Stereochemistry & Mechanism of Birch Reduction of Cyclic Allenes

R. Vaidyanathaswamy*, S.N. Moorthy & D. Devaprabhakara
Department of Chemistry, Indian Institute of Technology, Kanpur 208016
14 September 1976

The sodium-ammonia reductions of 1,2-cyclononadiene, 1,2-cyclodecadiene, 1,2-cyclodecadiene, 1,2-cyclodecadiene , 1,2-cyclodecadiene, 1,2-cyclodecadiene , 1-methyl-1,2-cyclodecadiene, 1-methyl-1,2-cyclodecadiene and 1-methyl-1,2-cyclodecadiene have been described. All the cyclic allenes get reduced smoothly t o give good yields of olefinic product(s). Analysis of the products after partial reduction, use of lithium instead of sodium and the relative rates of reduction of cyclic allenes and acetylenes rule out the possibility of the rearrangement of cyclic allenes to acetylenes before reduction. From the ratios of cis to trans olefins obtained in the presence and absence of a proton donor, potential routes for the formation of isometric olefins have been proposed. The synthetic utility of this reduction procedure is borne out in the reduction of some trisubstituted cyclic allenes which offers a convenient stereoselective synthesis of C9?, C10?and C13? trisubstituted cyclic olefins.

Birch reduction has been used for the elegant conversion of allens to olefins 1-9. Several potential pathways have been proposed for this conversion. For example, acetylene was shown to be the intermediate for the predominant formation of trans-olefins from acyclic allenes. This constrasts with the earlier observation made by Svoboda and coworkers10 that some of the cyclic acetylene undergo Birch reduction through isometric allenes. Since acetylene-allene equilibration has been well established both in cyclic¹¹ and acyclic systems12 by sodamide, we undertook a study of sodium-ammonia reduction of cyclic allens. The present paper describes in detail Birch reduction of C9-C14 cyclic allenes, the potential pathways through which the olefins arise and the synthetic utility of this reduction procedure. Both medium- and large-ring cyclic allenes were used for this study as the medium-ring allene is more stable than the corresponding acetylene, whereas in a large-ring the reverse holds good. Two cyclic acetylenes, cyclododecyne and cyclotridecyne and a 1, 3-diene, cis, cis-1, 3 -cyclodecadiene were also chosen as representative cases in order to examine the common path-ways, if any, in these reductions. Three of the trisubstituted cyclic allenes, though reduced primarily from the viewpoint of their synthetic utility, also contribute to the general mechanism of reduction of cyclic allenes.

The disubstituted allenes were synthesized by dibromocarbene addition 13 to respective olefins followed by treatment of dibromocarbene adducts with methylithium 14. The acetylenes15, 16 and 1, 3- diene17 were synthesized by known procedures. The synthesis of trisubstituted allenes has been published in a recent communication 9.

The reductions were carried out under two set of conditions: (i) in the presence of excess sodium and absence of a proton donor and (ii) in the presence of a proton donor like ethanol wherein there was not large excess of sodium. Our results are summarized in Tables 1 and 2.

The following features are apparent from our results: (a) the reduction of C9-allene is stereospecific in the presence or absence of a proton donor, (b) in general, there is a steady increase in the amount of trans-olefin as the ring size increases both under conditions (i) and (ii), (c) there is more trans-olefin formation under condition (ii) than under condition (i) for allenes and the reverse is true for acetylenes and (d) there is high regioselectivity as well as stereo-selectivity in the reduction of trisubstituted allenes.

Because of well-established evidence in the case of acyclic allenes that they undergo reduction partly through the isometric acetylenes 6, 18, it is imperative on our part to prove whether cyclic allenes undergo reduction directly or not. The following evidences are presented to show that acetylene is not an intermediate in these reductions. As a representative case, 1, 2-cyclodecadiene was chosen since it should have the greatest tendency to go over to cyclotetradecyne among the allenes we have studied. On partial reduction of this allene and quenching the product immediately after the blue colour dis-appeared (2 min), there was found no acetylene but only unreacted allene and olefins.

We also found that cycloterradecyne underwent reduction only to the extent of 5% within the time required for 1, 2-cyclodecadiene reduction. For C9?and C10? allenes the question of isomerization to acetylene is less likely11. Furthermore, Svoboda and coworkers10 have shown that cyclododecyne undergoes reduction mainly through its isometric allene, 1, 2-cyclodecadiene and the reductions of cyclic acetylenes are slow. We have also examined further the use of lithium metal in place of sodium since lithium amide is a weaker base and should suppress the formation of trans-olefin in case trans-olefin arises from acetylene. 1, 2-cyclodecadiene was chosen as the substrate for this study since it provides 1:1 isomer ratio and hence it should be most sensitive to this change. It is interesting to point out that there was no significant change in the isomer ratio when lithium was used. All these facts compel us to believe that the reductions of cyclic allenes do not go through their acetylene isomers.

Table 1 – Birch Reduction of Disubstituted Cyclic Allenes

Substrate Product composition (%) (a) Product composition (%) (b)
  cis trans cis trans
1,2-Cyclononadiene 100 - 100 -
1,2-Cyclodecadiene 100 - 66 34
1,2-Cycloundecadiene 73 27 45 55
1,2-Cyclododecadiene 75 25 46 54
1,2-Cyclotridecadiene 50 50 17 83
1,2-Cyclotetradecadiene 27 73 14 86
Cyclodecyne (c) - - 10 90
Cyclododecyne 5 95 27(d) 73(d)
Cyclotridecyne 5 95 13 87
cis, cis-1,3- Cyclodecadiene - - 100 -

(a) The reductions were carried out in the absence of a proton donor (condition i) and the yields were in the range 80-90%; (b) the reductions were carried out in the presence of ethanol as proton donor (condition ii) and the yields were in the range 80-90%; (c) the values are taken from Svoboda and coworkers10; and (d) the conversion was only to the extent 50% after 6 hr.

Table 2 – Sodium-Ammonia Reduction of Trisubstituted Allenes

Substrate Composition, (%)(a)
  cis-1-Methyl-cycloalkene Trans-1-Methyl-cycloalkene
1-Methyl-1,2-cyclononadiene 95(b) -
1-Methyl-1,2-cyclodecadiene 94(b) -
1-Methyl-1,2-cyclotridecadiene - 90(b)

(a) The yields are usually more than 70%.
(b) The major products have not been identified.

Earlier studies 6, 7 on sodium-ammonia reduction of allenes did not consider 1, 3- dienes as possible intermediates. This is not erroneous since a recent report clearly mentions that allenic hydrogens are more acidic than allylic hydrogen 12. Furthermore, it has been observed that in C9?, C10?and C11?systems, the formation of 1, 3-diene is kinetically least important11.

Chart 1 shows the possible routes (path A and path B) for the conversion of allenes to olefins. Since the product composition differs considerably in the presence and absence of alcohol, mechanisms of different nature may operate under these conditions. In the presence of alcohol there could be relatively fast protonation prior to second electron addition to the anion radical as shown in Chart 1 (path B)19. The allenes should undergo reduction directly as no isomerization is possible in the presence of a proton donor. The comparision of cisltrans ratio of 1,2-cyclodecadiene, 1,3-cyclodecadiene and cyclodecyne show that isomer ratios differ considerably and therefore a common intermediate between 1,2- and 1,3- dienes or 1,2- diene and acetylene is ruled out. The most reasonable pathway is the reduction of allene via vinylic and allylic radicals. Since 1, 3-cyclodecadine gives rise to cis-olefin exclusively under these conditions, 34% trans-olefin obtained in the reduction of 1, 2-cyclodecadiene could arise solely from vinylic radical. In this context it is profitable to mention that vinylic radical produced in a C10?system during sodium borohydride reduction of cis-organomercurial 20 gives rise to 22% trans-olefin. However, it is difficult to estimate the individual contribution of allylic and vinylic radicals at the present time in other cyclic systems. It is to be noted that the other cyclic systems. It is to be noted that the cisltrans ratios of olefins from cyclotridecyne and 1, 2-cyclodecadiene approach close to each other. This could be due to either common intermediate (vinyl radical) or due to the stability of the olefin formed irrespective of the intermediate.

In the absence of a proton donor comparison of reductions of cyclic allenes and acetylene to find a common intermediate becomes more difficult since the reductions of acetylenes are not direct. It is surprising to note that there is always a great preponderance for cis-olefin formation over its trans-isomer in C9-C12-allenes. This cannot be explained on the basis of relative stability of olefins21 alone. In the absence of a proton donor and excess of sodium, the allylic dianion (path A, Chart 1) may survive sufficiently longer time to dictate the stereochemical course of reduction. It has been shown on the basis of experimental evidences22 and theoretical calculations 23, 24 that a cis-allylic anion is kinetically more important than its trans-isomer. Hence two important forces namely the formation of cis-allyl-anion (electronic factor) and the stability of olefins (steric or geometric factor) have to be considered to understand the product composition. In C9-and C10-systems, both the forces favour the formation of cis-olefin. But in other systems, these two forces oppose each other. Hence there is a gradual increase of trans-isomer as the ring size increases. Comparison of isomer ratios from C12- and C13-acetylenes and allenes and allenes suggests that a major part of the cis-isomer could arise from cis-allylic dianion.

The high regioselectivity observed in the reduction of trisubstituted allenes further favours our suggestion that the allylic dianion must be the intermediate during the reduction of an allenic system in the absence of a protonating agent. We propose that the least substituted end of the dianion carries the greater amount of charge since the electron pair could occupy an orbital containing more s characters when concentrated there. Our results demonstrate this expected behaviour of the allylic dianion intermediate, and thereby provide a simple and an effective synthetic route to such trisubstituted olefins.

Experimental Procedure

The boiling points (b.p) are uncorrected. GLC analyses were made on a Varian Aerograph model 90-P instrument, on columns filled with 15% silicone rubber SE-30, 20% carbowax 20M and 15% carbowax 20M-5% silver nitrate as liquid phases supported on 60/80 mesh chromosorb P. Samples were separated by preparative GLC for IR and NMR analyses. All the IR spectra (Vmax in cm?¹) were recorded on a Perkin-Elmer model 700 and model 137 spectrophotometers as neat liquids between sodium chloride plates. The products were identified by comparison of GLC retention times and IR spectra with those of authentic samples. NMR spectra were recorded on a Varian A-60 spectrometer in carbon tetrachloride using TMS as an internal standard and the chemical shifts are given in ? values. 1, 2-cyclodecadiene², 1, 2-cyclodecadiene², 1,2-cyclodecadiene², cyclodecyne15, cyclotridecyne16, cis, cis-1,3- cyclodecadiene17, 1-methyl-1,2-cyclodecadiene9 and 1-methyl-1,2-cyclodecadiene9 were prepared by published procedures.

1,2-cycloundecadiene- Methyllithium prepared from lithium (14 g) and methyliodide (142 g) was treated with 11, 11- dibromobicyclo (8.1.0) – undecane25 (112 g) in dry ether (150 ml) under oxygen-free dry nitrogen atmosphere with stirring at ca-40° and worked up to get pure 1,2-cycloundecadiene (43 g); b.p. 92-94°/20 mm (Found: C, 88.12; H, 12.01.C11H18 requires C, 88.00; H, 12.00%); IR (neat):1940 (C=C=C) and 870 (=C-H); NMR (CCL4): 4.91 (2 H, m).

1,2-cyclodecadiene- Following the procedure of Doering and Hoffmann13, from cycloundecene (22 g), potassium (6 g), dry t-butanol (200 ml) and bromoform (38 g), there was obtained 12,12-dibromobicyclo [9.1.0] dodecane (38.8 g) . Its purity was checked by GLC; b.p. 92.94°/0.03 mm (Found: C, 44.32; H, 6.10. C12H20Br2 requires C, 44-45; H, 6.17%).

The dibromo-compound (31.3 g) prepared above was treated with methyllithium prepared from lithium (2.8 g) and methyl iodide (28.4 g) at ca-40° to give only 1,2-cyclodecadiene (12.4 g) as shown by GLC analysis; b.p.109°/10 mm (Found: C, 87-68; H 12.12. C12 H20 requires C, 87.80; H, 12.20%); IR (neat): 1960 (C=C=C) and 871 (=C-H); NMR (CCL4):4-890 (2H, m).

Cyclotetradecadiene- From cyclotridecene (26 g), bromoform (50.6 g), potassium (8 g) and t-butanol (250 ml), there was obtained. 14, 14-dibromobicyclo (11.1.0) tetradecane (33 g); b.p. 115-20°/0.02 mm (Found: C, 47.65; H, 6.70. C14H24Br2 requires C, 47.73; H, 6.82%). The product was found to be pure by GLC.

Treatment of the dibromo-compound (33 g) with methyllithium at ca- 40° prepared from lithium (5.6 g) and methyl iodide (28.4 g) yielded pure 1,2-cyclotetradecadiene (12 g) as examined by GLC; b.p. 78-79°/0.25 mm (Found: C, 87.23; H, 12.71. C14H24 requires C, 87.42; H, 12.58%); IR (neat): 1960 (C=C=C) and 873 (=C-H); NMR (CCL4): 4.92 (2H, m).

1-Methyl-1,2-cyclotridecadine- Following the one step synthesis of allene from olefin26, 1-methylcyclododecene (1.8g) was treated with carbon tetrabromide (3.32 g) and excess methyllithium prepared from lithium (0.56 g) and methyliodide (5.6 g) at ca – 65°. Fractional distillation of the product afforded GLC pure 1-methyl-1, 2-cyclotidecadiene (1.3 g), b.p.53°/0.05 mm (Found: C, 87.28, H, 12.43. C14H24 requires C, 87.50; H, 12.50%); IR (neat): 1935 (C=C=C).

The general procedure for sodium-ammonia reduction: (A) In the absence of alcohol – A 250 ml three-necked round-bottomed flask was equipped with a mercury sealed mechanical stirrer, a dry ice condenser leading to a mercury trap and an inlet for ammonia gas. Dry liquid ammonia (50 ml) was condensed into the flask. After the collection was over, the gas inlet was replaced by a stopper. Freshly cut sodium (0.016 g atom) was added and the blue solution stirred for 15 min. the diene or acetylene (2 mmoles) in dry ether (5 ml) was added by means of a dropping funnel replacing the stopper. Freshly cut sodium (0.016 g atom) was added and the blue solution stirred for 15 min. The diene or acetylene (2 mmoles) in dry ether (5 ml) was added by means of a dropping funnel replacing the stopper. Stirring was continued for 5 min more in the case of allenes and for 3 hr in the case of acetylenes. At the end of the reaction, solid ammonium chloride was added until the blue colour disappeared and ammonia was allowed to evaporate. The residue was diluted with water and extracted with ether twice. The combined ether extracts were washed with dil. Hydrochloric acid followed by water until neutral to litmus. The solution was dried (MgSO4) solvent removed and the residue distilled under reduced pressure. The distillate was subjected to chromatographic and spectral analyses.

(B) In the presence of alcohol- As described above dry ammonia (30 ml) was condensed into a 250 ml three-necked round-bottomed flask. The organic compound (2 mmoles) to be reduced was dissolved in absolute ethanol (15 ml) and ether (5 ml) and added dropwise into liquid ammonia. Sodium was cut into small pieces and introduced slowly with stirring so as to maintain a faint blue colouration. The addition of sodium was fast in the beginning, but had to be considerably slowed down later. After the reaction was over (2 hr in the case of allenes and 1, 3- diene and 6-hr for acetylenes), the reduction product was extracted with pentane and thoroughly washed with water. The pentane solution was dried (MgSO4), solvent removed and the residue distilled under reduced pressure. The distillate was analysed by GLC, IR and NMR.

Partial reduction of 1,2-cyclotetradecadiene- Partial reduction of 1,2-cyclotetradecadiene (1.92 g) with sodium (345 mg) in dry ammonia (10 ml) followed by quenching the product with water soon after the blue colour disappeared gave a mixture containing 1,2-cyclotetradecadiene (40%), cis-cyclotetradecene (15%) and trans-cyclotetradecene (45%). The product was completely free from cyclotetradecyne.


One of the authors (S.N.M) thanks the CSIR, New Delhi, for the award of a senior research fellowship.


  1. GARDNER, P.D. & NARAYANA, M., Chem., 26 (1961), 3518.
  2. DEVAPRABHAKARA, D & GARDNER, P. D., J. Am. Chem. Soc., 85 (1963), 648.
  3. BROWN, J.M., Chemy Ind., (1963), 1689.
  4. VAIDYANATHASWAMY, R. & DEVAPRABHAKARA, D., J. org. Chem., 32 (1967), 4143.
  5. SHARMA, S.N., SRIVASTAVA, R.K. & DEVAPRABJAKARA, D., Can. J.Chem., 46 (1968), 84.
  6. NAGENDRAPPA, G.N., SRIVASTAVA, R.K. & DEVAPRABHAKARA, D., J. org. Chem., 35 (1970), 347.
  7. VAIDYANATHASWAMY, R., JOSHI, G.C. & DEVAPRABHAKARA, D., Tetrahedron Lett., (1971), 2075.
  8. MOORTHY, S.N. & DEVAPRABHAKARA, D., Synthesis, (1972), 612.
  9. MOORTHY, S.N., VAIDYANATHASWAMY, R. & DEVAPRABHAKARA, D., Synthesis, (1975), 194.
  10. SVOBODA, M., ZAVADA, J. & SICHER, J., Colln Czech. Chem. Commun., 30 (1965), 413.
  11. MOORE, W.R. & WARD, H.R., J. Am. Chem. Soc. 85 (1963), 86.
  12. CARR, M.D., GAU, L.H. & REID, I., J. chem. Soc. Perkin Trans., 2 (1973), 672.
  13. DOERING, W. (VON) E. & HOFFMANN, A. K., J. Am. Chem. Soc., 76 (1954), 6162.
  14. SKATTEBOL, L., Acta chem. Scand., 17 (1963), 1683.
  15. ZIEGENBEIN, W. & SCHEIDER, W.N., Chem. Ber., 98 (1965), 824.
  16. NOZAKI, H., KATO, S. & NOYORI, R., Can. J. Chem., 44 (1966), 1021.
  17. VAIDYANATHASWAMY, R. & DEVAPRABHAKARA, D., Indian J. Chem., 13 (1975), 873.
  18. BENKESER, R. A. & TINCHER. C. A., J. org. Chem., 33 (1968), 2727.
  19. HOUSE, H. O. & KINLOCH, E.F., J. org. Chem., 39 (1974), 747.
  20. VAIDYANATHASWAMY, R., RAO, V. V. & DEVAPRABHAKARA, D., Tetrahedron Lett., (1971), 915.
  21. COPE, A. C., MOORE, P.T. & MOORE, W.R., J.Am. Chem. Soc., 82 (1960), 1744.
  22. BANK, S., SCHRIESHEIM, A. & ROWE, C.A. (Jr), J.Am. Chem. Soc., 87 (1965), 3244.
  23. BANK, S., J. Am. Chem.Soc., 87 (1965), 3245.
  24. HOFFMANN, R. & OLOFSON, R.A., J. Am. Chem. Soc., 88 (1966), 943.
  25. MOORE, W.R. & WARD, H.R., J. org. Chem., 27 (1962), 4179.
  26. UNTCH, K.G., MARTIN, D.J. & CASTELLUCCI, N.T., J. org. Chem., 30 (1965), 3572.