Mechanism of olefin rearrangements induced by iron carbonyls. Rearrangements in the bicyclo[6.1.0]nonene system

Received May 29, 1978 9,9-Dichlorobicyclo[6.1 .O]non-4-ene (28) rearranged to the bicyclonon-3-ene isomer 29 and ultimately to bicyclonon-2-ene (30) in the presence of Fe2(C0)9. The tetracarbonyliron(0) complex of the bicyclonon-3-ene isomer (33) was isolated and subjected to similar reaction conditions but failed to produce the rearrangement product 30. Furthermore, the Fe(C0)4 complex corresponding to compound 30 did rearrange with the iron center intact, producing complex 33. These results appear to suggest that two different mechanisms are operating here, one in which an iron carbonyl species is attached to the olefin throughout the rearrangement process and the other possibly not involving intermediacy of an iron carbonyl complex. The crystal and molecular structure of complex 33 was determined by single-crystal X-ray analysis and refined by full-matrix least-squares calculations to R = 0.092 over 1420 statistically significant reflections measured by diffractometer. Crystals are triclinic, space group Pi, a = 7.094 (4) A, b = 16.936 (8) A, c = 6.530 (4) A, a = 95.49 (S)', p = 100.16 (5)', y = 104.05 (5)O, U = 741.2 A3, and Z = 2. The coordinated alkene occupies one equatorial site of a trigonal-bipyramidal iron coordination geometry, mean Fe-C(a1kene) = 2.148 A, Fe-C(carbony1) = 1.794 A. The cyclooctene ring is in a C, chair-boat conformation The reaction of iron carbonyls with olefins, in which isomerization of the olefin is the predominant transformation, is one of the most extensively studied in the area of organoiron carbonyl chemi~try.~ The results of these studies have been strikingly uniform; because of this high degree of uniformity, the reaction has taken on a quality of predictability and has therefore come to be regarded as unexceptional and well understood. Indeed, only two mechanistic proposals have received serious consideration since the reaction was first observed, and the overwhelming weight of evidence has eliminated one of these as a viable p~ssibility.~ In general, the reaction involves reorganization of one olefin to another in the presence of catalytic* amounts of an iron carbonyl (usually Fe(CO)5, Fe2(C0)', or Fe3(C0)12), in an inert atmosphere and inert solvent such as hexane, by use of either elevated temperature or photochemical procedures. In all cases the original u framework of the olefin is undisturbed, so that the reaction is limited to olefin rearrangements which come about by a sequence of hydrogen shifts. The interaction of iron carbonyls with olefins was first observed in 1930 by Reihlen and co-workers in the synthesis of butadieneiron tricarbonyl,' but reports of reaction of iron carbonyls with monoolefins did not appear until 1955, when Asinger and Berg described the conversion of 1 -dodecene in the presence of Fe(CO)S to a mixture of the linear isomers of dodecene.I0 Similar trends have since been reported for l-undecene,lla* n-octenes, and n-hexenes,Ild always with the result that the double bond was seen to migrate preferentially to an internal position. Manuel has observed the reaction of various hexenes in the presence of catalytic quantities of Fe3(C0)12, in which, when conditions of equilibrium are achieved, product ratios reflect values expected on the basis of thermodynamic stabilities of the hydrocarbons themselves.6 Thus, at equilibrium 1-hexene produced a mixture of cis2-hexene (16%), trans-2-hexene (58%), and 3-hexenes (25%), values calculated on the basis of thermodynamic stabilities of which were 20,47, and 32%, respectively. Furthermore, under similar reaction conditions, both 4-methyl- 1 -pentene and 2-methyl- 1 -pentene yielded identical product mixtures composed of these hydrocarbons as well as 2-methyl-2-pentene and 4-methyl-2-pentene with ratios in excellent agreement with values based on predicted thermodynamic stabilities within