Conformational behaviour of medium-sized rings. Part 6. 5,6,11,12,17,18-Hexahydrotribenzoa,e,icyclododecene and its 2,3,8,9,14,15- and 1,4,7,10,13,16-hexamethyl derivatives. 2,3,8,9- and 1,4,7,10-Tetramethyl-5,6,11,12-tetrahydrodibenzoa,ecyclo-octene

摘要

1398 J.C.S. Perkin I Conformational Behaviour of Nledium-sized Rings. Part 6.l 5,6,11 ,I2,-17,18-Hexahydrotribenzoa,e,icyclododecene and its 2,3,8,9,14,15-and I,4,7,10,13,16-Hexamethyl Derivatives. 2,3,8,9-and I,4,7,1 O-tetra- methyl -5,6,11 ,I2-tetrahydrodi benzoa,ecyclo-octene By David J. Brickwood, W. David Ollis,rsquo; Julia Stephanidou Stephanatou. and J. Fraser Stoddart. Department of Chemistry, The University, Sheff ield S3 7HF The temperature dependences of (i) the broad-band decoupled 13C n.m.r. spectrum of 5,6,11,12,17,18-hexahydro-tribenzoa.e,icyclododecene (1) and (ii) the lH n.m.r. spectra of its 2,3,8.9,14,15- (2) and 1.4,7,10,13.16- (3) hexamethyl derivatives have been interpreted in terms of ring inversion between enantiomeric C, conformations. Conformational analysis on these molecules has been carried out with the aid of strain energy calculations on selected conformations of the parent hydrocarbon (1) and its 1.4.7.1 0,13.16-hexamethyl derivative (3).Useful correlations between calculated and experimental thermodynamic parameters were found. The temperature dependences of the lH n.m.r. spectra of the 2.3.8.9- (1 5) and 1,4,7,10- (1 6) tetramethyl derivatives of 5,6,11,12- tetrahydrodibenzoa,ecyclo-octene(14) have been interpreted in terms of interconversion of chair- and boat-like conformations. Strain energy calculations on selected conformations of the 1.4.7.1 O-tetramethyl derivative (1 4) have led to useful correlations between calculated and experimental thermodynamic parameters. THE conformational properties of the hexahydro-tribenzocyclododecene (1) were first discussed in a prescient publication in 1945 by Baker, Banks, Lyon, and Mana2 Inspection of molecular models led these authors to consider the four conformations (la-d) shown in Figure 1 as possible forms for the molecule in the crystalline state.A preliminary X-ray study indicated that the occurrence of conformations (la) and (Id), with trigonal symmetry, in the crystal was im- probable, without providing any information concerning the relative merits of conformations (lb) and (lc). It is significant that several years before the wide recognition of conformational analysis as a stereochemical discipline J3 1 Part 5, D. J. Brickwood, A. M.Hassan, W. D. Ollis, J. S. Stephanatou, and J. F. Stoddart, preceding paper. 2 W. Baker, R. Banks, D. R. Lyon, and F. G. Mann, J. Chem. SOC.,1945, 27. 3 D. H. R. Barton, Experientia, 1950, 6, 316; Topics Stereo- chem., 1971, 6, 1; 0. Hassel, Tidsskr. Kjemi Bergvesen og MetaZ- Zurgi, 1943, 3 5, 32; Topics Stereochem., 1971, 6, 11. the ready interconversion of conformations (la)-( Id) was seen to involve only those changes in stereochemistry commonly described nowadays as torsional processes. Hexahydrotribenzocyclododecene (1) is the parent hydrocarbon of a series of twelve-membered ring systems which include the trisalicylide derivatives (4)-(6). The recognition that tri-3,6-dimethylsalicylide (4), tri-o- thymotide (5), and tri-o-carvocrotide (6) all exist in solution in diastereoisomeric propeller and helical conformations encouraged us in 1972 to examine the conformational behaviour of the hydrocarbon (1) and its hexamethyl derivatives (2) and (3) in solution.In 1945, Baker et d2reported the synthesis of the hexahydrotribenzocyclododecene (1) from o-xylylene dibromide with sodium in (i) dioxan and (ii) ether 4 W. D. Ollis and I. 0. Sutherland, Chem. Comm., 1966, 402; A. P. Downing, W. D. Ollis, and I. 0.Sutherland, ibid., 1967, 171; A. P. Downing, W. D. Ollis, I. 0. Sutherland, J. Mason, and S. F. Mason, ibid., 1968, 329; A. P. Downing, W. D. Ollis, and I. 0. Sutherland, J. Chem. SOC.(B),1970, 24. containing small amounts of ethyl acetate. In our hands, a Wurtz reaction employing the solvent condi- tions in (ii) gave a low yield of the hydrocarbon (1) (la) (lb) 1FIGURE Reproductions of the photographs of molecular models shown in ref.2 for four ' strainless phases ' (la-d) of hexahydrotribenzocyclododecene (1) ; (la) with CSnsymmetry, (lb) with C, symmetry, (lc) with C, symmetry, and (Id) with C3, symmetry together with small amounts of the acyclic hydrocarbons (7),(lo),and (13). This reaction has been the subject of several investigations 5-8 and the yield of the tribenzo- cyclododecene (1) has been improved 59698 (i) by varying the relative concentrations of the reactants 598 and (ii) by incorporating tetraphenylethylene as ~atalyst.~*~ Tetra-hydrodibenzocyclo-octene (14) was obtained as the 295-14 major product in 45 yield along with 16 of the tribenzocyclododecene (1) and small amounts of o-ditolylethane (7) when we carried out the Wurtz reaction in tetrahydrofuran in the presence of tetraphenylethylene (cf.ref. 6). The 2,3,8,9,14,15-hexamethylderivative (2) of (1) and the 2,3,8,9-tetramethyl derivative (15) of (14) were obtained together with the acyclic hydrocarbons (8) and (1 1) by a tetraphenylethylene-catalysed Wurtz reaction on the dibromide (21) in tetrahydrofuran. The dibromide (21) was synthesized by a route involving F. Vogtle and P. Neuman, Angew. Chem. Internat. Edn., 1972, 11,73; Synthesis, 1973, 85. E. Muller and G. Roscheisen, Chem. Bey., 1957, 90, 543. 7 H.A. Staab, F. Graf, and B.Junge, Tetrahedron Letters, 1966, 743. * A. C. Cope and S. W. Fenton, J.Amer. Chem. SOC.,1951, 73, 1668. @ E. W. Randall and L. E. Sutton, J. Chem. SOC.,1958, 1266. lo Part 1, R. Crossley, A. P. Downing, M. Nbgridi, A. Braga de Oliveira, W. D. Ollis, and I. 0.Sutherland, J.C.S. Perkin I, 1973, 205. l1 W. D. Ollis, J. F. Stoddart, and I. 0. Sutherland, Telra-h~dron,1974, 30, 1903. (i) a Diels-Alder cycloaddition l5 between 2,3-dimethyl- buta-l,3-diene and diethyl acetylenedicarboxylate to give diethyl 4,5-dimethylcyclohexa-l,4-diene-1,2-dicarb-oxylate (17) together with small amounts of the aromatic oxidation product (18), (ii) oxidation of (17) with 2-chloro-2-nitropropane l6 in methanolic sodium methoxide to the known l6 dimethyl diester (19), (iii) reduction of (19) with lithium aluminium hydride to the diol (20), and finally (iv) treatment of (20) with phos- phorus tribromide.The 1,4,7,10,13,16-hexamethylderivative (3) of (1)and the 1,4,7,10-tetramethyl derivative (16) of (14) were obtained together with the acyclic hydrocarbons (9) and (12) and 1,2,3,4-tetramethylbenzene (26) by a Wurtz reaction on the dibromide (25) in ether-ethyl R2 (4)R' = R2 =Me { 5 ) R' = Me, R2 = CHMe2 (6 ) R' = CHMe2,R2 = Me acetate. The synthesis of the dibromide (25) was based on known procedures 17*18 involving (i) a Diels-Alder condensation l7 between 2,5--dimethylfuran and maleic l2 D. Montecalvo, M. St.- Jacques, and R. Wasylishen, J.Amer. Chem. SOC.,1973, 95, 2023.F. Sauriol-Lord and M. St.-Jacques, Canad. J. Chem., 1975, 53, 3768. lP N. L. Allinger and J. T. Sprague, Tetrahedron, 1975, 81.21; R. R. Fraser, M. A. Raza, R. N. Renaud, and B. B. Layton, Canad. J. Chem., 1975, 53, 167. 1s V. F. Kuckerov, N. Y. Grigor'eva, and I. I. Zemskova, Zhur. obshchei Khim.. 1961, 31,447 (Chem. Abs., 1961, '55, 22173f).E. Druckrey, M. Arguelles, and H. Prinzbach, Chimia (Switz.),1966,20, 432; G. Kaupp, ibid., 1971.25, 230. 17 M. S. Newman and B. T. Lord, J. Amer. Chem. Soc., 1944, 66, 733. E. Buchta and G. Loew, Annalen, 1955,597, 123. anhydride to give the adduct (22), (ii) acid-catalysed dehydration of (22) to yield 3,6-dimethylphthalic R' R3 (13) anhydride (23), (iii) reduction l8 of (23) with lithium aluminium hydride to the diol(24), and finally (iv) treat- ment l8of (24) with phosphorus tribromide.In this paper, results of studies on the conformational 19 N. L. Allinger, M. T. Tribble. M. A. Miller, and D. H. Wertz, J. Amer. Chem. SOC.,1971, 08, 1637; for a recent review, see N. L. Allinger,Ado. Phys. Org. Ckem., 1976,18, 1. J.C.S. Perkin I behaviour of the twelve-membered ring hydrocarbon (1) and its 1,4,7,10,13,16-hexamethylderivative (3) in solution by dynamic 13C and lH n.m.r. spectroscopy are compared with the conclusions reached on the basis of strain energy calc~lations.~~~ l9 In this way, the influence on the conformational behaviour of aromatic methyl substituents in the ortho-positions with respect to the twelve-membered ring can be assessed.In the 2,3,8,- 9,14,15-hexamethyl derivative (2) the methyl sub-stituents occupy simultaneously meta- and para-positions on the benzene rings relative to the twelve-membered ring. Since it has hydrogen atoms in all six orth-positions, and since it contains good lH n.m.r. probes in the form of the aromatic protons and methyl protons, COZE t "'0 R' Me I IC0,Et 01 (17) K. (18) R'= H, R2= Me, X = C02Et (19) R1=H.R2=Me,X = COzMe (20) R' = H,R2 = Me,X = CHzOH (21 1 R' =H,R2 = Me,X =CH2BrNo(24)R' = Me,R2 = H,X = CH20H (25)R' = Me,R2 = H,X = CHzBr Me (22) Me Me0 (23) (26) the hydrocarbon (2) may be considered as an ideal model for the parent hydrocarbon (1). This means that a comparison is possible between the results obtained in solution by dynamic 13C n.m.r.spectroscopy with the parent hydrocarbon (1) and by dynamic lH n.m.r. spectroscopy with the 2,3,8,9,14,1 amp;hexamethyl deriv- ative (2). Part of this investigation has been the subject of a preliminary communication 2o and has also been discussed briefly in a recent review l1 on the confonn- ational behaviour of some medium-sized ring systems. Finally, the availability from the syntheses of (2) and (3) of the 2,3,8,9- (15) and 1,4,7,10- (16) tetramethyl deriv- atives of tetrahydrodibenzocyclo-octene (14) has prompted us to examine the conformational properties of these ' 6,8,6 ' systems in solution and compare our results with those in the literature for (14).In particular, the 2,3,8,9-tetramethyl derivative (15), with its four ortho-hydrogen atoms on the benzene rings and its good lH n.m.r. probes in the form of the aromatic protons and methyl protons, may be regarded as a suitable model for the parent hydrocarbon (14). 50 D. J. Brickwood, W. D. Oliis, and J. F. Stoddart, J.C.S. Chem. Comm., 1973, 638. 1978 EXPERIMENTAL Apart from 13C n.m.r. spectroscopy, the general methods are described in Part 3.,l Broad-band decoupled 13C n.ni.r. spectra were recorded on a JEOL SP 100 spectrometer with dichlorodideuteriomethane as ' lock ' and tetramethylsilane as internal standard. Reaction between Sodium and o-Xylylene Dibromide (with T. J. GRANT).,-Sodium wire (35.4 g) was added to a solution of o-xylylene dibromide (93.4 g) in dry ether (600 ml) and the mixture was heated under reflux for 24 h.Additions of ethyl acetate (2 ml) were made initially, and at the end of 5, 15, and 20 h. After the reaction was complete, the resulting amorphous solid was removed by filtration. Evaporation of the filtrate gave an oily residue which was separated into four fractions by fractional distillation. Fraction 1 had b.p. 50-56" at 20 mmHg, fraction 2 b.p. 80-140" at 0.1 mmHg, fraction 3 b.p. 140-210" at 0.1 mmHg, and fraction 4 b.p. 21amp;260 at 0.1 mmHg. Fraction 1crystal-lised from ethanol as needles of o-ditolylethane (7) (0.11 g, I), m.p. 65" (lit.,, 66-66"), M+ 210, 7(CDCl,) 2.90 (8 H, s, aromatic), 7.18 (4 H, s, CH,CH,), and 7.72 (6 H, s, 2 x Me).Fraction 2 was purified by distillation to give 1,2-bis-(o- methylphenethy1)benzene (10) (0.25 g, l),b.p. 210" at 0.5 mmHg Found: C, 92.1; H, 8.25; M (mass spec.), 314. CZ4H,, requires C, 91.7; H, 8.3; M, 3141, -r(CDCl,) 2.74-2.96 (12 H, m, aromatic), 7.15 (8 H, s, 2 x CH,CH,), and 7.74 (6 H, s, 2 x Me). Column chro- matography of fraction 3 on silica gel using benzene-light petroleum (b.p. 60-80") (1 : 1) as eluant yielded 5,6,11,12,- 17,18-hexahydrotribenzoa,e,icyclododecene( 1) (0.022 g,O.lyo),m.p. 189" (lit.,, 184.5"), M+312, T(CDC1,) 2.60-2.92 (12 H, m, aromatic) and 6.97 (12 H, s, 3 x CH,CH,). Fraction 4 crystallised from ethanol to yield 2,2'-bis-(o- methylphenethy1)aa'-bibenzyl (13) (0.023 g, O.lyo), m.p. 69" Found: C, 90.9; H, 7.95; M (mass spec.), 418.C32H34 requires C, 91.9; H, 8.14, M, 4181, T(CDCI,) 2.76-2.98 (16 H, m, aromatic), 7.12 (4 H, s, bibenzyl CH,CH,), 7.17 (8 H, s, other CH,CH,), and 7.76 (6 H, s, 2 X Me). The reaction was also carried out in dry tetrahydrofuran with tetraphenylethylene as catalyst.6 Column chromato- graphy on silica gel using benzene-light petroleum (b.p. 60-80") ( 15 : 1) as eluant gave 5,6,11,12-tetrahydrodi-benzoa,ecyclo-octene (14) (45), m.p. 109-1 11" (lit.,s 110-112") in addition to o-ditolylethane (7) (2) and 5,6,11,12,17,18-hexahydrotribenzoa,e,icyclododecene(1) (16).Diethyl 4,5-Dimethyl~yclohexa- 1,4-diene- 1,2-dicarboxylate ( 17).16-2,3-Dimethylbuta-l,3-diene(820 mg) and diethyl acetylenedicarboxylate (1.7 g) were heated in an autoclave at 120 "C for 6 h.T.1.c. indicated that two products were formed. They were separated by column chromatography on silica gel using light petroleum (b.p. 60-8O0)-ethyl acetate (8: 1) as eluant. Fraction 1 corresponded to the diene (17) (2.05 g, 81), M+ 252, .r(CDCl,) 5.78 (4 H, q, J 7.8 Hz, 2 x CO,CH,CH,), 7.10 (4 H, s, methylene pro- tons), 8.35 (6 H, s, 2 x Me), and 8.71 (6 H, t, J 7.8 Hz, 2 x CO,CH,CH,). Fraction 2 was characterised as di-ethyl 4,5-dimethylbenzene- 1,2-dicarboxylate ( 18) ( 190 mg, 8), M+ 250, T(CDC1,) 2.55 (2 H, s, aromatic), 5.69 (4 H, q, J 7.5 Hz, 2 x CO,CH,CH,), 7.72 (6 H, s, 2 x Me), and 8.68 (6 H, t, J 7.5 Hz, 2 x CO,CH,CH,). The same products were obtained when 2,3-dimethylbuta- 1,3-diene (820 mg) and diethyl acetylenedicarboxylate (1.7 g) were heated under reflux in ethanol (15 ml) for 24 h.Dimethyl 4,5-Dimethylbenzene- 1,2-dicarboxylate ( 19).l6-The diene diester (17) (8.0 g) was added to M-sodium methoxide (500 ml) containing 2-chloro-2-nitropropane (40 ml) and the mixture was refluxed in nitrogen for 1 h. The solution was acidified with hydrochloric acid and then excess of acid was destroyed by addition of ethereal diazo- methane. Extraction with ether yielded the crude product as an oil. Crystallisation from methanol gave the diester (19) (7.2 g, 92), m.p. 54-55' (1it.,l6 56"). 1,2-BishydroxymethyE-4,5-dimethylbenzene(20) .-The di-ester (19) (1.3 g) was refluxed for 6 h with lithium aluminium hydride (650 mg) in dry ether (40 ml).The excess of hydride was destroyed by addition of moist ether followed by water. The mixture was filtered and the residual solid was washed thoroughly with dilute hydrochloric acid and then with water. The combined filtrates were extracted with ether, and the ethereal layer was dried (Na,S04) and evaporated. The crude product was recrystallised from light petroleum (b.p. SCr80") to give the did (20) (760 mg, 88y0), m.p. 95-98" Found: C, 72.5; H, 8.8; M (mass spec.), 166. ClOH14O2 requires C, 72.3; H, 8.5; M, 1661, T(CDC1,) 2.94 (2 H, s, aromatic), 5.41 (4 H, s, 2 x CH,OH), 6.58 (2 H, br s, 2 x OH), and 7.87 (6 H, s, 2 x Me). 1,2-Bisbromomethyl-4,5-dimethylbenzene(2 1) .-Phosphor- us tribromide (4.2 ml) in dry ether (56 ml) was added drop- wise with stirring to the diol (20) (3.5 g) dissolved in dry benzene (40 ml) and dry ether (40 ml).The mixture was stirred for 2 h and then poured into water (160 ml). The organic layer was washed with 10 sodium hydrogen carbonate solution (2 x 25 ml) and then with water (2 x 40 ml), then dried (Na,S04) and evaporated. Recrystallis-ation of the residue from ethanol gave the dibromide (21) (5.6 g, 92y0), m.p. 113-116" (decomp.) Found: C, 41.1; H, 4.3; Br, 54.5; M (mass spec.), 290. CloHl,Br, requires C, 41.1; H, 4.1; Br, 54.8; M, 2901, t(CDC1,) 2.89 (2 H, s, aromatic), 5.40 (4 H, s, 2 x CH,Br), and 7.79 (6 H, s, 2 x Me). Reaction between Sodium and the Dibromide (21).-A solution of the dibromide (21) (4.5 g) in dry tetrahydrofuran (20 ml) was added with stirring during 10 h to sodium (1.5 g) and tetraphenylethylene (260 mg) in dry tetrahydrofuran (30 ml) cooled to -80 "C in acetone-solid CO,.After the reaction was complete moist ether was added followed by water. Separation of the organic layer was followed by further extraction with chloroform ; the material from the organic solutions was subjected to column chromatography on silica gel using benzene-light petroleum (b.p. 60-80") (1 : 30) as eluant to give four components. Component 1 was recrystallised from ether-ethanol to afford bis-( 2,4,5- trimethyl$dzenyl)ethane (8) (96 mg, 5), m.p. 136-137" Found: M (mass spec.), 266.203 4.C,,H,, requires M, 266.203 51, 7(CDCl,) 3.03 and 3.06 (4 H, two equal intensity singlets, aromatic), 7.24 (4 H, s, CH,CH,), and 7.72 and 7.78 (18 H, 2 x s, 6 x Me). Component 2 was recrystal- lised from chloroform-ethanol to give 2,3,8,9-tetramethyl- 5,6,11,12-tetrahydrod~benzoa,e~yclo-octene(15) (225 mg, 11), m.p. 246-248", Found: M (mass spec.), 264.187 2. C2oH24 requires M, 264.187 91, T(CDCI,-CS,), 3.31 (4 H, s, aromatic), 7.09 (8 H, s, 2 x CH,CH,), and 7.91 (12 H, s, 4 x Me). Component 3 was characterised after recrystal- lisation from ether-ethanol as lJ2-dimethyl-4,5-bis-(2,4,5-trimethyZphenethy1)benzene (1 1) (136 mg, 6), m.p. 139- 140.5" Found: M (mass spec.), 398.296 6. c3oH38 requires 21 Part 3, W. D. Ollis and J.F. Staddart, J.C.S.Pevkin I, 1976, 926. J.C.S. Perkin I M, 398.297 31, .r(CDCl,) 3.01, 3.06, and 3.09 (6 H, 3 equal Reaction between Sodium and the Dibromide (25) .-Sodium intensity singlets, aromatic), 7.22 (8H, s, 2 x CH,CH,), wire (5.35 g) was added to a solution of the dibromide (25) and 7.74, 7.77, and 7.81 (24 H, 3 x s, 8 x Me). Component (14 g) in dry ether (100 ml) and the mixture was heated 4 was recrystallised from chloroform-ethanol to afford under reflux for 24 h. Additions of ethyl acetate (2 ml) needles of 2,3,8,9,14,15-hexamethyZ-5,6,11,12,17,18-hexa-were made initially and at the end of 5, 10, and 20 h. After hydrotribenzoa,e,icyclododecene(2) (242 mg, 11yo),m.p. 310" Found: M (mass spec.), 396.280 1. C3oH3, requires M, 396.281 71, s(CDC1,) 2.93 (6 H, s, aromatic), 7.07 (12 H, s, 3 x CH,CH,), and 7.76 (18 H, s, 6 x Me).Diels-AZder Adduct (22) of 2,5-Dimethylfuran and MaZeic Anhydride.17-2,5-Dimethylfuran (500 ml) was added drop- wise with vigorous stirring during 1 h to maleic anhydride (44 g) in dry ether (100 ml). Stirring was continued for 16 h and then the crystalline product was collected. The filtrate was cooled in an ice-bath for 2 h to yield additional product. The yellow crystalline adduct (22) (54.5 g, 66) had m.p. 70-72" (lit.,17 59-63"), M+ 194, T(CD,),CO 3.59 (2 H, s, olefinic), 6.65 (2 H, s, methine), and 8.35 (6 H, s, 2 x Me). 3,6-Dimethylphthalic Anhydride (23) .17-The Diels-Alder adduct (22) (54 g) was added in portions to a rapidly stirred solution of 90 sulphuric acid (600 ml) maintained between 0 and -6 "C.Stirring was continued for 1 h at 0 "C and then the temperature was allowed to rise to +10 "C. The mixture was then poured on to ice (1.5 kg) and the pre- cipitate was collected, washed with ice-water, and dissolved in water (800 ml) containing sodium hydroxide (40 g). Upon addition of glacial acetic acid (150 ml) a small quan- tity of 2,5-dimethylbenzoic acid (m.p. 132-134"; lit.," 132-134") was obtained. Crude 3,6-dimethylphthalic anhydride was isolated from the filtrate by adding 5~- hydrochloric acid until the solution was acidic. Re-crystallisation from benzene gave the pure anhydride (23) (34.5 g, 71), m.p. 140-141" (lit.,17 142-143'). Mf 176, 7(CD3),CO 2.30 (2 H, s, aromatic) and 7.40 (6 H, s, 2 x Me).1,2-Bishydroxymethyl-3,6-dimethylbenzene(24) .le-3, 6-Di-methylphthalic anhydride (23) (30 g) was extracted (Soxh- let) into refluxing dry ether (900 ml) containing lithium aluminium hydride (15 g). Heating under reflux was continued for 28 h and then excess of hydride was destroyed by addition of moist ether followed by water. The mixture was filtered and the residual solid was washed thoroughly with dilute hydrochloric acid and then with water. The combined filtrates were extracted with ether ; the ethereal layer was dried (Na,SO,) and evaporated. The crude product was recrystaIlised from light petroleum (b.p. 60- 80 "C) to give needles of the diol (24) (25.2 g, 87y0),m.p.69-70" (lit.,18 70"), M+ 166, .r(CDCl,) 2.95 (2 H, s, aromatic), 5.35 (4 H, br s, 2 x CH,OH), 6.10 (2 H, br s, 2 x OH), and 7.68 (6 H, s, 2 x Me). 1,2-BisbromomethyZ-3,6-dimethylbenzene (25). 18-Phos- phorus tribromide (30 ml) in dry ether (400 ml) was added dropwise with stirring to the diol (24) (25 g) dissolved in benzene (250 ml) and dry ether (250 ml). The mixture was stirred for 16 h, and then poured into water (1 1). The organic layer was washed with 10 sodium hydrogen carbonate solution (2 x 150 ml) and then with water (2 x 150 ml), dried (Na,SO,), and evaporated to yield a crystalline residue. Recrystallisation from ethanol gave the dibromide (25) (36 g, 94y0), m.p. 101-102" (lit.,le loo"), M+ 290, T(CDC1,) 2.90 (2 H, s, aromatic), 5.30 (4 H, s, 2 x CH,Br), and 7.65 (6 H, s, 2 x Me).* The program numbers (1-111) established in Part 3 21 are adhered to in this paper, and together with the additional programs (IV and V) described here, these programs will form the basis of a collection for reference in future Parts of this series. the reaction was complete, the inorganic material was removed by filtration. Evaporation of the filtrate gave an oily residue which was separated by fractional distillation at 5 mmHg. Fraction 1 (0.89 g) had b.p. 64-88" and fraction 2 (1.7 g) b.p. 90-220". Fraction 1 was redistilled at 4.5 mmHg to give 1,2,3,4-tetramethylbenzene(26) (370 mg, 6), b.p. 78-80" at 4.5 mmHg (lit.,17 78.5-80.6" at 10 mmHg). Fraction 2 was subjected to column chromatography on silica gel using benzene-light petroleum (b.p.60-80 "C) (1 : 19) as eluant to give four crystalline components. Component 1 was recrystallised from light petroleum (b.p. 40-60") to give bis-(2,3,6-trimethylphenyZ)-ethane (9) (110 mg, 9), m.p. 97-100" Found: M (mass spec.), 266.203 8. CZ0H,, requires M, 266.203 41, r(CDC1,) 3.09 (4 H, s, aromatic), 7.17 (4 H, s, CH,CH,), and 7.66, 7.72, and 7.75 (18 H, 3 equal intensity singlets, 6 x Me). Component 2 was recrystallised from light petroleum (b.p. 60-80 "C) to give 1,4,7,1O-tetramethyl-5,6,11,12-tetrahydro-dibenzoa,ecycZo-octene(16) (45 mg, 4), m.p. 128-130O Found: M (mass spec.), 264.187 7. C,,H,, requires M, 264.187 91, r(CDC1,) 3.29 (4 H, s, aromatic), 6.91 (8H, s, 2 x CH,CH,), and 7.76 (12 H, s, 4 x Me).Component 3 was recrystallised from light petroleum (b.p. 60-80 "C) to give 1,2-dimethyE-3,6-bis-(2,3,6-trimethyZphenyZethyZ)ben-zene (12) (63 mg, 3), m.p. 166-168" Found: M (mass spec.), 398.296 1. C,,H,, requires M, 398.297 31 T(CDCI,) 3.20-3.24 (6 H, m, aromatic), 7.16 (8H, s, 2 x CH,CH,), and 7.73, 7.78, 7.83, and 7.88 (24 H, 4 equal intensity singlets, 8 x Me). Component 4 was recrystallised from benzene-light petroleum (b.p. 60-80 "C) to give 1,4,7,10,- 13,16-hexamethyZ-5,6,11,12,17,18-hexahydrotribenzoa,e,i-cyclododecene (3) (4 mg, lye), m.p. 290" Found: M (mass spec.), 396.280 7. C3oH3, requires M, 396.281 71, .r(CDCl,) 2.95-3.05 (6 H, m, aromatic), 6.70-7.10 (12 H, m, 3 x CH,CH,), and 7.42, 7.50, and 7.54 (18 H, 3 equal intensity signals, 6 x Me).Determination of Rates of Conformational Changes Fry Dynamic 1H and 13C N.m.r. Spectroscopy.-The methods used are fully described in Parts 1-4.10121-23 The computer programs (coded in FORTRAN IV) used to generate the theoretical line-shapes are now described for the general methods 1-111. Method I. A program (I) * for exchange of nuclei be- tween two equally or unequally populated sites, A and B, with no mutual coupling. The aromatic protons of eight- membered ring compounds (15) and (16) both gave two singlet signals of unequal intensities at low temperatures and so spectral line-shapes were simulated using this program. Calculated and observed spectra are shown in Figures 2 and 3 for compounds (15) and (16), respectively.The methyl protons of compounds (15) and (16) both gave two singlets at low temperatures as well. In the case of com- pound ( 16), spectral line-shapes were also simulated using this program. Calculated and observed spectra are shown in Figure 3. Method 11. A program (IV) f for exchange of nuclei 22 Part 2, R. P. Gellatly, W. D. Ollis, and I. 0. Sutherland,J.C.S.Perkin I, 1976, 913. 23 Part 4, D. J. Brickwood, W. D. Ollis, and J. F. Stoddart, J.C.S. Pevkin I, 1978, 1385. between three equally or unequally populated sites, A, B, and C, with no mutual coupling. The modified form of the Bloch equations introduced by McConnell 24 can be ex-tended 25 to cases involving exchange between more than two sites with no mutual coupling.For a j site problem, the general equation (1) may be written for the complex magnetic moment Gj of the jth site. The assumption is made that all spectra were recorded under steady-state conditions so that all the time derivatives of the complex magnetic moments may be set to zer0.24926 ajGj -C,+j(RljG, -RjlGj) = -irHIMoPj (1) where a, = T 2j-1 -i(aj -a) (2) o is the observing frequency * and is therefore that of the rotational frequency $ of the applied field H,, aj is the Larmor frequency $ corresponding to the jth site, T2jis the AB B A B h 1 I 2.98 9-34 1403 SITE computer program. Values for site frequencies, wj, populations, pj, and relaxation times, T2jwere obtained initially from chemical shifts, relative intensities, and half- height peak widths for the singlet signals A, B, and C in the low temperature spectra.The absorption intensity at frequency 6.1 is proportional to the imaginary part of GA + GB + Gc. The soIutions to equations (3)-(5) for a range of values for w which cover the appropriate region of the spectrum give line-shapes for selected input values of the rate constants, KAB etc. All three twelve-membered ring compounds ( 1)-(3) exhibited low temperature spectra consistent with a conformation which renders constitution- ally identical atoms and groups in these compounds di- astereotopic. Thus, the methyl protons of compound (3) give rise to three singlet signals at room temperature and below.The spectral line-shapes associated with the high temperature spectra were simulated (see Figure 4) using this 2FIGURE Observed (full line) and computed (broken line) spectra of the aromatic protons of 2,3,8,9-tetramethy1-5,6,11,12-tetra-hydrodibenzoa,ecyclo-octene (15): (a) at -52 "C, kaB 215 s-l, PA0.6, PB 0.4; (b) at -57 "C, kAB 126 s-l, PA 0.6, PB0.4; (c) at -62"c,kaB 68s-', PA0.6, PB 0.4; (d) at -67 "c,kAB 34 S-l, PA0.6, PB 0.4; (e) at -71 "c,ko 15 S-', PA 0.6, PB 0.4 transverse relaxation time of nuclei in the jth site, pj is the population of site j expressed as a mole fraction, Mo is the nuclear magnetic moment per unit volume, y is the gyro- magnetic ratio for the nucleus under investigation and hlj etc. is the rate constant for exchange from site 1 to site j' etc.For an exchange process between three sites A, B, and C with no mutual coupling, the complex simultaneous equations (3)-(5) can be derived from the general equation (1) as follows: GA(~A+ RAB + ~AC)+ GB(-~BA) + Gc(-hca) ---yiHlLw0PLk (3) GA(-~AB)+ GB(~B+ ~BA+ ~BC)+ Gc(-~cB) = -iyH~MoPB (4) GA(-~AC)+ GB(--~BC)+ Gc(ac + ~CA+ ~CB) = --yHlMoPc (5) These equations were solved for the complex magnetic moments GA, Gg, and Gc by use of a MASTER THREE *Angular frequency units (radians s-l) are used for all frequencies in these equations. t We thank Professor I. 0.Sutherland for a copy of his program. 24 H. M. McConnell, J. Chem. Phys., 1958, 28, 430. program. Compound (2) exhibited low temperature spec- tral dependence for its aromatic protons whilst the quater- nary aromatic carbons and bismethylene carbons of com-pound (1) each gave rise to three singlets at low temper- atures in its broad-band decoupled 13C n.m.r.spectrum (see Figure 5). In all cases, spectral line shapes were simulated using this program. Calculated and observed spectra are shown in Figures 6-8. Method 111. A program (V) t for exchange of nuclei ' between four equally or unequally populated sites A, B, C, and D with no mutual coupling. This computer program (MASTER FOUR SITE) was written to solve the four complex simultaneous equations which can be derived from the genera1 equation (1) in an analogous manner to that described in method 11.The problem has already been discussed by us at some length in our publications on the conformational behaviour of the trisalicylide derivatives (4)-(6). This program was used to simulate the spectral 2s C. S. Johnson, Adv. Magnetic Resonance, 1965, 1, 33. 20 L. W. Reeves, Adv. Phys. Org. Chem., 1965, 3, 187; A. Allerhand, H. S. Gutowsky, J. Jonas, and R. A. Meinzer, J. Amer. Chem. SOG.,1966,88, 3185. line-shapes of the methyl protons of compound (3) in order to examine the possibility that a small amount of a second conformation in a diastereoisomeric mixture was populating a fourth site. Calculated spectra are shown in Figure 4 beside the observed spectra and the spectra calculated using Method I1 assuming only three-site occupancy.Strain Energy Calculations .-These were carried out on FIGURE3 Observed (full line) and computed (broken line) spectra of (i) the aromatic protons and (ii) the methyl protons J.C.S. Perkin I and 1,4,7,10,13,16-(3) Hexamethyl Derivatives.-The lH n.m.r. spectrum of the hexahydrotribenzocyclo-dodecene (1) at room temperature in carbon disulphide containing a small amount of deuteriochloroform showed (see Table 1)a multiplet for the aromatic protons TABLE1 Temperature-dependent I3C n.m.r. spectral parameters (25.14 MHz) for compound (1) and IH n.m.r. spectral parameters (100 MHz) for compounds (1)-(3) Com-Temp.pound Solvent ("C) Group 6(13C)or T~ (1) CD,Cla-CS, -104 C6H4 6 140.0 (A), 139.7 (B),(2:1) 139.3(C), 130.2,126.7 6 38.2 (A), 36.8 (B),36.0 (C)6 139.9 (ABC), 130.4,126.7 6 37.4 (ABC) T 2.92(br,s) T 6.64-7.64 (m) T 2.74-3.00 (m) T 7.07 (s)(2) CDCI,-CS, T 2.82 (s) (A), 2.91 (s)(1: 1) (BC) = T 6.60-7.80 (m) T 7.74(br,s) T 3.03(s)(ABC) T 7.15 (s) T 7.80 (s)(3) CDCl, T 2.95-3.05 (m) T 6.70-7.10 (m) T 7.42(S) (A), 7.50(s) (B),7.54(s) (C) T 3.04 (s) T 6.91 (br,s) T 7.50(s) (ABC) a Sites are designated A, B, and C for the three-site systems.Sites that represent three time-averaged signals are designated ABC. b Contains a few drops of CDCl,. c The chemical shifts of the B and C protons are coincident. and a singlet for the CH,-CH, protons which became a broad unsymmetrical multiplet at -80 "C indicating the presence of a conformation devoid of three-fold sym-metry.Molecular models suggest that compounds (1)-octene (16): (a) at -29 "C, kAB 129 s-l for (i) RAB 124 s-l for (ii), PA0.12,PB 0.88; (b) at -40 "C, km 20 s-1 for (i) and (ii), PA0.12,PB0.88;(c)at -51 "C, ~AB2.0s-l for (i), k*~2.2s-l for (ii), PA 0.12,PB 0.88; (d) at -61 "C, kAB 0.36s-1 for (i), km 0.36S-' for (ii), PA 0.12,PB 0.88 compounds (I), (3), and (16) using a program * (coded in FORTRAN) based upon the procedure reported by Allinger et al.19 Details of the force field employed have been given in a recent review.11 RESULTS AND DISCUSSION The temperature-dependent n.m.r. spectra and the conformational properties of the twelve-membered ring compounds (1)-(3) are presented first, and the discus- sion of the corresponding results for the ' 6,8,6 ' systems (15) and (16) follows.The Temperature-dependent N.m.r. Spectra and the of 1,4,7,10-tetramethyl-5,6,11,12-tetrahydrodibenzoa,ecyclo-(3) could adopt (see Figure 9) two conformations in the ground state, one with C, symmetry (27a) and the other with D, symmetry (28a). These two conformations (denoted by C, and 0,)are topologically analogous to the helical (C,) and propeller (C,) conformations of the trisalicylides.4 Since the C, conformation (27a) has only one C, axis of symmetry it follows that the CH, groups of the two homotopic CH,CH, groups are di- astereotopic while the CH,CH, group which is bisected by the C, axis contains homotopic CH, groups. Thus, this conformation (27a; R1 = R2 = H) should give rise to two superimposable ABCD systems and one AA'BB' system (i.e.an unsymmetrical multiplet) for the CH,CH, protons in the low temperature lH n.m.r. spectrum of (1). The D, conformation (28a; R1= R2 = H) has three C, axes and one C, axis of symmetry which render all the CH, groups of the CH,CH, groups homotopic. This Conformational Properties of 5,6,11,12,17,18-Hexahydro-conformation (28a; R1 = R2 = H) should therefore tribemoa,e,icycZododecene (1) and its 2,3,8,9,14,15- (2) give rise to three coincident AA'BB' systems (ie. a * See footnote on p. 1403. symmetrical multiplet) for the CH,CH, protons in the lH 1978 1405 n.m.r. spectrum of (1) at low temperatures. One of ring and (ii) good lH n.m.r.probes are potentially present two conclusions can be drawn from the fact that an in the form of the methyl protons and aromatic protons. unsymmetrical multiplet is observed for the CH,CH, At +20 "C, the lH n.m.r. spectrum of compound (2) in protons at -80 "C: either (i) compound (1) exists in deuteriochloroform-carbon disulphide (1: 1) consisted solution entirely in the C, conformation (27a; R1= of three singlets for the aromatic, CH,CH,, and aryl- R2 = H), or (ii) it exists as a mixture of the C, and D, methyl protons (see Table 1). Interest was centred ABC ABC ABCD 73OC 56" C = 7*42 I 7-42 / L T 7.42 ) bsol; 7.50 7-52 7-50 7-52 7.50 7-52 FIGURE4 Observed (full line) and computed spectra of the methyl protons of 1,4,7,10,13,16-hexamethyl-5,6,11.12,17,18-hexahydro-tribenzoa,e,icyclododecene(3)using (i) program IV for exchange of protons among three equally populated sites, A, B, and C (----) and (ii) program V for exchange of protons among three equally populated sites, A, B, and C, and a fourth site D (pD0.8) coincident with C (-* ---) (in both cases the input values for all the rate constants, ~ABetc.were the same and so they will be referred to collectively as k) : (a) at +73 "C, k 106 s-l for (i), k 63 s-l for (ii); (b) at + 56 "C, k 16.9s-l for (i), k 13 s-l for (ii); (c) at + 45 "C, k 6.3 s-l for (i), k 4.8 s-l for (ii); (d) at +38 "C, k 3.7s-l for (i),k 2.9 s-l for (ii); (e) at +8 "C, k 0.01 s-l for (i), k 0.07 s-1 for (ii) conformations, (27a; R1 = R2 = H) and (28a; R1 = around the signal for the aromatic protons because they R2= H).responded to temperature changes over a shorter range In order to obtain a better understanding of the than did the signals for the CH,CH, and aryl-methyl conformational behaviour of the hexahydrotribenzo-protons. As the temperature was lowered to -90 "C cyclododecene (1), the 2,3,8,9,14,15-hexamethylderiv-the singlet for the aromatic protons separated into two ative (2) was studied as a suitable model. This com- singlets with relative intensities 1 :2 (see Figure 6). pound (2) was chosen for two reasons: (i) the methyl Over this temperature range, the signal for the aryl- groups are sufficiently remote from the twelve-membered methyl protons broadened and the CH2CH, signal ring not to influence the conformatioan behaviour of the became a broad unsymmetrical multiplet as in the low 1406 temperature lH n.m.r.spectra for compound (1). The most obvious interpretation of the low temperature lH II qe,Si(b)(b)--I04I04 'C'C cs2 I I I 1 200 150 Kx) 50 5FIGURE The broad-band decoupled 13C n.m.r. spectra of 5,6,11,12,17,18-hexahydrotribenzoa,e,icyclododecene(1) at (a) +20 "C and (b) -104 "C in CD,Cl,-CS, (2: 1); see Table 1 for chemical shift data n.m.r. spectrum of (2) is to assign the two singlets in the aromatic region to the three pairs of diastereotopic ABC I, ~2.82291 6FIGURE Observed (full line) and computed (broken line) spectra of the aromatic protons of 2,3,8,9,14,15-hexamethyl-5,6,11,12,17,18-hexahydrotribenzou,e,icyclododecene (2)using program IV for exchange of protons between three equally populated sites, A, B, and C (the input values for all the rate constants, kAB etc.were the same and so they will be referred to collectively as k) : (a) at -62 OC. k 94.1 s-I; (b) at -72 "C, k 32.2s-l; (c)at -76 "C, k 17.4 s-l; (d) at -81 "C, k 10.6 0;(e) at -90 OC, k 7.5 s-l aromatic protons in the C, conformation (27a; R1= H, R2= Me). This interpretation recognises that the chemical shifts of two of the three pairs of diastereotopic aromatic protons are apparently coincident. For this J.C.S. Perkin I reason, the aromatic protons give rise to two singlets with relative intensities 1 : 2rather than to three singlets of equal intensity.ABC ABC 6 6 7FIGURE Observed (full line) and computed (broken line) spectra of the quaternary aromatic carbons of 5,6,11,12,17,18-hexahydrotribenzoa,e,icyclododecene (1) using program IV for exchange of broad-band decoupled carbons between three equally populated sites, A, B, and C (the input values for all the rate constants, kAg elc. were the same and so they will be refer-red to collectively as k) : (a) at -82 "C, k 19.3 0;(b) at -88 "C, k 9.3s-l; (c) at -92 "C,k 5.0s-l; (d) at -104 "C, k 0.52 s-l -I------638.3 36-8 356 FIGURE8 Observed (full line) and computed (broken line) spectra of the dimethylene carbons of 5,6,11,12,17,18-hexa-hydrotribenzoa,e,icyclododecene (1) using program IV for exchange of broad-band decoupled carbons between three equally populated sites, A, B, and C (the input values for all the rate constants, kAB etc.were the same and so they will be referred to collectively as k) : (a) -82 "C, k 110 s-l; (t)at -88 "C, k 4.3 s-l; (c)at -92 "C, k 1.9 s-l; (d) at -104 C, k 0.34s-l Although the conformational interconversion and inversion processes involving the C, (27a; R1= H, R2= Me) and D, (28a; R1= H, R2= Me) conform- ations and their enantiomers (denoted by C,* and D3*) can be discussed in terms of the equilibria DRamp; C, C2* D3*, the site exchange between- sites A, B, and C involving the three pairs of diastereotopic aromatic protons (R1= HA, HB, or Hc) of the C, and C,* conformations (27a and b; R1= H, R2= Me) can only be fully understood by reference to Figures 10 and R2 R2 H R' / 3 ~ 1 R' C, conformation (270) R2 R2 R' D, conformation (28a) FIGURE The ground-state C, (27a) and B, (28a) conformations 9 of compounds (1)-(3). For (l),R1= R2= H., for (2), R1= H, R2= Me; for (3), R1= Me, R2 = H.The enantiomeric ground-state conformations are denoted by C2* (27b) and D,* (2) in the text 11. The designations HA,HB, and Hc of the three pairs of diastereotopic aromatic protons of the C, conform-ations (27a and b) in Figure 10 do not reflect the relative chemical shifts of these protons. Individual assign-ments are, of course, quite arbitrary. Despite the fact that the D, and D,* conformations (28aand b; R1= H, R2= Me) are not detectable by lH n.m.r.spectroscopy at low temperatures they must be considered as possible intermediates in the site exchange processes involving the C, and C,* conformations (27a and b; R1= H, R2= Me). Formally, the site occupied by the homo- topic aromatic protons (R1= HD) of the D, conform-ations (28a and b; R1= H, R2= Me) is designated as site D in Figures 10 and 11 even although its occupancy has not been detected experimentally. In Figure 10, each formula represents a ground state conformation on the conformational itinerary. The conformations are drawn such that the mean plane of the twelve-membered ring lies in the plane of the paper and the niethylene groups are indicated as being oriented above (0)or below (0)the mean plane.Formally at least, D, conformations (28a and b; R1= H, R2= Me) can be interconverted with C, conformations (27a and b; R1 = H, R2= Me), and C, eC,* ring inversions can occur by processes which involve simultaneous torsion about the two carbon-carbon bonds (e.g. 4a,5 and 6,6a) linking a particular dimethylene group (e.g. 44 to two aromatic rings (e.g. those defined by 1,2,3,4,4a,18a and 6a,7,8,- 9,10,10a). These processes, when executed with mole- cular models, are reminiscent of pedalling motions. Pedalling of any one of the three homotopic dimethylene bridges (1,2, or 3) of the D, conformation (28a; R1= H, R2= Me) leads to a C, conformation (27a; R1= H, R2= Me). However, since the sites of the three pairs of diastereotopic protons designated HA, HB, and Hc are different with respect to the six designated aromatic protons (H-1,4,7,10,13,16) according as to whether the C, axis bisects dimethylene bridge 1, 2, or 3, it is con- venient to identify the C, conformations (27a; R1= H R2= Me) as C2-I, C,-2, and C,-3, respectively.Pedal-ling of either of the two homotopic dimethylene bridges (2 or 3 for C2--J,I or 3 for C,-2, and 1 or 2 for C,-3) of C2-1, C,-2, or C,-3 effects C,+C,* ring inversion such that C,-1 is inverted to give either C,*-3 or C2*-2, C,-2 to give either C,*-3 or C2*-1, and C,-3 to give either C,*-2 or C2*-1. Pedalling of any of the three dimethylene bridges, I, 2, or 3 bisected by the C, axis in C,*-I, C,*-2, and C,*-3, respectively, leads to the D3* conformation (28b; R1 = H, R2= Me).The com- plete site exchange scheme for HA, HB, Hc, and HD is summarised by the cubic array diagram in Figure 11. The designation of sites within the square brackets reading from left to right correspond to H-1,4,7,10,13,16. Since the D,conformations (28a and b; R1 = H, R2= Me) are undetected intermediates in the low temperature lH n.m.r. spectra, further discussion of the conform- ational behaviour of compound (2) can be restricted to a consideration of the C, conformations (27a and b; R1= H, R2= Me). Figures 10 and 11 show that the C, == C,* inversion process is associated with a first-order rate constant k which in turn may be related to the rate constants for the aromatic protons' site exchanges HA HB, HA HC, HB *HA, HB HG.Hc +HA, Hc -HB. Thus, the rate constants for C, T-C2* ring inversion at different temperatures were determined (see Figure 6) by comparing experi- mental IH n.m.r. spectra for the aromatic protons with theoretical spectra generated by the line-shape pro-cedure described in method I1 (see Experimental section). Values for the free energy of activation for the C, === C2*ring inversion process were determined at the various temperatures. Table 2 records the average value (10.1 kcal mol-l) for AGI (C, C,*) in compound (2)* During our investigations on the conformational behaviour of compound (2), an FT 13C n.m.r. spectro-meter became available to us. Consequently, we decided to investigate the temperature dependence of the broad-band decoupled 13C n.m.r.spectrum of the parent compound (1) in order to establish that (i) the 2,3,8,9,14,- 15-hexamethyl derivative (2) is indeed a good model for Pedal 2 Pedah Pedal--1 Pedal 3 J.C.S. Perkin I pound (1) also adopts predominantly C, conformations (27a and b; R1= R2 = H) in solution. Of course, small contributions (2) from the D3 conformations (28a and b; R1= R2 = H) would not necessarily be detected by variable temperature broad-band de-coupled 13C n.m.r. spectroscopy. As before, the site Pedal 3 -1 10 Conformational itinerary and site-exchange scheme for different nuclei and groups of nuclei involving the C2(27a), C,*FIGURE (27b),D,(28a), and D,*(28b) conformations of compounds (1)-(3).For (l),R1= R2= H., for (2), R1= H, R2= Me; for (3). = a methylene group above the mean plane of the ring. 0= a methylene group below the mean plane R1 = Me, R2 = H. of the ring (1) and (ii) line-shape analysis of broad-band decoupled 13C n.m.r. spectra constitutes a reliable means of obtain-ing activation parameters for conformational changes in medium-sized ring compounds. At +20 "C, the broad- band decoupled 13C n.m.r. spectrum of compound (1) in dichlorodideuteriomethane-carbon disulphide (2 : 1) con-sisted of three singlets for the aromatic carbon atoms and one for the CH,CH, carbons (see Table 1 and Figure 5). As the temperature was lowered to -104 "C, the singlet for the CH2CH2 carbon atoms (C-5,6,11,12,17,18) separ- ated into three singlets of approximately equal intensities.The low-field singlet for the quaternary aromatic carbons (C-4a,6a,lOa,12a,16a,18a)also separated into three singlets of approximately equal intensities at -104 "C. Over this temperature range the signals for the other two sets of aromatic carbons remained as singlets. Taken together, these observations indicate that com- exchanges between sites A, B, and C involving (i) the three pairs of diastereotopic quaternary aromatic carbons and (ii) the three pairs of diastereotopic CH,CH, carbons of the C, and C2* conformations (27a and b; R1= R2= H) can only be fully appreciated by reference to Figures 10 and 11. The situation is analogous to that already discussed for the three pairs of diastereotopic aromatic protons in compound (2).Rate constants for C, +C2* ring inversion at various temperatures were determined by comparing the experimental broad-band decoupled 13C n.m.r. spectra for (i) the quaternary aromatic carbons (see Figure 7) and (ii) the CH,CH, carbons (see Figure 8) with theoretical spectra generated by the line-shape procedure described in method I1 (see Experimental section). Values for the free energy of activation for the C, == C2*ring inversion process were determined at the various temperatures. The average value for AGt (C, =+ C2*)was found (Table 2) to be (i) 9.9 kcal mol-1 derived from the quaternary aromatic carbons and (ii) 10.1 kcal mol-l derived from the CH2CH, TABLE2 Thermodynamic parameters associated with C, +C,* ring inversion in compounds ( 1)-( 3) Com-pound R1 R2 Solvent N.m.r.probe #eb AGt/kcal mol-l (1) H H CD,Cl,-CS, (2: 1) C-4a,6a,10aJ12a, C-5,6,11,12,17, 16aJ18a 18 9.9 f0.2 10.1 f0.1 (2) (3) H Me Me H CDCl,-CS, CDCI,(1:1) H-1,4,7,10,13,16 Me-1,4,7,10,13, 16 10.1 f0.3 17.4 f0.3 f a Details of l3C and 1H chemical shifts are given in Table 1. The site exchanges involving sites A, B, and C are described in Figures 10 and 11. Line-shape analysis was carried out using program IV described in method 11. e Value from 13C n.m.r. line-shape analysis (see Figure 7). d Value from 13C' n.m.r. line-shape analysis (see Figure 8). *Value from 'H n.m.r.line-shape analysis (see Figure 6). f Value from 'H n.m.r. line-shape analysis (seeFigure 4). carbon atoms. Not only is the agreement between these two values for the same conformational process in com- pound (1) obtained from use of two different 13C n.m.r. probes satisfying, but the close correspondence with the average value (10.1kcal mol-1) for AGX (C, +C2*) from from lH n.m.r. line shape analysis of compound (2) is en- couraging. Clearly, the 2,3,8,9,14,15-hexamethylderiv-ative (2) is a good model for the tribenzocyclododecene (1). Finally, application of dynamic 13C n.m.r. spectro- scopy obviously leads to results which agree with those c -3 ABCCbA CCBAAB -FIGURE Cubic array diagram showing the site exchanges for 11 different nuclei and groups of nuclei involving C, (27a).C,* (27b). D, (28a), and D,* (28b) conformations of compounds (1)-(3).-The sites indicated in square brackets read from left to right starting with the lowest and proceeding to the highest numbered nucleus or groups of nuclei in formulae (1)-(3) obtained by dynamic lH n.m.r. spectroscopy (cf. ref. 27). This observation is reassuring. t The figures quoted in Table 3 are more refined than those reported previously." The values (kcal mol-l) now gives for the total strain energies ET for the C, and D, ground-state conform- ations are in good agreement 2.36 (2.96 11) and 5.96 (5.87 ")whereas those for TS (C, ===C,*) and TS (D D2*)are different 11.60 (9.74 11) and 19.11 (15.79 ")I.In order (i) to assess the relative importance of the D, conformations (28a and b; R1 = R2 = H) which are not R' TS (D3===C2) conformation (29) bsol;R "/ bsol;R2 TS (C2e C,*, conformation (30) FIGURE12 The transition state conformations TS (C, + C,*) (29) and TS (D, C,) (30) for ring inversion and inter- conversion in compounds (1)-(3). For (l), R1= R2 = H; for (2), R1 = H, R2 = Me; for (3), R1= Me, R2 = H. The portions of the molecules indicated by thickened bonds are coplanar in each case (see footnote8 in Table 3) detectable experimentally and (ii) to investigate the nature of the transition state geometry in the C,+ C2* ring inversion process, strain energy calculations on selected conformations of the tribenzocyclododecene (1) were carried out using a molecular mechanics pro-gram (see Experimental section) based on the procedure developed by Allinger et a1.l9 The results for the ground- state conformations (Figure 9) of the C, (27a; R1 = R2 = H) and D, (28a; R1 = R2 = H) type and for the trial transition states (Figure 12) TS (C2eC,*) (29; R1 = R2 = H) to C, @C,* ring inversion and TS (0,----L C,) (30; R1 = R2 = H) to D, T-C, ring interconversion are given in Table 3.t In order for a ground-state conformation to reach a transition-state, pedalling of a dimethylene bridge is required until a conformation is attained in which the dimethylene bridge becomes coplanar with one of the aromatic rings (see footnote g in Table 3). The data for the individual interactions in each conformation show (Table 3) that the major contributions towards the total strain energies of both transition states arise mainly from angle strain and non-bonded interactions. The angle strain is aqsociated almost entirely with the twelve-membered ring CtC angles and the transannular non-bonded inter- actions are associated with the carbon and hydrogen atoms of the dimethylene groups.Both transition-state 27 B. E. Mann, Progr. NMR Spectroscopy, 1977, 11,95. conformations show considerable relaxation of non-bonded interactional strain at the expense of increased angle strain. In the case of the ground-state conform- ations, the D,conformation (28a; R1 = R2 = H) experi- ences (i) larger non-bonded transannular interactions in- volving the dimethylene groups and (ii) more angle strain in the twelve-membered ring than does the C, conform-ation (27a; R1 = R2= H).These results are consistent with indications from space-filling molecular models of closer packing in the C, conformation (27a; R1 = R2 = H) than in the D,conformation (28a; R1 = R2= H) of the two sets of three transannularly-interacting hydro-gens above and below the mean planes of the twelve- membered rings in both these conformations. The calculated strain energy difference (Aamp; 3.6 kcal mol-l) between the C, (27; R1 = R2 = H) and D,(28; R1 = R2 = H) conformations (Table 3)is consistent with the J.C.S. Perkin I the basis of symmetry considerations alone. This conclusion can be reached either by (i) noting that the symmetry number is reduced from 2 to 1 on going from the ground state to transition state conformations or (ii)appreciating that a dimethylene group can be pedalled in two energetically equivalent ways during C, C2* ring inversion.If the difference (Aamp; = 9.24 kcal mol-l) in the strain energies between the C, (27a; R1 = R2= H) and the TS (C, C2*)(29; R1 = R2 = H) conformations is equated with AH$ then the entropy difference of Rln 2 means that a free energy of activation (AG'.) of 8.8 kcal mol-l can be calculated for C, === C,* ring inversion. The relatively close correspondence between this value and the experimentally determined values (Table 2) of 9.9 and 10.1 kcal mol-l suggests that the selected TS (C, +C2*)(29; R1= R2 = H) con-formation corresponds closely to the actual transition TABLE3 Calculated strain energies (ET/kcal mol-l) of various conformations of 5,6,11,12,17,18-hexahydrotribenzoua,e,i-cyclododecene (I) Conformation E, O-C Ee a-e Ed a-d Es a-c.e Eamp; a-c,r ET C2fC2*(27; R' = R2 = H) 0.11 2.07 0.28 0.13 -0.23 2.36 D, 3 D,*(28; R' = R2= H) 0.25 4.61 0.12 0.10 0.88 5.96 TS(C2 C2*)(29; R' = R2=H) g 0.40 7.69 0.88 0.62 2.01 11.60 TS(D,* C,)( 30; R' = R2=H) 0 0.64 9.72 1.83 0.46 6.46 19.11 The following energy terms (J.F Stoddart, ' Organic Chemistry, Series One, Structure Determination in Organic Chemistry, ed. W. D. Ollis, Butterworths, London, 1973, p. 1) have been used: E, (bond length strain), Ee (angle strain), Ed (torsional strain) Ea (out-of-plane strain in aromatic rings), Ed (non-bonded interactional strain) ; total strain energy ET = E, + Ee + Ed + Es + E,*, hcalculations based upon the following force constants.Bond length strain: aromatic km 1 102, KCH 729; aliphatic Rm 663. RCH 655 kcal mol-l A-2, Angle strain : aromatic kmc 144, RWH 108; aliphatic kw 115, kcc~94, kHcH74 kcal mol-l radian-,; all angle strain reduced by a factor of 0.7. Calculations based upon the following equ,qibrium bond lengths and bond angles: amp- atic C-C 1.395, C-H 1.09; aliphatic C-C 1.54, Ar-C 1.50, C-H 1.09 A. Aromatic CCC 120deg;, CeH 120"; aliphatic CC 111.5' CCH 109.5", HCH 106". The torsional strain associated with C-C bonds was treated as a three-fold barrier of height 3.0 kcal mol-l.Aromatic C-C bond twisting was calculated according to Boyd et a2. (see footnote e). 6 Out-of-plane strain associated with aromatic rings was calculated according to Boyd el al. (R. H. Boyd, J. Chem. Phys., 1968, 49, 2574; C. Shieh, D. McNally, and R. H. Boyd,Tetrahedron, 1969, 29, 3653). Non-bonded interactions based upon the Hill equation as summarised in E. L. Eliel, N. L. Allinger,S. J..Angyal, and G. A. Morrison, ' Conformational Analysis,' Wiley-Interscience, New York, 1965, ch. 7. g Transition-state geo- metries are defined by holding a CH,CH, unit in the plane of one of the aromatic rings. For the TS (C, C,*) (29; R1= R2= H) conformation, atoms, 5, 4a, 4, 3, 2, 1, 18a, 18, 17, and 16 were ' held ' coplanar. For the TS (D,+C,) (30; R1= R2= H) conformation, atoms 12, 12a, 13, 14, 15, 16, 16a, 17, 18, and 18a were ' held ' coplanar.See Figure 12. failure to observe any D,conformational signal in the low temperature broad-band decoupled 13C n.m.r. spectra of the tribenzocyclododecene (1). Entropy differences arising from the different symmetries of the two con- formations also favour the C, conformation (27a; R1 = R2= H). This conformation has a symmetry number of 2 and consequently will be higher in entropy by Rln 3 cal deg-l mol-l than the D,conformation (28a; R1 = R2= H) which has a symmetry number of 6. If AET is equated with AH, then the entropy difference of Rln 3 means that the free energy difference (AG) between the two conformations at room temperature is 4.2kcal mol-l. This value for AG corresponds with ca.0.1 of the D, conformations (28a and b; R1 = R2= H) in equilibrium with the C, conformations (27a and b; R1 = R2 = H) in solution. Thus, it is not surprising that the D, con-formations (28a and b; R1= R2 = H) of the tribenzo- cyclododecene (1) are not detected by dynamic 13C n.m.r. spectroscopy. Compared with the C, conformations (27a and b; R1 = R2= H), the transition state con- formation TS (C, *C2*) (29; R1= R2 = H) is favoured on entropy grounds by Rln 2 cal deg-l mol-l on state for C, C,* ring inversion. This investigation demonstrates that use of molecular mechanics based on strain energy calculations is a reliable complementary procedure with medium-sized ring hydrocarbons for (i) assessing the relative importance of experimentally inaccessible ground-state conformations and (ii) defining the geometry of transition-state conformations.It was also of interest to ascertain the effect on the , conformational behaviour in solution of the twelve- membered ring of introducing six methyl substituents into the ortho-positions of the aromatic rings of the hexahydrotribenzocyclodecene (1). Accordingly, we have investigated the temperature dependence of the lH n.m.r. spectrum of the 1,4,7,10,13,16-hexamethylderiv-ative (3). At +20 "C, the spectrum in deuteriochloro- form consisted of multiplets for the aromatic and CH,CH, protons, and three singlets for the aryl-methyl protons (see Table 1). Interest was centred on the signal for the aryl-methyl protons since the three singlets observed at room temperature coalesced to give one singlet as the temperature was raised (see Figure 4) to +SO "C.Over this temperature range, the multiplets for the aromatic 1978 1411 and CH,CH, protons both coalesced to give singlets interconversion. However, in view of the results of the However, close inspection of the three singlets for the strain energy calculations, the values employed for aryl-methyl substituents at +20 "C indicates that the two these rate constants were the same as those established singlets located at lower field are slightly broader than simultaneously for C, HC2*ring inversion (see below). the high-field singlet. Two interpretations of this Good matches were obtained (see Figure 4) between observation may be considered: either (i) the three theoretical and experimental spectra provided the con- pairs of diastereotopic methyl protons (R1 = MeA, tribution from the D,conformations (28a and b; R1 = MeB, or hlec) of the C, and c2*conformations (27a and Me, R2 = H) was not allowed to exceed 0.8Yo.Rate b; R1 = Me, R2 = H) are exhibiting non-identical constants for the C, == C2*ring inversion process at relaxation times, or (ii) there is a small amount (5) various temperatures were determined and the average of the D,and D,*conformations (28a and b; R1 = Me, value for AGT (C, eC2*) was found to be 17.5 kcal R2 = H) whose homotopic methyl protons (R1 = MeD) mol-l. However, after this detailed consideration, are giving rise to a signal coincident with that of Mec interpretation (ii) discussed in the preceding paragraph in the C, conformations (27a and b; R1 = Me, R2 = H) was effectively ruled out by the observation that there and exchanging with MeA and MeB.were no significant changes in the room temperature Strain energy calculations have been performed on the line-shape for the aryl-methyl protons down to -40 "C. TABLE4 Calculated strain energies (ET/kcal mol-1) (1. of various conformations of 1,4,7,10,13,16-hexamethyl-5,6,11,12,17,18-hexahydrotribenzoa,e,icyclododecene (3) Conformation E, Ee E4 E6 Enb ET C, = C,* (27; R1= Me, R2= H) 0.35 2.71 0.76 0.24 0.05 4.11 C D,= D,*(28; R1= Me, R2= H) 0.31 2.47 1.53 0.65 0.00 4.96 TS(C, +C,*) (29; R1= Me, R2= H) a 0.87 15.29 1.43 0.76 2.44 20.79 d TS(D, +C,) (30; R1= Me, R2= H) a 0.96 12.99 1.87 0.98 5.05 21.85 See footnotes a--f in Table 3.Transition-state geometries are defined by holding a CH,CH, unit in the plane of one of the aromatic rings. For the TS (C, +C,*) (29; R1= Me, R2= H) conformation, atoms 5, 4a, 4, 3, 2, 1, 18a, 18, 17, and 16 were 'held ' coplanar. For the TS(D, C,) (30; R1= Me, R2= H) conformation, atoms 12, 12a, 13, 14, 15, 16, 16a, 17, 18, and 18a were ' held ' coplanar. See Figure 12. c The C, conformation (27a; R1= Me, R2= H) is also favoured by entropy. On the basis of symmetry considerations alone, the C, conformation (27a; R1= Me, R2= H), which has a symmetry number of 2, will be Rln 3 cal deg-l mol-l higher in entropy than the D,conformation (28a; R1= Me, R2= H) which has a symmetry number of 6.If the difference in strain energies (Aamp; 0.85 kcal mol-') between the two ground-state conformations is equated with AH, then the entropy difference of Rln 3 means that the free energy difference (AGcarc.) between the two conformations is ca. 1.5 kcal mol-I. If the difference in strain energies (AETr 16.68 kcal mol-l) between the C, conformation (27a; R1= Me, R2= H) and the tran-sition state TS (C, C2*)(29; R1= Me, R2= H) is equated with AH$ then an approximate estimate for AGtcalc.of 16.3 kcal mol-l follows from recognising that the transition state is favoured on entropy grounds by Rln 2 cal deg-l mol-l (2.e.the CH,CH, unit can be pedalled in two energetically equivalent ways during C, +C,* ring inversion). ground-state conformations (27a and b; R1 = Me, By implication, interpretation (i) must provide the R2 = H) and (28a and b; R1 = Me, R2 = H) and on the answer to the problem.Comparison (see Figure 4) probable transition states (see Figure 2) TS (C, +C2*) of the experimental lH n.m.r. spectra for the aryl- (29; R1 = Me, R2 = H) for C, C2* ring inversion methyl protons with theoretical spectra generated and TS (D,-----L C,) (30; R1 = Me, R2 = H) for D, by line-shape equations based upon a three-site exchange C, ring interconversion. The results (Table 4) agree process method I1 in the Experimental section with the conclusion that the C, conformation (27a; R1 = amongst MeA, MeB, and Mec gave a value for AGZ of Me, R2 = H) is more stable (AGcalc. = ca.1.5 kcal 17.4 kcal mol-l for C, +C2*ring inversion. There mol-l) than the D,conformation (28a; R1 = Me, R2 = is encouraging agreement between this value and H) but suggest that the latter may contribute slightly the AG'Calc. value of 16.3 kcal mol-l determined from to the conformational equilibrium. If this is the case, strain energy calculations (see Table 4) assuming that the then the fact that the calculated energy barriers transition state for C, ===C2* ring inversion corresponds AETI(,~+c,) 16.89, AET~(~~-17.74 kcal mol-l for to TS (C, C2*) (29; R1= Me, R2 = H). D, C, ring interconversion are of the same order of A comparison (Table 2) between the AGI (C, ===C2*) magnitude as the C, +C2* ring inversion barrier values for the hexahydrotribenzocyclododecene(1)and (AETI 16.68 kcal mol-l) suggests that an exchange process its 2,3,8,9,14,15- (2) and 1,4,7,10,13,16- (3) hexamethyl involving a fourth site might be occurring in the temper- derivatives shows that approximately the same barrier is ature range +20 to +SO "C.This possibility i.e. associated with C, +C2* ring inversion in compounds interpretation (ii) above was explored by generating (1)and (2) whereas a much higher barrier (plus ca. 7 kcal theoretical spectra using line-shape equations based on a mol-l) is associated with the same ring inversion process four-site exchange process (method 111 in the Experi- in compound (3). This observation leads to the conclu- mental section) amongst MeA, MeB, Mec, and MeD with sion that the presence of aryl-methyl substituents in the the chemical shifts of Mec and MeD being made coinci- meta-positions with respect to the dimethylene bridges dent.It was found that the line-shapes were insensitive does not contribute towards an increase in the energy ta the magnitude of the rate constants for D3 C, ring barrier associated with C, C2*ring inversion. Thus, 1412 the assumption that the 2,3,8,9,14,15-hexamethylderiv-ative (2) is a good model for the parent hydrocarbon (1) is vindicated. In contrast , the presence of aryl-methyl substituents in the ortho-positions with respect to the dimethylene bridges increases significantly the energy barrier associated with C,+C2* ring inversion as a result of (i) nonbonded interactions between the methyl groups and adjacent dimethylene bridges, and (ii) non- bonded interactions between juxtaposed methyl groups.By way of comparison, it is interesting that the activ- ation parameters for the conformational changes in the trisalicylides (4)-(6) , for which analogous transition states have been proposed,4 were found * to correlate in a similarly predictable manner with the varying steric demands of the alkyl substituents occupying the ortho- positions. Thus, the free energies of activation for ring inversion and interconversion processes are considerably less for tri-3,6-dimethylsalicylide (4) than they are for tri-o-thymotide (5) or tri-o-carvocrotide (6).However, whereas the hydrocarbons (1)-(3) prefer to adopt the less symmetrical C, conformation (27) in solution, both helical (C,) and propeller (C,) conformations are present in solutions of the trisalicylides (4)-(6) with the more symmetrical propeller always preferred. Finally, in this discussion on the conformational behaviour of twelve-membered ring hydrocarbons , it is worth reflecting upon the pioneering work of Baker et aL2 In fact, the C, conformation (27a; R1 = R2 = H), established by dynamic n.m.r. spectroscopy and strain energy calculations in the present investigation to be the most stable ground-state conformation, corresponds with form (lb) in Figure 1. This is (i) one of four forms (la- d) in Figure 11 proposed as a ' strainless phase ' of hexahydrotribenzocyclododecene (1) by Baker et aLz and (ii) one of the two forms (lb and c) in Figure 13 which satisfies the preliminary X-ray crystallographic evi-R (31) R =H (32)R = Br R R dence (see the introductory section) that (1) lacks trigonal symmetry, at least in the crystal.To our knowledge, the only other related investigations in more recent times are those describing the conformational behaviour of the all-cis-l,5,9-cyclododecatriene(31)28 and its lJ5,9-tribromo-derivative(32).29 The crown (33), saddle (34) , symmetrical s-trans-(35) , and unsym- metrical s-trans-(36) conformations shown in Figure 13 were all considered as possibilities by Untch et aZ.28v29for compounds (31) and (32) in solution.On the basis of (i) U.V. spectroscopy of (31), (ii) dipole moment measure- ments on (32), and (iii) the magnitudes of the vicinal coupling constants for the CH,CH, protons of (32), the 28 K. G. Untch and D. J. Martin, J. Amer. Chem. Soc., 1965,87, 3518. J.C.S. Perkin I presence of crown (33) and saddle (34) conformations in significant amounts was excluded. From examination of molecular models it was deduced that the unsym- metrical s-trans-(35) is less strained than the symmetrical /R R/bsol;R R Crown (33) Saddle (34) R / H R-HH R Unsymmetrical s-trans (36) Symmetrical s-trans (35) FIGURE13 Possible ground-state conformations crown (33).saddle (34), symmetrical s-trans (35), and unsymmetrical s-tvans (36) for cis,cis,cis-cyclododeca-1,5,9-triene(31) and its tribromo-derivative (32)proposed by Untch et aLZ8e29 For (31),R = H; for (32), R = Br