitted to molecular sieves in CH₂Cl₂ in the absence of the Lewis acid. Up to this point the role of the molecular sieves is still unclear. Using a stoichiometric amount of NbCl₅, a complex mixture of products was obtained unless the reaction is run at -78 ºC. At this temperature the reaction was slower but the selectivity was the same. In view of the better results obtained with CH₂Cl₂ and CH₃CN, these solvents were chosen for further studies.

InCl₃ showed a greater selectivity as confirmed by the results shown in Table 2. The use of molecular sieves improved the yields, but lowered the selectivities (entry 3). The reaction did not work in ether. A different result was achieved using TaCl₅ (Table 3). In the presence of a catalytic amount of this Lewis acid, citonellal is slowly consumed (entry 1) whereas when 1.0 equivalent is used the yield of the products was high (entry 2). In both cases, neoisopulegol was the major isomer although in low selectivity.

The isomeric ratio was not improved at lower temperatures (Table 4). NbCl₅ showed almost complete conversion at -40 ºC, even when used in smaller amounts (1 mol% or 5 mol%). Surprisingly neoisopulegol was slightly favored (entry 1), whereas InCl₃ and TaCl₅ showed lower conversions under the same conditions (entries 2 and 3). This indicates that NbCl₅ is the more active Lewis acid.

In order to get a better insight into the influence of different Lewis acids on the cyclization, the reactions were monitored by gas chromatography and some interesting observations could be made.

Using a catalytic amount of NbCl₅, citronellal is consumed in 80 minutes (Figure 1). The relative ratio of isomers varied from 1.2:1 to 2.2:1 in 24h. Using an equimolar amount of this Lewis acid, several other products with higher retention times were detected after 10 minutes of reaction and almost complete decomposition of products was observed after 2h.

In the presence of InCl₃, citronellal is consumed in 60 min with 0.1 equivalent (Figure 2) and in less than 10 min with 1.0 equivalent (Figure 3). Small amounts (4-8%) of a third isomer could be detected. The isomeric ratio remains constant with a catalytic amount (3.3:1) whereas with a stoichiometric amount of the Lewis acid it raises from 3.3:1 in 1h to 6.75:1 in 7h and lowers back to 4.75:1 with prolonged reaction time (24h). This result may be due to decomposition of the isomers to products of higher retention times which are present in significant amounts after 3h (45%).

Using TaCl₅, the consumption of citronellal is not complete even after 22h with 0.1 equivalent (Figure 4) whereas with 1.0 equivalent the reaction takes 30 minutes to go to completion (Figure 5). The isomeric ratio remained constant (near 1.4:1 favoring neoisopulegol) in both cases even after 23h.

In order to be sure that no isomerization was taking place under the reaction conditions, isopulegol and neoisopulegol were separated by column chromatography and separately submitted to the Lewis acids (10 mol% and 100 mol%) in CH₂Cl₂. After 24h the crude reaction mixtures were analysed by GC using biphenyl as internal standard. In all cases no isomerization was observed, but by using 1 equivalent of InCl₃ two products of higher retention time were observed probably due to decomposition and/or dimerization of the isomers whereas when 1 equivalent of NbCl₅ was employed the isomers were almost completely destroyed and many peaks appeared in the chromatogram. In all other cases the isomers were recovered intact after the work-up.

The second type of ene reaction we studied was the cyclization of compound 6 which was prepared by the Knoevenagel condensation of citronellal with diethylmalonate (Scheme 2) according to a known procedure.¹²

1,7-Dienes undergo the intramolecular ene reaction less readily than the corresponding 1,6-dienes. The temperatures required are higher and the reactions proceed in lower yields except for activated components. That is the case for compound 6, which has two electron withdrawing groups. This reaction is usually very selective and the main product has all the substituents in equatorial positions. Indeed, for all Lewis acids studied the yields and selectivities were very good (Table 5). There was no need for molecular sieves in this reaction. These results are similar to those obtained with ZnBr₂.¹²

Conclusion

NbCl₅, TaCl₅ and InCl₃ efficiently catalyzed intramolecular ene reactions. Whereas NbCl₅ was the most active, InCl₃ was the most selective Lewis acid. TaCl₅ showed inverse selectivity slightly favoring neoisopulegol in the cyclization of citronellal. The best conditions for this reaction turned out to be the use of 0.1 equivalent of InCl₃ in CH₂Cl₂ at room temperature for 1h. The yields obtained in these reactions were invariably good but the selectivities were low except in the cyclization of compound 6 in which compound 7 could be obtained almost as the only isomer.

Experimental

General

All reactions involving Lewis acids were performed under argon in a flame-dried flask. Ethereal solvents were distilled from sodium benzophenone ketyl. All other solvents were distilled from CaH₂ prior to use. NbCl₅ was supplied by Companhia Brasileira de Mineração e Metalurgia (CBMM). TaCl₅ and InCl₃ were purchased from Aldrich. (R)-Citronellal (Fluka) (GC t 20.48 min) was distilled prior to use. Compound 6 was prepared from (R)-citronellal according to a procedure described by Tietze et al.¹²

IR spectra were recorded on a BOMEM Hartman & Braun — Michelson MB series 100 LASER FT-IR. NMR spectra were recorded on a Varian Mercury Plus 300 spectrometer. Chemical shifts are reported in ppm from tetramethylsilane as internal reference. Data are reported as follows: chemical shift, multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, dd = doublet of doublets, td = triplet of doublets, m = multiplet, etc), coupling constant and integration.

Column chromatography was performed on silica gel (70-230 mesh). GC analyses were carried out in a Shimadzu GC chromatograph, model GC-17A, employing a CBPI PONA-M50-042 column (50 m x 0.15 mm x 0.20 µm). The following parameters were used: initial temperature = 80 ºC; final temperature = 180 ºC; initial time = 1 min; final time = 30 min; detector temperature = 250 ºC; injector temperature = 250 ºC; temperature gradient = 10 ºC/min. Biphenyl was used as internal standard (GC t 25.00 min).

General procedure for the cyclization of (R)-citronellal. To a stirred suspension of the Lewis acid (0.13 mmol) in the specified solvent (4 mL) under an argon atmosphere, at 0 ºC, was added (R)-citronellal (1.3 mmol). The reaction was stirred at room temperature and the consumption of the reagent was monitored by GC and/or TLC. The reaction is generally finished within 1h. Then, sat. NH₄Cl (4 mL) was added and the reaction mixture was extracted with CH₂Cl₂ (3 x 10 mL). The combined organic phases were dried (Na₂SO₄), filtered and concentrated at reduced pressure. The crude product was then purified by column chromatography (10% EtOAc in hexanes). The spectroscopic data are identical to those reported for compounds 2^4,13,14 and 3.¹⁴

Isopulegol, (2). IR (film) n_max/cm^-1: 3405, 3073, 2949, 2925, 2869, 1645, 1455, 1375, 1051, 1027; 887; ¹H NMR (300 MHz, CDCl₃): d 0.95 (d, J 6.5 Hz, 3H), 1.10-2.10 (m, 9H), 1.71 (s br, 3H), 3.45 (td, J 10.1 and 4.4 Hz, 1H), 4.85 (br s, 1H), 4.90 (m, 1H). ¹³C NMR (75 MHz, CDCl₃): d 19.1, 22.1, 29.6, 34.3, 42.6, 54.0, 70.3, 112.7, 146.6. GC t 20.79 min.

Neoisopulegol, (3). IR (film) n_max/cm^-1: 3465, 3085, 2943, 2925, 2861, 2845, 1643, 1455, 1375, 1024, 890; ¹H NMR (300 MHz, CDCl₃): d 0.88 (d, J 6.4 Hz, 3H), 1.10-2.10 (m, 9H), 1.79 (s br, 3H), 3.99 (q, J 2.5 Hz, 1H), 4.79 (br s, 1H), 4.95 (br s, 1H). ¹³C NMR (75 MHz, CDCl₃): d 21.8, 22.1, 31.3, 34.7, 40.9, 52.7, 66.3, 111.2, 147.2. GC t 21.04 min.

General procedure for the cyclization of compound 6. To a stirred suspension of the Lewis acid (0.05 mmol) in CH₂Cl₂ (4 mL) under an argon atmosphere, at 0 ºC, was added compound 6 (0.52 mmol). The reaction was stirred at room temperature for 4h. Then, sat. NH₄Cl (4 mL) was added and the reaction mixture was extracted with CH₂Cl₂ (3 x 10 mL). The combined organic phases were dried (Na₂SO₄), filtered and concentrated at reduced pressure. The crude product 7 was then purified by column chromatography (10% EtOAc in hexanes). The spectroscopic data are identical to those reported for compound 7.¹²

IR (film) n_max/cm^-1: 3060, 1750, 1735, 1645, 1155, 1035, 895; ¹H NMR (300 MHz, CDCl₃): d 0.90 (d, J 6.5 Hz, 3H), 0.94-1.84 (m, 7H), 1.27 (t, J 7.0 Hz, 3H), 1.28 (t, J 7.0 Hz, 3H), 1.65 (s br, 3H), 2.03-2.17 (m, 2H), 3.51 (d, J 3.1 Hz, 1H), 4.17 (q, J 7.0 Hz, 2H), 4.18 (q, J 7.0 Hz, 2H), 4.73-4.76 (m, 1H), 4.78-4.80 (m, 1H); ¹³C NMR (75 MHz, CDCl₃): d 14.0, 14.1, 18.9, 22.4, 32.3, 32.7, 34.6, 36.5, 39.7, 48.6, 53.2, 60.6, 61.0, 112.4, 147.5, 168.6, 169.7; GC t (major isomer) 35.10 min; t (minor isomer) 36.12 min.

Acknowledgments

The authors thank the Instituto de Química, Universidade de Brasília, for financial support, FINEP-CT INFRA nº 0970/01, CBMM for NbCl₅ samples and Prof. Inês S. Resck (UnB) for high field NMR spectra.

References

1. Snider, B. B. In Comprehensive Organic Synthesis; I. Fleming, ed.; Pergamon: Oxford, 1991, vol. 2; p. 527; Mikami, K.; Shimizu, M.; Chem. Rev. 1992, 92, 1021; Snider, B. B.; Acc. Chem. Res. 1980, 13, 426; Oppolzer, W.; Snieckus, V.; Angew. Chem. Int. Ed. Engl. 1978, 17, 476; Hoffmann, H. M. R.; Angew. Chem. Int. Ed. Engl. 1969, 8, 556. For an excellent review on asymmetric ene reactions, see: Dias, L. C.; Curr. Org. Chem. 2000, 4, 305.

2. For representative examples of this reaction, see: Jacob, R. G.; Perin, G.; Loi, L. N.; Pinno, C. S.; Lenardão E. J.; Tetrahedron Lett. 2003, 44, 3605; Chuah, G. K.; Liu, S. H.; Jaenicke, S.; Harrison, L. J.; J. Catal. 2001, 200, 352; Yadav, G. D.; Nair, J. J.; Langmuir 2000, 16, 4072; Arata, K.; Matsuura, C.; Chem. Lett. 1989, 1797. See also refs. 4, 5, 6 and 13.

3. Takeshi, I.; Yoshiki, O.; Yoji, H.; European Patent EP1225163, July 24, 2002.

4. Nakatani, Y.; Kawashima, K.; Synthesis 1978, 147.

5. Moreira, J.; Corrêa, A. G.; J. Braz. Chem. Soc. 2000, 11, 614.

6. Aggarwal, V. K.; Vennall, G. P.; Davey, P. N.; Newman, C.; Tetrahedron Lett. 1998, 39, 1997.

7. Koèovský, P.; Ahmed, G.; Šrogl, J.; Malkov, A. V.; Steele, J.; J. Org. Chem. 1999, 64, 2765.

8. Part of this work was presented at the X Brazilian Meeting on Organic Synthesis (BMOS), August 2003, São Pedro, SP, Brazil.

9. Andrade, C. K. Z.; Azevedo, N. R.; Oliveira, G. R.; Synthesis 2002, 928; Andrade, C. K. Z.; Azevedo, N. R.; Tetrahedron Lett. 2001, 42, 6473; Andrade, C. K. Z.; Oliveira, G. R.; Tetrahedron Lett. 2002, 43, 1935; Andrade, C. K. Z.; Matos, R. A. F.; Synlett 2003, 1189; Andrade, C. K. Z.; Kalil, P. P.; Rocha, R. O.; Alves, L. M.; Panisset, C. M. A.; Lett. Org. Chem. 2004, 1, 109.

10. Review on NbCl₅ applications in organic synthesis: Andrade, C. K. Z.; Curr. Org. Synth. 2004, 4, 333!

11. Ranu, B. C.; Eur. J. Org. Chem. 2000, 2347.

12. Tietze, L. F.; Beifuss, U.; Ruther, M.; J. Org. Chem. 1989 54, 3120.

13. Jensen, B. L.; Malkawi, A.; McGowan V.; J. Chem. Ed. 2000, 77, 1474.

14. Kropp, P. J.; Breton, G. W.; Craig, S. L.; Crawford, S. D.; Durland, Jr., W. F.; Jones, III, J. E. J.; Raleigh, J. S.; J. Org. Chem. 1995, 60, 4146.

Received: November 12, 2003

Published on the web: September 7, 2004

Publication Dates

• Publication in this collection
  24 Jan 2005
• Date of issue
  Dec 2004

History

• Received
  12 Nov 2003
• Accepted
  07 Sept 2004