J. Chem. SOC., Furuduy Trans. 1, 1988, 84(10), 3233-3242 Electron Spin Resonance Studies of Cycles and Bicycles Finlay MacCorquodale and John C. Walton" University of St. Andrews, Department of Chemistry, St. Andrews, Fife KY16 9ST The e.s.r. spectra of cyclohexylmethyl radicals have shown that two conformations, one with the CH; group equatorial and one with the CH; group axial, can be distinguished. The H, h.f.s. of the axial conformer is large because rotation about the C,-C; bond is hindered by axial hydrogens at C(3) and C(5) in the ring. This proved to be a very useful property enabling the conformations and ring-inversion barriers of cyclohexanes, cyclohexenes and related radicals to be studied by e.s.r. spectroscopy. In addition the various conformers of larger rings (up to 15- membered) also show different spectra.Their preferred conformations and the dynamics of 'corner migration' in the medium rings have been investigated. The e.s.r. spectra of cycloheptenylmethyl radicals showed the presence of a minor conformer which suggested that transannular cyclisation might be important. Product analysis confirmed that bi- cyclo[3.2. lloctane can be obtained in good yield. The stabilisation energy of cyclopropylmethyl radicals was determined from exchange-broadened spectra ; several cyclic homoallyl-type radicals were shown to have essentially zero stabilisation. Hydrogen abstraction from small strained bicycloalkanes, including bicyclo[n . 1 . Olalkanes, spiro[2. nlalkanes and spiro[3. nlalkanes yields the corresponding strained bicycloalkyl radicals, and their rearrangements have been followed by e.s.r.spectroscopy. Bicyclo[ 1 . 1 . 1 Jpentane and bi- cyclo[2.1. llhexane are unusual in that bridgehead radicals are formed. Bicyclo[2.2. Olhexane also shows significant bridgehead reactivity and provides the first example of an SH2 reaction involving a four-membered ring. Bicyclo[3.2. Olheptane, in which four- and five-membered rings are fused together, does not undergo this SH2 reaction with halogens. The rates of rearrangement of several cycloalkyl and bicycloal kyl radicals have been determined by kinetic e.s.r. spectroscopy. The chemical properties that alicyclic rings can show are inherent in their molecular structure and shape. The conformational options open to a particular ring system are specially important in determining the type of chemistry it can display.N.m.r. spectroscopy has been widely exploited in the study of cyclohexanes and medium ring~,l-~ but the closely related technique of e.s.r. spectroscopy has been far less used. The main reason for this lies in the fact that most cyclic radicals so far studied contain the planar radical centre in the ring. This drastically alters the ring's conformation in comparison with that of the parent alicyclic compound, so that the results are of very limited significance for the general chemistry of stable cyclic molecules. This applies, for example, to the e.s.r. work on nitroxide conformations4 and that on mono- and bi-cyclic semidiones. A cycloalkylmethyl radical (1) contains the small, non-polar CH; group which causes minimal perturbation of the adjacent ring.Study of these radicals enables the e.s.r. 32333234 E.S.R. of Cycles and Bicycles method to be successfully applied to conformational analysis. Effectively, the CH; group acts as a ‘spin probe’ of both the ring conformation and the dynamic processes in which the ring takes part. Because the CH; group has such a minor influence itself, the results are of direct interest to all chemists working with alicyclic rings. Six- and Seven-membered Rings The starting point in this e.s.r. approach to conformational analysis is the cyclohexylmethyl radical. At low temperatures only one radical can be observed,‘ with (at 140 K) a(2H,) = 21.5 G, a(H,) = 30.4 G. This is the chair conformer (2) with the CH; group in the equatorial position.At higher temperatures a second conformer can be o b s e r ~ e d , ~ . ~ which has (at 182 K) a(2H,) = 21.5 G, together with a much larger H, h.f.s. of 41.2 G. That this is the axial conformer (3) was confirmed in two ways. First, the radical derived from trans-4-t-butylcyclohexylmethyl bromide, which must be the equatorial conformer, had an H, h.f.s. very similar to that of (2), whereas the radical derived from cis-4-t-butylcyclohexylmethyl bromide (4) in which the But group necessarily occupies the equatorial position, had an e.s.r. spectrum very similar to that of (3) and with a(H,) = 41.9 G at 140 K.’, * Secondly, the 2-adamantylmethyl radical (5), in which the rigid structure ensures that the CH; group is axial in one ring while the H, is equatorial in this same ring, had an H, h.f.s.of 41.1 G at 140 K. (5 1 The surprisingly large difference in the H, h.f.s. of the axial and equatorial radicals has its origin in the different rotational energy functions of the C a C ; bonds. The equatorial radicals have a(H,) values similar to those of open-chain analogues such as isobutyl [a(H,) = 31.7 G at 140 K].‘ In the axial radicals the barrier is much higher because of steric hindrance from the syn-axial hydrogens at C(3) and C(5) [see (6)]. It is well known that axial substituents at C( 1) experience steric repulsion from syn-axial hydrogen^,^ but the effect on the rotation of the substituent at C(l) had not previously been discovered.F. MacCorquodale and J. C. Walton 3235 The axial radicals showed six-line multiplets from long-range splittings at low temperatures.As the temperature was increased the lines broadened, coalesced at ca. 210 K and finally sharpened into an eight-line pattern at 280 K. We attribute this exchange-broadening to restricted rotation about the C,-C; bond, and a barrier of ca. 6 kcal mol-1 can be estimated from the coalescence temperature.? This is a surprisingly high barrier and it highlights the importance of this kind of syn-axial interaction. In cycloalkylmethyl radicals it has the important consequence of making axial and equatorial conformers distinguishable, and hence enables the e.s.r. method to be extended to a wide range of ring systems. The equatorial preference of the CH; group was quantified by measurements of the concentrations of the two conformers at a series of temperatures.The conformational free-energy difference of the CH; group (- AGioo) was found to be 0.71 kcal mol-'.f This value gives a rough measure of the 'size' of the group and thus CH; is 'larger' than OH and OMe, for which -AGioO values of 0.52 and 0.60 kcal mol-' have been determined,ll but 'smaller' than CH, (-AGioo = 1.74 kcal mo1-')l2 or other alkyl groups.13 Two cyclohexenylmethyl radicals have been studied by e.s.r. spectroscopy. The quasi- equatorial (7) and quasi-axial conformers (8) of the cyclohex-2-enylmethyl radical can be distinguished by their H, h.f.s., which were 30.6 and 32.6 G, respectively, at 140 K.14 (7 1 (8 ) The conformational free-energy difference ( - AGioo) was found to be 0.17 kcal mol-', i.e.much less than in cyclohexylmethyl radicals, as would be expected. The spectra showed exchange broadening in the range 170-280 K, from which the cyclohexene ring inversion barrier was found to be ca. 5.6 kcal mol-', i.e. in reasonable agreement with the free energy of inversion as measured by n.m.r. for cyclohexene, uiz:'5*16 AGS = 5.3-5.4 kcal mol-'. Cyclohex-3-enylmethyl radicals were also observed in two conformations, the quasi- equatorial (9) and quasi-axial (10) which had H, h.f.s. of 28.8 and 32.3 G, respectively, at 140 K.'* This pair of radicals existed as an equimolar mixture over a wide temperature b (10) f Assuming log(A/s-') = 13.0 and Au = 0.8 G together with the usual formula for the exchange rate $ I cal = 4.18 J.constant. lo3236 E.S.R. of Cycles and Bicycles range, i.e. -AG:oo = 0 to within the experimental error. In both types of cyclo- hexenylmethyl radical the CH; group experiences steric repulsion from only one axial (or quasi-axial) hydrogen at C(5), and hence the axial and equatorial conformers are closer in energy. The lower H, h.f.s. for both (8) and (lo), as compared with axial cyclohexylmethyl radicals, are due to the same cause. The cycloheptylmethyl radical shows only one conformer even at very low temperatures. Since pseudorotation is very rapid in the cycloheptane ring,l this is presumably the fast-exchange average spectrum over all possible ring conformations. The cyclohept-4-enylmethyl radical showed the presence of two conformers, the major one (80%) having a(H& = 33.7 and the minor one (20%) having a(H,) = 42.9 G at 200 K.17 The major conformer can be identified as the equatorial chair form (lla).The H, h.f.s. of the minor radical suggests that it is an axial conformer, but there are three main candidates, i.e. [(ll b H l l d)], because both the boat and twist-boat conformations are believed to be stable conformations not greatly above the chair in energy.18. l9 Chair conformers interconvert via the boat and twist-boat conformers and the fast equilib- rium20 ensures that this radical has access to the axial boat form (llc). In this conformation the radical centre is immediately above the double bond in a very favourable position for transannular cyclisation. Reaction of the bromide precursor with ButSnH gave good yields of bicyclo[3.2.lloctane (lz).’’ Similarly, good yields of cyclised products were obtained from cyclo-oct-4-enylniethyl bromide. Medium and Large Alicyclic Rings For these larger rings n.m.r. spectroscopy and X-ray diffraction rarely give unambiguous results, and new methods, such as that provided by the CH; ‘spin probe’, are very desirable. We found that cycloalkylmethyl radicals having nine- to fifteen-membered rings (excluding cyclotetradecyl) show only two conformers which are readily distinguishable by e.s.r. spectroscopy. We label the conformer with the smaller H, h.f.s. quasi-equatorial (QE) and the conformer with the larger HB h.f.s. quasi-axial (QA). Cycloalkylmethyl radicals with ten-membered, or larger, rings had QA conformers with a(HB) values in the range 38.3-40.4 G and QE conformers with a(H,) values between 27.7 and 32.0 G at 140 K.20F.MacCorquodale and J. C. Walton Table 1. Conformer ratios for nine- to fifteen-membered rings 1 .o 1.3 (139) 1.3 1.6 (184) 0.75 0.8 (172) 1 .o 0.7 (21 1) 0.6b 0.5 (195) 9 P331 10 [2323] 11 WI 12 [3333] 13 P461 15 1 .O" 1.1 (231) [ 3 3 33 31 3237 a Statistical ratio for the indicated conformation. Statistical ratio for the [13333] conformation is 1.25. " Same ratio for other quinquangular conformers. The main cyclodecane conformer has a structure designated quadrangular [2323] in Dale's notation.21 There are thus three possible sites for the CH; 'spin probe', i.e. (13a)-(13c), neglecting conformations with this group on the 'inner edges', which will be severely crowded.Rotation of the CH; group will be most strongly impeded at the corner (13a), because of 1,3-interaction with the syn-H-atom on C(3), so we assign structure (13a) to the QA radical. Conformers (136) and (13c) are expected to give e.s.r. spectra of the QE type, and the spectrum does show some signs of two different QE radicals.20 There are eight corner and six outer-edge sites, so the statistical ratio of conformers [13a]/([l3b] + [13c]) would be 1.33, which is in reasonable agreement with the measured [QA]/[QE] ratio (see table 1). The most important conformations of cycloundecane22 and cycl~dodecane~~~ 24 are probably triangular, [335] (14) and quadrangular [3333] (15), respectively. Both these conformations should give rise to just two types of radical and, as table 1 shows, the3238 E.S.R.of Cycles and Bicycles statistical ratio of corner to outer-edge conformations is in reasonable agreement with the experimental [QA]/[QE] ratio. Five low-enthalpy conformations have been found for cyclotridecane by molecular- mechanics 22i 25 An X-ray diffraction structure of a 13-membered ring containing nitrogen showed that the main conformer had the quinquangular [ 133331 However, the e.s.r. data are in better agreement with the triangular [346] structure (table 1). Furthermore, the corner sites next to the one-bond edge in the quinquangular conformer would be expected to give rise to distinctly different spectra, but none were observed. Cyclopentadecane exists as a mixture of about five quinquangular forms21* 22 all having corner/outer-edge statistical ratios of 1 .O.This is reasonably close to the experimental [QA]/[QE] ratio (table 1). The one-bond, corner sites were expected to furnish novel corner radicals, and the e.s.r. spectrum did indeed show unusually broad lines for the QE radical. Dale26 has shown that interconversion of large-ring conformers takes place by a process of ‘corner migration’. The activation energies for this process in nine- to fifteen- membered rings were determined from the exchange broadening in the e.s.r. spectra. The e.s.r. activation energies were of similar magnitude to the free energies of activation (AGI) estimated from the I3C n.m.r. coalescence temperatures (where available). Our measurements also agreed quite well with available barriers calculated by force-field methods.20 The e.s.r.approach to conformational analysis via the CH; ‘spin probe’ gives a number of valuable insights and should prove useful for examining rings containing heteroatoms. Cyclobutylmethyl and Related Rearrangements Cycloalkylmethyl radicals with three- and four-membered rings rearrange rapidly by p- scission to give alkenyl radicals. Thus, the cyclopropylmethyl radical rearranges extremely rapidly and has been widely used as a fast radical The cyclo- Q-7r - 0 - 0 Scheme 1.F. MacCorquodale and J. C. Walton 3239 Table 2. Kinetics of @-scission in cyclobutylmethyl and related radicals radical k/s-l (25 “C) kre, Elkcal mol-I log (Als-l) (16) 4.7 x 103 I .o 12.2 12.6 (17) 2.8 x 104 6.0 11.5 12.9 (19) 3.9 x 1 0 4 8.4 12.5 13.8 cyclopropylme t hyl a 1.3 x lo8 2.8 x 104 5.9 12.5 a Data from ref.(27). butylmethyl radical (16) rearranges at a rate which is convenient for study by both e.s.r. spectroscopy28 and tri-n-butyltin hydride red~ction.~’ This system was chosen for a study of the effect of increasing resonance stabilisation in the rearranged radical. Thus, the kinetics of the rearrangements of cyclobutylmethyl (16), which gives a primary radical, 3-methylenecyclobutylmethyl(17), which gives an ally1 radical (18) and cyclobut- 2-enylmethyl (19), which gives a pentadienyl radical (20), (scheme 1) were studied by both the kinetic e.s.r. method and the tin hydride reduction method. Excellent correspondence between the two methods was obtained in each case and the rate constants and Arrhenius parameters are listed in table 2.The kinetic e.s.r. data were converted to absolute rearrangement rate constants by use of Fischer’s parameters for the combination of t-butyl radical^,^" corrected for solvent viscosity differences. The tin hydride results were converted to absolute rate constants by using the laser flash photolysis result for the rate of H-abstraction from Bu,SnH by primary alkyl radical^.^' The excellent agreement between the two types of measurement is good evidence of the reliability of these two reference rate constants. The rearrangements given in table 2 provide some of the most trustworthy free-radical clocks. The remarkable result, which the data in table 2 reveal, is that product stabilisation of over 12 kcal mol-l in the case of 2-allylallyl (18) and over 20 kcal mol-1 in the case of pentadienyl (20) has comparatively little effect on the rate and activation energy of ring fission.These experimental results suggest that scission of the CP--C, bond is not synchronised with development of resonance stabilisation in the product radical. Similar non-perfect synchronisation of the development of resonance stabilisation, with proton loss, has been observed in carbanion-forming reactions.32 The three radical re- arrangements of scheme 1 were studied using the semiempirical MNDO method,33 and the computed progress of the rearrangement was plotted in the form of a More O’Ferrall-Jencks diagram.34. 35 These diagrams for radicals (17) and (19) showed the type of curvature typical of non-perfect synchronisation, i.e.the MNDO calculations indicate that these rearrangements are extreme cases in which ring fission precedes development of resonance stabilisation. Thus, even large resonance stabilisation energies in the product radicals cause only a very small lowering of the rearrangement activation energy. Spiroalkyl and Bicycloalkyl Radicals Hydrogen abstraction from spiro[2. nlalkanes produces spiro[2. nIalk-2-yl radicals. Only the two lowest members of the series can be observed by e.s.r. spectroscopy because rearrangement by /3-scission is too rapid for the The spiro[2.2]pent-2-yl radical is a o-radical, like cyclopropyl, and it does not rearrange even at temperatures as high as 380 K. Spiro[2.3]hex-2-y1 radicals (21) can be observed at ca.150 K, but rearrange to cyclobutenylethyl radicals (22) at higher temperatures. The rate of ring fission is ca. an order of magnitude slower than in cyclopropylmethyl radicals : probably because of the creation of the strained cyclobutene ring in (22). The Ha h.f.s. in radical (21) is3240 E.S.R. of Cycles and Bicycles unusually low at 19.6 G,36 suggesting that spin density is delocalised into the Walsh orbitals of the adjacent cyclopropane ring. The orientation of the p orbital containing the unpaired electron is favourable for this type of delocalisation. Thus, the hydrogens at C(2) in spiro[2.3]hexane should be activated towards hydrogen abstraction. Experiments in which the concentration of the spiro[2. nlalk-2-yl radical, relative to that of other secondary radicals derived from the ring, was measured by e.s.r.spectroscopy indicated significant activation. A similar study of spiro[3.3]heptane (23), which would not be expected to show any activation at C(2), gave essentially equal concentrations of the two possible H-abstraction products (after statistical correction). In order to quantify the extent of this cyclopropylmethyl (cpm) stabilisation the rotation barrier about the CB-C; bond in cyclopropylmethyl radicals was examined. Exchange broadening due to restricted rotation was observed at low temperature^,^' and from the matching of experimental and computed spectra a barrier of 2.8 kcal mol-1 was obtained. This implies a small but chemically significant stabilisation in cpm radicals of ca. 2.5 kcal mol-l.Bicyclo[n . m . llalkanes can undergo H-abstraction either at the bridgehead or at the methylene groups in the bridges. The simplest member of this series, bicyclo- 1.1. llpentane (24) preferentially undergoes bridgehead abstraction to give a radical with a very large y - h . f . ~ . ~ ~ This indicates significant cross-ring orbital overlap, which may explain the unusual bridgehead reactivity. The bridgehead radical from the next member of the series, bicyclo[2.1. llhexane (25), has been observed previo~sly~~ and the large y- h.f.s. from the remaining bridgehead hydrogen suggested that H-abstraction from the bridgehead of (25) might also be favoured. We found that the only radical detectable by e.s.r. in this case was bicyclo[2.l.l]hex-2-yl derived by abstraction of one of the bridge (C,) hydrogens.However, the less selective bis(trimethylsily1)aminyl radicals did give significant bridgehead attack; see scheme 2. For bicyclo[2.2. llheptane (norbornane) (26) the bridgehead radical shows only a small h.f.s. from the remaining bridgehead hydrogen4' and, as would be expected on this basis, bridgehead hydrogen abstraction was negligible. We thank NATO for a travel grant, without which much of this work could not have been undertaken.F. MacCorquodale and J. C. Walton 324 I 69.9 100 50 22.5 0 0 2 9 2.25 + 27.7 20. 0 BU'O. 50 CI 1.3 0.7 H 0.7 H 3.9 >270' 100 Bu'O* 100 Br 71 ( Me3Si )zN' 0 3 100 Br* 97 (Me,S i IzN Scheme 2. Bicyclo[n. m . llalkyl radicals. References 1 F. A. L. Anet and R. Anet, in Dynamic Nuclear Magnetic Resonance Spectroscopy, ed.L. M. Jackman and F. A. Cotton (Academic Press, New York, 1975), chap. 14, p. 543. 2 J. Dale, Top. Stereochem., 1976, 9, 199. 3 U. Berkert and N. L. Allinger, Molecular Mechanics, ACS Monograph 177 (American Chemical Society, Washington, D.C., 1982), p. 89. 4 R. Briere, H. Lemaire and A. Rassat, Bull. SOC. Chim. Fr., 1965, 3273; R. E. Rolfe, K. D. Sales and J. H. P. Utley, J. Chem. SOC., Perkin Trans. 2, 1973, 1171; R. Chiarelli and A. Rassat, Tetrahedron, 1973, 29, 3639. 5 G. A. Russell, in Radical Ions, ed. E. T. Kaiser and L. Kevan (Wiley-Interscience, New York, 1968), chap. 3, p. 87. 6 M. L. Kemball, J. C. Walton and K. U. Ingold, J. Chem. SOC., Perkin Trans. 2, 1982, 1017. 7 K. U. Ingold and J. C. Walton, J. Am. Chem. SOC., 1985, 107, 6315. 8 K.U. Ingold and J. C. Walton, J. Chem. SOC., Perkin Trans. 2, 1986, 1337. 9 E. L. Eliel, N. L. Allinger, S. J. Angyal and G. A. Morrison, Conformational Analysis (Interscience, New York, 1967), chap. 2, p. 36. 10 G. A. Russell, G. R. Underwood and D. C. Lini, J. Am. Chem. SOC., 1967, 89, 6636. 11 J. A. Hirsch, Top. Stereochem., 1967, 1, 199. 12 H. Booth and J. R. Everett, J. Chem. Soc., Chem. Commun., 1976, 278. 13 H. Booth and J. R. Everett, J. Chem. SOC., Perkin Trans. 2, 1980, 255. 14 J. C. Walton, J. Chem. SOC., Perkin Trans. 2, 1986, 1641. 15 F. A. L. Anet and M. Z. Haq, J. Am. Chem. Soc., 1965, 87, 3147. 16 F. R. Jensen and C. H. Bushweller, J. Am. Chem. SOC., 1969, 91, 5774.3242 E.S.R. of Cycles and Bicycles 17 F. MacCorquodale and J.C. Walton, J. Chem. Soc., Chem. Commun., 1987, 1456. 18 N. L. Allinger and J. T. Sprague, J . Am. Chem. SOC., 1977, 94, 5734. 19 D. N. J. White and M. J. Bovill, J . Chem. Soc., Perkin Trans. 2, 1977, 1610. 20 K. U. Ingold and J. C. Walton, J . Am. Chem. Soc., 1987, 109, 6937. 21 J. Dale, Acta Chem. Scand., 1973, 27, 11 15. 22 F. A. L. Anet and T. N. Rawdah, J . Am. Chem. SOC., 1978, 100, 7810. 23 J. D. Dunitz and H. M. M. Sheaver, Helv. Chim. Acta, 1960, 43, 18; J. D. Dunitz, Perspect. Struct. 24 F. A. L. Anet and T. N. Rawdah, J. Am. Chem. Soc., 1978, 100, 7166. 25 B. H. Rubin, M. Williamson, M. Takeshita, F. M. Menger, F. A. L. Anet, B. Bacon and N. L. 26 J. Dale, Acta Chem. Scand., 1973, 27, 1130. 27 D. Griller and K. U. Ingold, Acc. Chem. Res., 1980, 13, 317. 28 K. U. Ingold, B. Maillard and J. C. Walton, J . Chem. Soc., Perkin Trans. 2, 1981, 970. 29 A. L. J. Beckwith and G. Moad, J . Chem. Soc., Perkin Trams. 2, 1980, 1083. 30 H. Schuh and H. Fischer, Znt. J. Chem. Kinet., 1976, 8, 341. 31 C. Chatgilialoglu, K. U. Ingold and J. C. Scaiano, J. Am. Chem. Soc., 1981, 103, 7739. 32 C. F. Bernasconi, Acc. Chem. Res., 1987, 20, 301. 33 J. J. P. Stewart, QCPE No. 455, 1983. 34 R. A. More O’Ferrall, J . Chem. SOC. B, 1970, 274. 35 D. A. Jencks and W. P. Jencks, J, Am. Chem. SOC., 1977, 99, 7948. 36 C. Roberts and J. C. Walton, J . Chem. SOC., Perkin Trans. 2, 1985, 841. 37 J. C. Walton, Magn. Reson. Chem., 1987, 25, 998. 38 B. Maillard and J. C. Walton, J. Chem. Soc., Chem. Commun., 1983, 900. 39 T. Kawamura and T. Yonezawa, J. Chem. Soc., Chem. Commun., 1976, 948. 40 T. Kawamura, M. Matsunaga and T. Yonezawa, J. Am. Chem. Soc., 1975, 97, 3234. Chem., 1968, 2, 1. Allinger, J. Am. Chem. SOC., 1984, 106, 2088. Paper 8/01525B; Received 18th April, 1988