Res. Chem. lntermed. Vol. 19, No. 8, pp. 797-805 (1993) 9VSP 1993 ISOMERISATION OF DECATETRAENES AND D I M E T H Y L D E C A T E T R A E N E S VIA P E N T A D I E N Y L AND 3-METHYLPENTADIENYL RADICALS J.C. Walton" Department of Chemistry, University of St. Andrews, St. Andrews, Fife KYI6 9ST, UK and K.U. Ingold Steacie Institute for Molecular Sciences. National Research Council of Canada, Ottawa, Canada KIA OR6 Recieved 25 June 1993; accepted 23 July 1993 Abstract--The thermal isomerisations of E,E-deca-l,3,7,9-tetraene and E,E-3,7-dimethyldeca-l,3,7,9tetraene take place via the intermediacy of pentadienyl and 3-methylpentadienyl radicals, respectively, rather than by concerted Cope type rearrangements. The pentadienyl radicals isomerise to pairs of E- and Z-pentadienyl radicals which recombine by end to end and end to centre, but not centre to centre, coupling to give mixtures of isomeric decatetraenes. The relative free energies of formation of these decatetraenes were derived from their equilibrium proportions and compared with relative enthalpies of formation calculated by the empirical MM2(87) method and the semi-empirical AM1 and PM3 SCF MO methods. None of these theories was successful at predicting and rationalizing the experimentally observed enthalpy changes. However, they all were somewhat better in dealing with branching than with cis/trans isomerisation. Frontal strain in the decatetraene plays an important part in influencing their stability. INTRODUCTION Allyl radicals undergo a rapid end to end combination reaction to give hexa-l,5-diene, 1, in both gas [1] and solution phases [2]. In the latter phase the reaction is essentially diffusion controlled [3] and has been used in several synthetic procedures for hexa-l,5diene [4-7]. On thermolysis 1 rearranges by a 3,3-shift in a degenerate isomerisation (Cope rearrangement) which is believed to be concerted for 1 and many derivatives [8], i.e. does not involve allyl radicals as intermediates Eq. (1). The next higher vinylogous radical in the open chain series of delocalised polyenyl radicals is pentadienyl 2. EPR spectroscopy [9], combined with other evidence [10], indicates that pentadienyl radicals combine at, or close to the diffusion controlled limit in solution. However, the products formed from coupling of two pentadienyl radicals are more complex than in the allyl case for two reasons. First, pentadienyl has significant central unpaired electron density at C(3) [9] and potentially can react by end to centre 798 J.C. Walton and K.U lngold (1) la lb and centre to centre coupling. Second, the E,E-pentadienyl radical 2a readily isomerises to the E,Z-pentadienyl radical 2b, the Arrhenius activation barriers for the forward and reverse conformational changes being [11] 11.7 and 9.3 kcal mol ~ respectively. The potential products from dimerisation reactions are summarised in Scheme 1. Because the combination reactions leading to dimer formation are very rapid, equilibria between the dimers will be set up at temperatures high enough to induce thermal dissociation of the carbon sp3-sp3 bonds in the decatetraenes. The thermodynamic stabilities, and hence proportions, of the decatetraenes will be influenced by stereoelectronic effects. We have studied the equilibria experimentally for two sets of tetraenes, starting in one case from E,E-deca-l,3,7,9-tetraene, 3, and in the other case from the 3,7-dimethyl analogue in which an additional impediment to central coupling was provided by the methyl substituents [12]. Of particular interest is the ability of modem theories to predict the influence of some rather subtle steric effects on the thermodynamic stabilities of isomeric hydrocarbons. Thus, these experiments provided a unique opportunity to test whether current theoretical models could rationalise the observed equilibria. We calculated the 3 2a -" 6 II 4 5 2b tt 8 Scheme 1 7 799 lsomerisation o f Decatetraenes and Dimethyldecatetraenes thermodynamic parameters o f the twelve decatetraenes using two empirical methods (BGC and MM2(87)) and two semi-empirical SCF MO methods (AM1 and PM3). RESULTS The equilibria between tetraenes 3 to 8 were studied by thermolysis of 3 at 475 K in tetradecane solution. The development of the product mixture was monitored, as a function o f time, by GC-MS. The decatetraene structures were established by IH and 13C N M R spectroscopy o f components separated by preparative GLC [12]. The ratio o f products became constant after ca. 1.5 h, due to establishment of the equilibria. The proportions o f individual decatetraenes, their equilibrium constants with 3, and the derived relative free energies are given in Table 1. Table 1 Thermolysis of E,E-deca-l,3,7,9-tetraene (3) at 475 K in hexadecane. Tetraene mol %~ K (475) hAG (475) kcal mol"~ 3 56 1.0 0.0 4 6 0.11 2.1 5 <2 <0.04 >3.1 6 23 0.411 0.8 7 15 0.268 1.2 8 <0.5 <0.01 >4.3 "% Total product at equilibrium (5h). A similar experimental study was carried out of the equilibria set up when a mixture o f E,E-3,7-dimethyldeca-l,3,7,9-tetraene, 9, and the E,Z-isomer, 10, was thermolysed at 473 K in hexadecane. Equilibrium was established after ca. 4 h and the product proportions and derived thermodynamic data are given in Table 2. The Z,Z-tetraene 11 was not detected, but we cannot rule out the possibility that it was unresolved from the peaks o f 9 or 10 on the product chromatogram. 800 J.C. Walton and K.U lngold 9 12 10 13 11 14 Table 2 Thermolysis ofamixture ofE, E-3,7-dimethyldeca-l,3,7,9-tetraene at 473 K in hexadecane. (9) and the E, Z-isomer (10) Tetraene tool %" K (473) AAG (473) kcal mol "~ 9 42.9 1.0 0 10 54.1 1.3 -0.25 11 n.o. b . 12 1.6 0.037 3.1 13 1.4 0.033 3.2 14 <0.5 <0.01 >6.3 . . . ~% Total product after 4.75h at 473 K. bNot observed (see text). DISCUSSION T h e p r o d u c t p r o p o r t i o n s indicated that, for the unsubstituted decatetraenes, e n d to e n d a n d e n d to centre coupling were b o t h important. Dimers containing a cis double b o n d , i.e. 4 and 7 were important, but n o n e o f 5, containing two cis double bonds, w a s identified. Isomerisation of Decatetraenes and Dimethyldecatetraenes 801 For the dimethyldecatetraenes, end to end dimerisation predominated and the end to centre dimers, 12 and 13 formed a minute proportion of the total. However, the E,Zisomer 10 and the E,E-isomer 9 were almost equal in concentration at equilibrium. Probably this is because the enthalpy difference between the tri-substituted 7,8-double bonds in 9 and 10 is small, i.e. it makes little difference if a methyl or a vinyl group is cis to the hexadienyl substituent at the other end of the double bond. In view of this it is surprising that none of the Z,Z-dimethyldecatetraene 11 was detected. We cannot rule out the possibility that some of this isomer could have been hidden under other peaks on the chromatograms. For both sets of tetraenes none of the centre to centre dimers (8, 14) could be detected; presumably because frontal strain in the dimers disfavours this type of coupling. A possibility that must be considered is that concerted Cope rearrangements of the decatetraenes contribute to the overall equilibrium. This electrocyclic 3,3-shift could isomerise 3 to 8 and a second such rearrangement could then regenerate 3 (Eq. (2)). 9 3 ! 8 (2) 3' However, neither 6 nor 7 can be formed from 3 by Cope rearrangements. Thus the full spectrum of products cannot be accounted for in terms of Cope rearrangements. Another piece of evidence which militates against the Cope mechanism is that for thermolysis of 3, radical 2 was directly detected by EPR spectroscopy and, for the 9 + 10 mixture, the 3-methylpentadienyl radical was spectroscopically identified under thermolysis conditions similar to those employed in the isomerisation studies [12]. There is therefore an intriguing mechanistic contrast between the thermal rearrangements of hexa-l,5-dienes on the one hand, which are usually concerted Cope type reactions, and of deca-l,3,7,9tetraenes on the other hand which prefer discrete radical intermediates. The activation energy of the Cope process is greatly lowered by the presence of an oxy-anion at C(3) of the hexa-l,5-diene (Oxy-Cope rearrangement). It would be of interest to see if such a substituent at C(5) of a deca-l,3,7,9-tetraene would increase the rate of the Cope rearrangement so as to make it competitive (or dominant) with the free radical process and hence open up a route to branched tetraenes of type 8 and 14. The enthalpies (AHt~ and entropies (S ~ of formation of the decatetraenes were calculated by the empirical Group Contributions method (BGC) of Benson and coworkers [13,14]. In this method the total enthalpy is obtained by summing contributions from standard groups of atoms which are assumed to contribute to the same extent in different molecules. Obviously, this assumption would not hold for molecules with strained or distorted structures, and therefore a limited number of structure factors were worked out to correct for this. In the case of the decatetraenes the frontal strain can only be allowed for in terms of the number of gauche interactions present. Thus, good 802 Jc. Walton and K.U. lngold precision in the AHf~ values can hardly be hoped for. The BG thermodynamic data is listed in Table 3. The empirical MM2(87) method uses a force field which adapts for individual molecular structures and has been steadily upgraded and improved [15-17]. version o f this method is generally hydrocarbons rather well. believed to handle conjugated The '87 n-systems in We also computed the AHf~ values and structures o f the decatetraenes using the AM1 and PM3 semi-empirical S CF M O methods [ 18,19] with full optimisation with respect to all geometrical degrees o f freedom. The AHf~ values from these three methods are also listed in Table 3. Table 3 Computed thermodynamic data for decatetraenes? BGC AHfo So MM2(87) AM 1 PM3 AH? AH? AHf~ 3 47.3 116.3 42.9 44.6 50.3 4 48.3 116.3 45.1 45.3 51.1 5 49.3 116.3 47.1 46.2 51.7 6 52.7 117.6 52.6 54.5 58.7 7 53.7 117.6 54.6 55.7 59.6 $ ' 58.0 118.9 66.6 66.5 69.9 9 34.9 134.7 33.5 34.4 34.1 10 34.9 134.7 34.6 34.4 34.9 11 34.9 134.7 35.9 33.4 35.1 12 43.5 134.4 42.5 47.7 48.7 13 43.5 134.4 44.1 52.2 45.8 14 53.1 133.7 61.4 68.6 64.5 Decatetraene 'AI-If~ values in kcal mol~, S~ values in cal Kt mol-~. Within each group o f tetraenes the S ~ values varied by < 2.5%. hydrocarbons, another. Thus, for these the enthaply and free energy changes will be linearly related to one All four computational methods found that, within each group o f tetraenes, the lsomerisation of Decatetraenes and Dimethyldecatetraenes 803 Figure 1. Correlation of computed enthalpy of formation differences with experimental free energy differences for decatetraenes. Squares: MM2(87). Circles: AM1. Triangles: PM3. Arrows originate at the experimental lower limits of the AAG(475) values for 8 and 14. 804 Jc. Walton and K.U Ingold AHf~ values increased with the degree of branching and the number of cis double bonds present, i.e. in the order 3 --> 8 and 9 ---> 14. The only exceptions being AHf~ derived by the AM1 method and AHf~ found by the PM3 method. However, all four methods found very similar AHf~ values for the linear dimethydecatetraenes 9 to l l . Clearly, all four theories predict that 11 should make a significant contribution to the equilibrium. As indicated above, the absence of experimental evidence for 11 could be because it was hidden under another chromatographic peak. All four theories predict that the centre to centre dimers 8 and 14 have much the highest AHf~ values in their group, and this is in good accord with the zero amounts detected. The BGC values for AHf~ and AHf~ were considerably smaller than those calculated by the other methods. It is probable that the gauche structure factors of the BGC method underestimate the frontal strain in these heavily branched isomers. Largely because of this, the BGC AHf~ values correlated rather poorly with the other three sets of computed data; for example, the linear regression line versus the MM2(87) data had a non-zero intercept (12.2 kcal tool -1) and a gradient (0.73) which deviated significantly from the optimum value of 1.0. On the other hand the AM1 and PM3 AHf~ correlated well with the MM2(87) data giving near zero intercepts, gradients near unity and correlation coefficients (r 2) of 0.927 and 0.956 respectively. The small variations of S Owithin each group of tetraenes (Table 3) indicated that AG should be proportional to AH for this restricted set of compounds. The correlations of the experimental AAG data (Tables 1 and 2) with AAHf~ [= AHf~ - AHf~ and AHf~ - AHf~ where n refers to the other isomers] computed by the MM2(87), AM1 and PM3 methods are shown in the figures. Despite a general trend towards higher AAHf~ for higher experimental AAG, it would appear that current empirical and semiempirical methods for calculating the enthalpies o f formation o f fairly simple hydrocarbons could stand improvement! The optimum structures of the tetraenes were also computed but, in the absence of experimental structural data, we only mention that the calculated lengths of the central sp3-sp 3 bonds in the decatetraenes increased with the extent of branching. For example (AM1 data) from 1.518 in 3 to 1.555 A in 14. This is consistent with the idea that increase in frontal strain is of major importance in controlling the dimerisation equilibria. EXPERIMENTAL Full details of the thermolysis procedure, product characterisation and quantitative analysis have already been published [12]. Acknowledgements KUI and JCW thank NATO for a travel grant which helped to make this research possible. lsomerisation of Decatetraenes and Dimethyldecatetraenes 805 REFERENCES 1. 2. 3. 4. D.G.L. James and M.S. Kambanis, Trans. Faraday Soc. 65, 1350 (1969). H.-G. Korth, H. Trill, and R. Sustmann, J. Am. Chem. Soc. 103, 4483 (1981). H.-G. Korth, P. Lommes, W. Sicking and R. Sustmann, Int. J. Chem. Kinet. ~ 267 (1983). R.F. Garwood, N. Ud Din, C.J. Scott, and B.C.L. Weedon, J. Chem. Soc., Perkin Trans 1, 2714 14. 15. (1973). J. Grignon, C. Servens, and M. Pereyre, J. Or~anometal. Chem. ~ 225 (1975). R. Sustmann and R. Altevogt, Tetrahedron Lett. ~ 5167 (1981). H. Higuchi, T. Otsubo, F. Ogura, and H. Yamaguchi, Bull Chem. Soc. Jpn. ~ 182 (1982). For a review see J.J. Gajewski, Hydrocarbon Thermal Isomerisations, Academic Press, New York, 1981, p. 166. A.G. Davies, D. Griller, K.U. Ingold, D.A. Lindsay, and J.C. Walton, J. Chem. Soc. Perkin Trans. 2 633 (1981); D. Griller, K.U. lngold, and J.C. Walton, J. Am. Chem. Soc. ~ 758 (1978); R. Sustmann and H. Schmidt, Chem. Ber. 112. 1440 (1979). J.C. Walton, J. Chem, Soc. Perkin Trans 2 173 (1989). I. Maclnnes and J.C. Walton, J. Chem. Soc., Perkin Trans 2 1073 (1985). P.N. Culshaw, J.C. Walton, L. Hughes, and K.U. Ingold, J. Chem. Soc., Perkin Trans. 2 879 (1993). S.W. Benson, F.R. Cruickshank, D.M. Golden, G.R. Haugen, H.E. O'Neal, A S. Rogers, R. Shaw, and R. Walsh, Chem. Rev. ~ 279 (1969). S.W. Benson, Thermochemical Kinetics, 2nd, Ed., Wiley, New York, 1976. N.L. Allinger, Adv. Phys. Or~. Chem. ~ 1 (1976); J. Kao and N.L. Allinger, J. Am. Chem. Soc. 16. 17. 18. 19. 99, 975 (1977). J.Y. Sprague, J.C. Tai, Y. Yuh, and N.L. AIlinger, J. Compt. Chem. 8, 581 (1987). Y. Yuh and N.L. Allinger, QCPE, No. 501, University of Indiana, Indiana, 1988. M.J.S. Dewar, E.G. Zoebisch, E.F. Healy, and J.J.P. Stewart, J. Am. Chem. Soc. 107, 3902 (1985). J.J.P. Stewart, QCPE, No. 455, University of Indiana, Indiana, 1987. 5. 6. 7. 8. 9. 10. 11. 12. 13.