摘要:

1976 913Conformational Behaviour of Medium-sized Rings. Part 11.l HeterocyclicAnalogues of 5,6,7,12-Tetrahydrodibenzo[a,d]cyclo-octene (1.2.4.5-Di-benzocyclo-octa-1.4-diene)tBy Robert P. Gellatly, W. David Ollis,' and Ian 0.Sutherland. Department of Chemistry, The University,Sheffield S3 7HFThe temperature dependence of the n.m.r. spectra of a number of heterocyclic analogues (2) of 5.6,7,12-tetrahydro-dibenzo[a,d]cyclo-octene has been interpreted in terms of the interconversion of chair- and boat-like conforma-tions. These interpretations have been supported by strain energy calculations, and there is a useful correlationbetween the activation parameters determined by variable temperature n.m.r. spectroscopy and the results providedby these calculations. The 1 2-oxodibenz[c,f]azocine derivative (21) apparently exists preferentially in a boatconformation as a consequence of a weakly bonding interaction between N(6) and the C(12) carbonyl function.IN Part I we discussed the conformational behaviourof 5,6,11,12-tetrahydrodibenzo[a,e]cyclo-octene (1) andsome heterocyclic analogues.These compounds showedtemperature dependence of their n.m.r. spectra whichcould be associated with chair- and boat-like conform-ations undergoing inversion and interconversion a trelatively low rates on the n.m.r. time scale. Since thiswork was completed a number of papers have appearedconcerning conformational studies of the hydrocarbon(1),2 some heterocyclic analogues of (I),3 the relatedhydrocarbon cis,~is-I,5-cyclo-octadiene,~ and its syn-3,7-dibromo deri~ative.~ It is clear from these resultsthat the lJ5-cyclo-octadiene system can show interestingconformational behaviour involving a rigid chair con-formation and a mobile boat conformation.6 It isevident from the examination of models and fromgeometrical considerations that the 1 ,.l-cyclo-octadienesystem might show similar conformational behaviour.We have therefore examined some heterocyclic analogues(2) of 5 , 6,7 , 12-tetrahydrodibenzo [a& cyclo-octene usingn.m.r.line-shape methods which have been demon-strated to be a powerful technique for solving the subtleproblems posed by the conformational behaviour ofcyclic compounds in solution.The cyelic sulphides (2a),8 (2d), (Zg), and (Zi) wereprepared by the reaction of the corresponding bis(bromo-methyl) compounds (3a-d) with sodium sulphide.Bis- (2-bromomethylpheny1)methane (3a) was preparedby a published proced~re,~ with slight variations asnoted in the Experimental section.The bis(bromo-methylphenyl) ether (3b) and sulphide (3c) were pre-pared by controlled side-chain bromination of thecorresponding di-o-tolyl derivatives. The sulphones(2b),s (Zl), (Zj), and (3d) were readily prepared byoxidation of the corresponding sulphides with peraceticacid. The amines (2c), (Zf), (Zh), and (2k) were preparedby the reaction of the corresponding bis(bromomethy1)compound (3) with benzylamine in benzene. The-f Throughout this paper we have used the nomenclaturerecommended by I.U.P.A.C.rule A-21.4, as in Part I.Part I, R. Crossley, A. P. Downing, M. NbgrAdi, A. Braga deOliveira, W. D. Ollis, and I. 0. Sutherland, J.C.S. Perkin I , 1973,205.8 D. Montecalvo, M. St. Jacques, and R. Wasylishen, J . Amer.Chem. SOC., 1973, 95, 2023.A. Saunders and J. M. Sprake, J.C.S. Perkin 11, 1972, 1660.F. A. L. Anet and L. Kozerski, J . Amer. Chem. Soc., 1973,95,3407.compounds (2a-k) were formulated as shown on thebasis of the synthetic procedures used, their molecularformulae (analysis and molecular weight or high reso-lution molecular weight), and their spectral character-istics (Experimental section and Table 1).i 3 )a; X = C H 2b ; X Z Oc ; x = sd ; X = SO2e; X = CO111; x zb; X zc ; xd ; Xf ; X =g j x :h j X zi ; X =j ; X zk ; X =1 ; x =mi X =n ; X =e ; X =( 2 )CH2,Y = SCH2,Y = S O 2CH2,Y = N.CH2Ph0 , Y = s0 , Y = s o 20 , Y : N-CHzPhY = sS,Y : N - C H 2 P hSO2,Y : sY = s o 2SO2,Y z NsCH2PhC0,Y I: N.CH2PhCH2,Y = NMeC H 2 , Y = N.CMe36-Benzyl-6,7-dihydrodibenz[c,f]azocin-l2(5H)-one (21)was prepared in rather low overall yield from 2,2'-dimethylbenzophenone by bromination to give thebis(bromomethy1) ketone (3e) lo which reacted withbenzylamine to give the amine (21).The azocinestructure (21) for this product is consistent with themolecular formula (analysis and molecular weight) andits spectral characteristics.EXPERIMENTALUnless otherwise stated i.r. spectra were determined fordispersions in KBr using a Perkin-Elmer 137 spectrometerR. K.Mackenzie, D. D. MacNicol, H. H. Mills, R. A.Raphael, F. B. Wilson, and J. A. Zabkiewicz, J.C.S. Perkin 11,1972, 1632.a J. D. Dunitz and J. Waser, J . Amer. Chem. Soc.. 1972, 94.6645.For reviews, see G. Binsch, Topics Stereochem., 1968, 3, 97;I. 0. Sutherland, Ann. Reports N.M.R. Spectroscopy, 1971, 4, 71.8 D. D. Emrick and W. E. Truce, J . Org. Chem., 1961,26,1329.E. D. Bergmann and 2. Pelchowicz, J . Amer. Chcm. Sot.,1963, 75, 4281.10 M. P. Cavaand J. A. Kuczkowski, J . Amer. Chem. SOC., 1970,92, 6800914 J.C.S. Perkin Iand U.V. spectra for ethanolic solutions using a Perkin-Elmer 137 UV spectrometer. N.m.r. spectra were measuredfor solutions in deuteriochloroform using a Varian HA-100spectrometer.Low resolution mass spectra were deter-mined with an A.E.I. MS12 spectrometer and high reso-lution spectra with an A.E.I. MS9 spectrometer. M.p.swere measured with a Reichert hot-stage apparatus.Microanalyses were determined by the University ofSheffield Microanalytical Service.- Preparative t.1.c. was carried out using Merck silicagel G. SoIutions were dried over anhydrous magnesiumsulphate and were evaporated under diminished pressureusing a rotary evaporator.Bis-( 2-bromonzethyZpheny2)lnetliane (3a) .-The method ofBergmann and Pelchowicz was used with the followingmodifications. (a) The methylation of 2,2'-methylene-dibenzoic acid was carried out using diazomethane ratherthan sulphuric acid-methanol; and (b) the bromination of2,2'-methylenedibenzyl alcohol was carried out using boilinghydrobromic acid for 20 min, cooling, and collecting theprecipitated product.The diphenylmethane derivative(3a) had m.p. 93-93" (lit.,@ 93-94') (Found: C, 51.0; H,3.9; Br, 45.25. Calc. for C,,H,,Br,: C, 50.85; H, 4.0;Br, 45.2%).7,12-Dihydro-5H-dibenzo[c,fJthiocin (2a) .-The method ofEmrick and Truce * was used ; the product crystallised frombenzene-ethanol (1 : 1); m.p. 193-194" (lit.,8 194.5")(Found: C, 79.9; H, 6.2%; M', 226. Calc. for C,,H,,S:C, 79.6; H, 6.2%; M , 226).7, l2-Di~~d~o-5H-dibenzo[cJf3thiocin SS-Dioxide (2b) .-The method of Emrick and Truce was used; the productcrystallised from benzene-toluene (4 : 1) ; m.p. 259-260"(lit.,8 258-259") (Found: C, 69.7; K, 5.3; S, 12.55%;M+, 258.Calc. for Cl5Hl4O2S: C, 69.8; H, 5.4; S, 12.4% ;M , 258).6-Benzyl-5,6,7,12-tetra~~ydvodibenz[c,f]uzocine (2c) .-Themethod of Pala et al.I1 was used; the product crystallisedfrom ethanol; m.p. 110-112' (lit.,,, 113-114") (Found:C, 88.1; H, 6.9; N, 4.8%; M+, 299. Calc. for C,,H,,N:C, 88.3; H, 7.0; N, 4.7%; M , 299).5H,7H-Dibenz[b,g][l,5]oxutlziocin (2d) .-A mixture ofbis-(2-bromomethylphenyl) ether l2 (3b) (0.5 g) and sodiumsulphide nonahydrate (1.5 g) in 95% aqueous methanol(80 ml) was heated under reflux with stirring for 46 h.The mixture was evaporated and the residual gum extractedwith hot ethanol. The extract was cooled giving theox~thiocin (2d) (76 mg, 25y0), m.p. 75-76' (Found: C, 73.6;H, 5.1; S, 14.1%; M+, 228.C,,H,,OS requires C, 73.7;H, 5.3; S, 14.0%; M , 228).5H,7H-Dibem[bJg] [ 1,5]oxuthiocin SS-Dioxide (2e) .-Amixture of the oxathiocin (2d) (96 mg) and hydrogenperoxide (0.5 g ; 300/,) in acetic acid (1 ml) was heated underreflux for 1 h. The solution was allowed to cool and theprecipitate collected giving the sulphone (2e), whichcrystallised from ethanol; yield 41 mg (37y0), m.p. 152-154" (Found: C, 64.4; H, 4.4; S, 12.6%; M+, 260.C,,H,,O,S requires C, 64.6; H, 4.6; S, 12.3%; M , 260);hmax 220 nm (E 24 800).6-Benzyt-6, 7-dihyd~o-5H-cZibenz[b,g] [ 1 , 5]oxuzocine (2f) .-A solution of benzylamine (0.9 g) in benzene (10 ml) wasadded dropwise over 30 min to a solution of bis-(2-bromo-methylphenyl) ether (3b) (1.0 g) in boiling benzene (10 ml).Heating was continued for a further 2.5 h, then the mixturel1 G.Pala, A. Montegani, and E. Zugna, Tetmhedvon, 1970, 26,1276.was cooled and filtered and the filtrate evaporated. Theresidual solid crystallised from ethanol giving the oxuzocine(2f) (86 mg, lo%), m.p. 137-139" (Found: C, 83.8; H,6.5; N, 4.65%; M+, 301. C2,H,,N0 requires C, 83.7; H,6.3; N, 4.65%; M , 301).Bis-( 2-bvomomethylfihenyl) Sulfihide (3c) .-A stirred mix-ture of di-o-tolyl sulphide ,, (12.5 g) , N-bromosuccinimide(21.0 g), and dibenzoyl peroxide (0.1 g) in carbon tetra-chloride (60 ml) was irradiated for 6 h with a tungsten lamp.The mixture was filtered, and the filtrate was evaporatedgiving a yellow gum which was repeatedly extracted withhot light petroleum (b.p.40-60"). The extract wasconcentrated and cooled giving the sulphide (3c) (7.18 g ,33y0), m.p. 68-70". A sample recrystallised from lightpetroleum (b.p. 40-60") had m.p. 69-70' (Found: C,44.9; H, 3.45; Br, 42.9; S, 8.5. C,,H,,Br,S requires C,45.2; H, 3.2; Br, 43.0; S, 8.6%); T (CDC1,) 5.25 (s,2 x CH,Br).5HJ7H-Dibenzo[b,g][1,5]ditkiocin (2g) .-A mixture ofbis-(2-bromomethylphenyl) sulphide (3c) (0.5 g) andsodium sulphide nonahydrate (1.5 g) in 95% aqueousmethanol (80 ml) was heated under reflux for 24 h withstirring. The mixture was evaporated and the residuedistributed between benzene and water. The benzenelayer was dried and evaporated. The residual solidcrystallised from ethanol giving the difhiocipr (2 g ) (156 mg,47y0), m.p.128-130' (Found: C, 68.7; H, 5.3; S, 26.6%;M+, 244.0380. C14H12S2 requires C, 68.9; H, 4.9; S,26.2% ; M , 244.0384).6-BenzyE-6,7-dihyd~o-5H-dibenzo[b,g] [ 1,5]thiazocine (211).-A solution of benzylamine (0.47 g) in benzene (5 ml) wasadded dropwise over 1.5 h to a solution of bis-(Zbromo-methylphenyl) sulphide (3c) (0.5 g) in boiling benzene(6 ml). After heating for a further 1 h, the mixture wasfiltered and the filtrate evaporated. The residue waspurified by t.1.c. with benzene as solvent giving the thiazocine(2h) (76 mg, 18%), which crystallised from 95% ethanol;m.p. 91-92" (Found: M+, 317.1249. CzlHlsNS requiresM , 317.1238).Bis-( 2-bromomethyZphenyZ) Sulplione (3d) .-A mixture ofbis-(2-bromomethylphenyl) sulphide (3c) (0.5 g) andhydrogen peroxide (1.14 g ; 30%) in acetic acid (9 ml) washeated under reflux for 3 h.The mixture was evaporatedand the residue purified by t.1.c. with benzene as solvent.The sulphone (3d) (106 mg, 20%) crystallised from ethanol;m.p. 128-131" (Found: C, 41.5; H, 3.1; Br, 39.6; S,8.1%; M+, 404. C,,Hl,Br,O,S requires C, 41.6; H, 3.0;Br, 39.6; S, 7.9%; M , 404).5HJ7H-Dibenzo[b,g][1,5]dithiocin 12,12-Dioxide (2i).---Amixture of bis-( 2-bromomethylphenyl) sulphone (3d) (0.5 g)and sodium sulphide nonahydrate (1.38 g) in 95% methanol(80 ml) was heated under reflux with stirring for 40 h. Themixture was evaporated and the residue distributedbetween chloroform and water. The chloroform layer wasdried and evaporated. The residual gum was purified byt.1.c. with chloroform as solvent, giving the dithiocin (Zi),which erystallised from ethanol; m.p.160-166" (134 mg,39%) (Found: M+, 276.0277. C,,H,,O,S, requires M ,276.0279) ; A-. 250 nm (E 7 300).5H,'IH-Dibelzlo[b,g][ 1,5]dithiocin 6,6,12,12-Tetraoxide(2j),-A mixture of the dithiocin dioxide (2i) (60 mg) andhydrogen peroxide (0.5 ml; 30%) in acetic acid (2 ml) wasl2 R. Shapiro and D. Slobodin, J. Org. Chem., 1969, 34, 1165.l8 M. Balasubramanian and V. Baliah, J. Ckem. SOC., 1956,1261; F. Mauther, Ber., 1906, 39, 36931976 915heated under reflux for 26 h. The mixture was cooled andthe precipitate collected, giving the disulpholze (2j), whichcrystallised from ethanol; yield 29 mg (38%), m.p. 293'(Found: C, 54.6; H, 4.05; S, 20.8%; M'-, 308.C,,Hl,04S, requires C, 54.5; H, 3.9; S, 20.8%; 144, 308);6-BenzyZ-6,7-dihydro-5H-dibenzo[b,g][1,5]thiazocine SS-Dioxide (2k).-A solution of benzylamine (0.44 g) inbenzene (5 ml) was added dropwise over 2 h to a solution ofbis-( 2-bromomethylphenyl) sulphone (3d) (0.5 g) in boilingbenzene (10 ml).After heating for a further 30 min, themixture was filtered, and the filtrate evaporated giving agum which was purified by t.1.c. with benzene as solvent.The lhiazocine dioxide (2k) crystallised from benzene ; yield07 mg (22%), m.p. 164-166" (Found: Mf, 349.1141.C2,H,,02NS requires M , 349.1137); A,,, 245 nm (E15 200).6-BenzyZ-6,7-dihydrodiben~[c,fla,-ocin-l2(5H)-one (21) .-Aniixture of 2,2'-dimethylbenzophenone (0,5 g), AT-bromo-succinimide (1.0 g), and dibenzoyl peroxide (0.01 g) incarbon tetrachloride (40 inl) was irradiated for 2 h at roomtemperature using a tungsten lamp.The mixture wasfiltered and the stirred filtrate treated dropwise at 0 "C witha solution of benzylamine (0.72 g) in carbon tetrachloride (51x11). The mixture was then filtered, and the filtrate evapor-ated giving a gum (0.97 g) which was separated by t.1.c.with chloroform as solvent. The major component crystal-lised from benzene giving the azocine (21), m.p. 106-107"(29mg, 4%) (Found: C, 84.4; H, 6.2; N, 4.3%; M+, 313.C,,H,,NO requires C, 84.35; H, 6.1; N, 4.5%; M , 313);A,,, 221 (E 24 600) and 265 nm (1 300).Deternzination of Exchange Rates by N.m.r. Spect~oscofiy.~-The methods used were fully described in Part 1.l Ingeneral, exchange rates were determined a t a single tem-perature only (Table 2) so that only free energies of activ-ation could be calculated. The computer programs (codedin FORTRAN IV) used to generate theoretical line-shapesare described for four general methods (1-IV).A program * was used that was suitable forthe calculation of n.m.r.line-shapes for two AB systems(9 1,Bl and A2,BZ) undergoing exchange of hydrogen nucleibetween the pairs of sites designated A2 and B2, A1 and-\2, and El and B2. In all cases the exchange rate (k,)between the sites A2 and B2 was fast compared with theexchange rates (k12 and hzl) between the sites A1 and A2and B1 and B2.7 Thus nuclei in the sites A2 and B2 giverise to a single line (ABZ) at the average site chemical shifta t temperatures where the sites A1 and B1 give rise to atypical four-line A B system [for an example see Figurel(a)].This situation was observed in the spectra ofcompounds (2d, f , and h) a t low temperatures (down to-- 110 "C) and in the spectra of compounds (2e and g) a tintermediate temperatures (e.g. -52 and $20 O C , re-spectively). In these cases the input parameters forspectrum simulation (chemical shifts, coupling constant,.Il, and population, pl) for the system A1,Bl were readilyobtained from the low temperature spectra, but for thesystem ,12,B2 the only critical input parameters were thepopulation (p,) and average chemical shift provided thatthe exchange rate k , was large (e.g.> lo4 s-l). The inputrelaxation times were based upon low temperature spec-A2B2 would give spectral line-shapes identical withA2 and B1,- B2.220 (E 18 300) and 262 nm (5 000).Method I.* We thank Dr. W. Deloughry for writing this program. t In cases in which k , % k,, the site exchanges B1and A1those based upon A1tral line-widths in the usual way.1~~ Agreement betweenobserved and calculated spectral line-shapes could usually beobtained quite readily [e.g. Figure 1, (a) and (b)], but in twocases, compounds (2i and k), the population, p,, of the sitesA2 and 3 2 was so low that the singlet espected from thesesites at temperatures where K, CQ but h,, _t 0 couldnot be distinguished in the experimental spectra (p2<ca.0.05). In these two cases it was assumed that theaverage chemical shift for the sites A2 and B2 was thatfound for analogous compounds (see Table l), and withk , 00 the populations p1 and p z were adjusted to givea good match between observed and computed spectralline-shapes. In general the spectra observed for these twocompounds, (2i and k), as k,, increased were similar to thoseof a single AB system based upon the sites ,41 and B1, but15 5 5.75 6-25 6.5FIGURE 1 Observed (full line) and computed (broken line)spectra of the C-5 and C-7 methyene protons of the oxathiocinderivative (2d) : (a) at -67" C , (b) a t -26.5 "C; k,, = 5.5 s-1PI = 0.69, 9, = 0.41 (sites are labelled in this and later figuresin accord with the labelling system used in Tables 1 and 2)the low-field (Al) doublet was additionally broadened ascompared with the high-field (Bl) doublet, which wascloser to the averaged chemical shift of A2 and B2 (seeFigure 2 for an illustration). The exchange rates K,, andk,, are related by P1kl2 = P2k2, and for p , <PI a similarline-shape could be obtained either by the method outlinedabove or by treating the spectrum as a single coalescingAB system (A1 and B1) with the exchange rate K = 0.5 K I 2 .However, this simplified treatment fails, for the reasonsdescribed above, as p , increases.This result is of generalinterest in that i t demonstrates that the widely usedtreatment of spectral line-shapes for inverting ring systems 7which ignores the presence of intermediate conformationswith a low population, for example, the twist-boat conform-ations of cyclohexane, is in accord with the above com-putational results.The n.m.r.spectra of the thiocin derivative (2a) weresimulated in a rather similar manner, but in this case thelow intensity signals (AB2 and CD2) of the boat conform-ation were just detectable (Figure 3) and spectral simulationwas based upon both pairs of AB systems. Thus thesystems A1,Bl and A2,B2 (Figure 3 and Table 1) shown bythe thiocin derivative (2a) correspond to its C-5 and C-7methylene group protons and the systems C1,Dl andC2,D2 correspond to its C-12 methylene protons.J. W. Cook, J . Chem. SOC., 1930, 1091916 J.C.S. Perkin IThe n.m.r. spectra of the azocine derivative (2c) atlow temperatures also showed signals associated with theA ........ BI b)FIGURE 2 Observed (full line) and computed (broken line)spectra of the C-6 and C-7 methylene protons of the dithiocinderivative (2i): (a) at 36 "C, (b) at 86 "C; the broken linecorresponds to the line-shape simulation by method I withk,, = 6.5 s-1, p 1 = 0.96, p a = 0.06, and the dotted line to theline-shape sirnulabon by method I11 with k = 13 s-1chair conformation (A1,Bl from the C-6 and C-7 methyleneprotons and C1,Dl from the C-12 methylene protons) and arapidly inverting boat conformation (AB2 from the C-5and C-7 methylene protons and CD2 from the C-12methylene protons). Computed spectra were based uponC-5 and C-7 (A1,Bl) methylene group protons. Theprogram PLOTTER I1 Is was used to generate theoreticalline-shapes.This program sums the spectra of two ABsystems, A1,Bl and Cl,Dl, with relative summed inten-sities 1.0 and 0.5 respectively and with exchange betweenthe pairs of sites A1 and B1 and C1 and D1 with the samerate, k . This program could also be used to simulate thelow temperature spectra of the compounds (2e and g)which both showed two AB systems of different intensitiesat low temperatures. The low intensity AB system(A2,B2) in both cases showed evidence, as the temperatureA1 81I I I5-50 575 64 6-25 't:FIGURE 3 Observed (full line) and computed (broken line)spectra of the C-5, C-7, and C-12 methylene protons of thethiocin derivative (2a) at (a) -31 "C and (b) +20 "C; R i a =9.31 s-l, 00 (method I), p1 = 0.91, p2 = 0,09; theupper computed spectrum shows the separate signals from thesystems A l , B l , and C1,Dl and the lower the summedspectrum of these two systemswas raised, of mutual site exchange with a rate k, and gavethe typical line-shapes 7 of a coalescing AB system.The6.0 6 *25 65 6-75 ZFIGURE 4 Observed (full line) and computed (broken line) spectra of the C-6 and C-7 methylene protons of the oxathiocinderivative (2e) at -88 "C, with k,, = k2, = 0 s-l and k2 = 6.0 s-l, p1 = 0.73, $2 = 0.27these two pairs of AB systems as in the case of (2a). Inaddition, the N-benzyl methylene group of this azocinederivative (2c) gave two singlet signals corresponding tothe chair and boat conformations; line-shapes in this casewere simulated by method IV.Compound (2b) gave, a t low temperatures,two superimposed AB systems from the C-12 (C1,Dl) andhigh intensity AB system (A1,Bl) remained unchanged inthis temperature range (kiJ k,,, and k2, ---b 0) and did notcontribute to the line-shape changes of the system A2,B2.The use of this program is illustrated in Figures 4 and 5,which show the appearance of the observed and calculatedspectra for compounds (2e and g) with k 2 > 0 and k1, = Method 11.16 J.R. Fletcher, Ph.D. Thesis, Sheffield, 19701976 917K,, = 0 and also the appearance of the spectrum ask , + 00 and k,, > 0 [Figure 5(b)].The disulphone (2j) gave only a single ABsystem at low temperature, and spectral line-shapes wereMethod 111.B1I I tI It I5.0 5-25 6.25 6.5 ZFIGURE 5 Observed (full line) and computed (broken line)spectra of the C-6 and C-7 methylene protons of the dithiocinderivative (2g): (a) a t -70 "C, with RII = A,, = 0, Kt = 8.05s-l; (b) a t +55 "C with k , 00, klP = 20.26 s-1, p1 = 0.81,p , = 0.19simulated using a program16 PLOTTER I suitable forsimulating the spectral line-shapes of a single AB systemundergoing mutual site-exchange.The N-benzyl methylene group of com-pounds (2c, f, and h) gave two singlet signals of unequalintensities at low temperatures and spectral line-shapes weresimulated using a program 16 NMREX 2 SITE suitable forexchange between two unequally populated sites with nomutual coupling. Exchange rates obtained using thisprogram agreed well with those obtained from the line-shapes of the spectra of the C-5 and C-7 methylene groupprotons, confirming the general validity of the approachMethod IV.AA15I6 72Af\I I6.0 6.25 6.5 675YFIGURE 6 Observed (full line) and computed (broken line)spectra of the N-benzyl and C-5 and C-7 methylene protons ofthe thiazocine derivative (2h): (a) at -46 O C ; (b) a t -14 "Cwith A,, = 9.28 s-1, k , -, 00 (method I) and km = 9.28 s-1, PA = 0.76, #B = 0.26 (method IV)using method I.Calculated and observed spectra areshown in Figures 6 and 7 for compounds (2f and h).Styrain Energy CaZcuZations.-These were carried outl6 I. R. Gault, Ph.D. Thesis, Sheffield, 1970.l7 (a) N. L. Allinger, M. T. Tribble, M. A. Miller, and D.H.Wertz, J. Amer. Chem. SOC., 1971, 08, 1637; (b) E. M. Engler,J. D. Andose, and P. von R. Schleyer, ibid., 1973, 95, 8003.using a FORTRAN program based upon the procedurereported by Allinger et aZ.170 The program proved reason-ably efficient, and the iterative procedure was continueduntil the root mean square atomic motion between successiveiterations was consistently less than 0.001 A in each co-ordinate. Free movement of all co-ordinates was allowedfor minimum energy conformations, but for transitionstates appropriate atomic co-ordinates remained fixedthroughout the minimisation procedure. Generally energyminimisation required 10-16 min computing time with anICL 1907 computer. Many of the conformations examinedinvolved large angle deformations, and it was clear fromthe results that force constants obtained from valence forcefield treatments for angle deformation are too large whenlarge deformations are considered.All bending forceconstants were therefore reduced empirically by multiplic-ation by 0.7 (cf. ref. 17). This modification of bendingforce constants has subsequently been found unnecessaryfor cases involving smaller angle deformation, and thisagrees quite well with a modified expression for angle straindeformation which includes both quadratic and cubicterms.17b This difficulty in the correct calculation of strainenergy resulting from angle deformation is a seriouslimitation of the calculations described in this paper.A61nAB2I t I I I5.25 6.5 6.25 6.5 6-75 ZFIGURE 7 Observed (full line) and computed (broken line)spectrum of the N-benzyl and C-5 and C-7 methylene protonsof the oxazocine derivative (2f) at -9 "C; kIa = 3.80 s-1,k,--+ 00 (method I) and RAB = 3.76 s-l, PA = 0.67, PB =0.33However, we feel that, although our results may lackquantitative significance, they are nevertheless of con-siderable help in understanding complex conformationalchanges.18RESULTS AND DISCUSSIONThe n.m.r.spectra of the heterocyclic analogues(2a-k) of 6,6,7,12-t etrah ydrodibenzo [a,4 c yclo-octene inmost cases showed temperature dependence associatedwith the presence of two conformational species inequilibrium, together with slow conformational in-version of one or both species. The spectral changesobserved for the aromatic protons were complex and willnot be discussed.The changes associated with thesignals of the C-5 and C-7 methylene protons, and whererelevant those of the C-12 and the N-benzyl methyleneprotons, are summarised in Tables 1 and 2. Table 1summarises the chemical shifts and coupling constantsof the high- and the low-temperature spectra. Table 2gives details of site exchanges affecting the signal line-shapes, with associated rate constants and activationparameters. These are derived by comparison ofobserved and computed spectral line-shapes (see Figures1-7) using methods I-IV (see Experimental section).W. D. Ollis, J. F. Stoddart, and I. 0. Sutherland, Tetra-hedron, 1974, 80, 1903918 J.C.S. Perkin ITABLE 1Temperature-dependent spectral parameters ( 100 MHz) for compounds (2a-1)YSSO,NCH,PhSSO,N*CH,PhSNCH,PhSNCH,PhNCH,Phso,Temp.Solvent ("C)CDCl, - 31C,D,N 095CDC1, - 30CDCl,CDCl,-CS,(2: 1)CDCl,CS,CDCl,CDCl,CDCl,CDCl,CDC1,CDCl,- 5740- 88-3 - 2940- 8520- 468036363636GroupArCH,*SArCH,.SArCH,ArArCH,ArArCH,*S 0,ArCH,ArArCH,*SO,ArCH,ArArCH,.NArCH,*NArCH,ArN*CH,PhN*CH,PhArCH,*SArCH,*S&CH,-SArCH,*SO,ArCH,SO,ArCH,.NN.CH,PhArCH,*NN*CH,PhArCH,*SArCH,*SArCH,-NN*CH,PhArCH,NN*CH,PhArCH,*SArCH,.SO,ArCH,-NN-CH,PhArCH,*NN-CH,PhN.m.r. parameters[T (coupling constant)]5.58 (Al), 6.16 (Bl) (J 15.0 Hz)6.39 (s) (AB2)5.76 (Cl), 6.22 (Dl) (113.1 Hz)5.72 (s) (CD2)4.09 (Al), 5.26 (Bl)(J 14.6 Hz)6.27 (Cl), 6.97 (Dl) (J 13.0 Hz)6.16 (s) (ABl)6.76 (s) (CD1)5.36 (Al), 5.98(Bl) (J 15.0 Hz)6.26 (s) (AB2)6.50 (Cl), 6.28 (Dl) (J 12.9 Hz)5.66 (s) ( 0 2 )6.80 (s) (A)6.19 (s) (B)6.64 (Al), 6.32 (Bl) (J 14.7 Hz)6.49 (s) (AB2)6.27br (s) (AB12)4.88 (Al), 5.83 (Bl) (J 14.6 Hz),6.68br (s) (AB12)6.34 (Al), 6.17 (Bl) (J 14.7 Hz), 6.43 (s) (AB2)6.84 (s) (A), 6.50 (s) (B)6.2br (s) (AB12)6.70 (s) (AB)6.11 (Al), 6.36 (Bl) (J 14.2 Hz),6.17 (A2), 6.70 (B2) (J 13.7 Hz)6.06 (Al), 6.32 (Bl) (J 14.2 Hz), 6.36 (s) (ABZ)4.96 (Al), 6.06 (Bl) (J 14.6 Hz), 6.29 (s) (AB2)6.78 (s) (A), 6.49 (s) (B)6.66 (s) (AB12)6.67 (s) (AB)4.46 (Al), 6.25 (Bl) (J 14.9 Hz) [6.6O(AB2)] b3.77 (Al), 6.89(Bl) (J 14.5 Hz)4.46 (Al), 6.03 (Bl) ( J 16.0 Hz) E6.10 (AB2)] b6.83 (s)6.26 (s) (AB2) c6.39 (s)5.60 (A2), 6.19 (B2) (J 13.9 Hz)The designations Al, B1, etc.correspond t o the site exchanges cited in Table 2. The designations ABI, AB2 refer t o coalescedThe designation AB12 refers t o the single coalesced signal from the ABb The square brackets indicatee This signal issignals from the AB systems A1,Bl or A2,B2 respectively.systems A1,Bl and A2,B2.that this signal was inferred from line-shape calculations but was not visible in the spectrum above the noise level.designated AB2 since it may be assigned t o a rapidly inverting boat conformation.The designation AB refers to the coalesced singlet signals A and B.An examination of molecular models, supported bystrain energy calculations (see Experimental section andTables 3 and 4), shows that the compounds (2) canadopt two types of conformation, both of which arealmost free from angle strain. The first of these is therigid chair-like conformation (4) which has C, symmetry.The C-5 and C-7 methylene protons of this conformation,H1 and H2 [see structures (4)] are exchanged between thesites A1 and B1 by the confarmational inversion (4a) I-(4b); it is therefore necessary when discussing then.m.r. results to consider these two identical conform-ations (4a and b) separately.They are designated Cand C* in the discussion that follows. The conform-ations of the compounds (21, and the relationshipsbetween conformations, are also conveniently consideredin the terms of the torsional angles associated with thesingle bonds of the eight-membered ring, and these aredescribed using the conventional + and - notation,lJQillustrated in the Newman projections (5a-c).Thesigns of the torsional angles about the single bonds ofW. Klyne and V. Prelog, Experientia, 1960, 16, 621; J. B.Hendrickson, J. AIster. Chem. SOC., 1962, 84, 3366; 1964, 86,4854; 1967,89,7047.the eight-membered ring of the compounds (2) areaccordingly listed below each of the conformational( l a ) C ( + - + - t - )( 4 b ) C * ( - + - - t - + )GCompd. X Y SolventS CDC1,so2 C,D,NN*CH,Ph CDCl,S CDC1,so2 CDC1,- cs, (2 : 1)NCH,Ph CDC1,S CDCl,NCH,Ph CDCl,S CDC1,so9 CDCl,N*CH,Ph CDCl,N*CH,Ph CDCl,TABLE 2Site exchanges and activation parameters for conformational changes of compounds (2a-1)Method Site exchanges a TI'CI A l , m A 2 ,B1,- B2,D1 .S+z D2I11 AIL--=--Bl,C1,- D1I A l e A 2 .B1.- B2D1,- D2I Al.-A2,B1* B2I Al=A2,B1- B2I11 A2- B2I Al,-A2,B1,- B2I Al-A2,Bf B2I11 A 2 , e B 2I Al,-A2,B1,- B2I Al,-A2,c1- c2,c1,- c2,IV A-BIV A.* BIV A-:BB1 -B2I1 Al,-Bl1 A l m A 2 ,B1- B2A2 ,- B2204449-26.6- 52- 88-9-966- 70- 14- 14866060kbls-19.34.656719.29.619.28.62.767.93.966.03.81.93.820.310.18.19.34.69.313.06.611.612.86.4-110 >132AG t Ikcalmo1-115.816.215.917.017.417.013.513.811.912.210.014.715.114.717.317.810.913.914.313.919.319.818.017.918.3< 7.8P I0.910.930.930.690.730.670.810.890.760.96ca. 1.00.94ca.0.0P 20.090.070.070.410.270.330.190.110.250.05ca. 0.00.06ca. 1.0AGc/kcalmol-l Process1.36 C -+ Boatc.-c*c-c*( - l l O ) a C . e = + = C *1.34 C + BoatC + Boat0.180.440.370.940.840.572.11.8< -2.3C ---+ Boatc-c*c.-c*C .__)c BoatBoat ,- BoatC _I) BoatCI-/ BoatC + BoatBoat- BoatC + BoatC+ BoatC+ BoatcL--'c*cr-c*c\-c*c-c*c.*c*C ,-+ Boatc*-c*Boata Details of chemical shifts and coupling constants are given in Table 1. The AB systems A1,Bl and C1,Dl refer to the (2-5, and C-7, and C-(major conformation) and A2,B2 and C2,D2 to the boat (minor) conformation.C*) .= 0.6k(C __)I Boat) if the inversion processconformation (see Figure 8 ) , but if the two degenerate pathways, C __+ TB, are involved in the inversion then for a single C ,-+ TB processC*).1, k(CAt the temperature in column 7 unless otherwise stated. The value given refers to the process C I_+C Boat. d From the spectruJ.C.S. Perkin Iand Y, and these energies have been calculated for thedibenzothiocin derivative (2a) (X = CH,, Y = S) andthe dibenzazocine derivative (2m) (X = CH,, Y =NMe). These calculations (Tables 3 and 4) show thatthe C, conformations B and B* (6a and b) involve, asdiagrams t used in this paper in the order of bonds 4a,6,5,6, 6,7, 7,7a, lla,12, and 12,12a [numbering schemeshown in diagrams (2) and (4a)l.The second conformational type includes a family ofboat-like conformations (Boat) 1 and two pairs of theseTABLE 3Calculated strain energy a,b (Eslkcal mol-l) of various conformations of 7 , l 2-dihydro-5H-dibenzo[c,fJ thiocin (2a)ConformationC - 88.2 + 88.2 1.87 0.15 1.18 1.95 0.04 -1.45 3.19B + 72.0 -71.9 9.67 0.35 7.70 0.48 0.12 1.03 5.19TB - 59.0 - 59.0 5.21 0.24 4.88 0.40 0.66 -0.97 3.29Boat * - 22.9 - 84.0 5.10 0.24 2.60 2.21 0.12 -0.07 4.11Boat f - 27.5 - 80.7 5.30 0.24 3.14 1.82 0.17 -0,05 4.11Boat f 0 - 97.7 5.37 0.22 1.61 3.65 0.02 -0.03 3.88Boat f +21.46 - 109.1 5.61 0.26 1.13 3.53 0.07 0.63 4.23TS1 h 0 +60.0 19.43 0.56 16.70 2.14 0.16 -0.12 4.224 6 .6 4 6 . 7 ES E B Ee E4 a*d E d a ENBI~ ENBIB'Boat f +41.38 - 113.0 5.84 0.31 1.94 2.66 0.06 0.97 4.57TS2 h -112.1 +103.9 33.27 0.75 27.90 3.79 0.08 0.76 4.47TS3 h 0 0 21.69 0.29 16.71 4.38 0.62 -0.41 3.39a Calculations based upon the following force constants. Bond stretch : aromatic km 1 102, Ka 729; aliphatic km 663, haH 665,kos 463 kcd A-2. Angle deformation: aromatic kwa 144, KCOH 108; aliphatic ROW 116, k w ~ 94, KHOH 79, kcso 100, KSCH 89 kcalr a d i a 0 , all angle strain reduced by a factor of 0.7. Equilibrium bond lengths and bond angles assumed as follows : aromatic C-C1.395, C-H 1.09; aliphatic Ar-C 1.50, C-S 1.80, C-H 1.09 A; aromatic CCC 120°, CeH 120"; aliphatic CeC 111.5", CeH 109.5",HeH 108". e For C-S bonds a 3-fold barrier of height 2.1 kcal mol-l is assumed.For aromatic C-C bond twisting and out-of-plane deformation Ed and Ed were calculated according to ref 18. Non-bonded interactions based upon the Hill equation assummarised in ref. 16. f Boat conformations defined by torsion angles about 6,6- and 6,7-bonds, 45,s fixed by fixing appropriateatom co-ordinates. 0 Minimum energy boat conformation obtained by starting with (b5,6 -27.6" 4 6 . 7 -86.1' and allowing freemovement of co-ordinates. TSl defined by keeping atoms 12a, 4a, 5,6, and 7 coplanar, TS2 by keeping atoms 4a, 7a, l l a , 12, and12a coplanar, and TS3 by keeping atoms 4a, 5, 6, 7, and 7a coplanar.TABLE 4Calculated strain energy (Eslkcal mol-l) of various conformations of 5,6,7,12-tetrahydro-6-methyIdibenz[c,flazocine(2; X = CH,, Y = NMe)Conformation 45.8 480 7 E8 E B Ee E4 EA ENBI ENBIBC - 83.7 +83.7 4.30 0.16 1.92 3.99 0.10 -1.88 3.42B +83.0 - 83.0 5.94 0.20 3.84 2.83 0.17 -1.10 4.22TB - 58.4 - 58.4 5.62 0.33 6.48 0.17 0.26 -0.61 4.13Boat b -67.8 - 59.3 5.54 0.32 6.32 0.19 0.40 -0.69 4.07Boat - 38.6 - 70.1 6.23 0.32 3.96 1.65 0.40 -0.10 4.97TS 1 B (TS 1A) 63.5 0.0 22.68 0.68 17.04 4.57 0.07 0.23 5.36TS2 111 112 33.36 0.67 23.87 8.36 0.01 0.66 5.23TS3 0 0 27.29 0.33 16.20 9.62 1.94 -0.69 4.02Q Calculations based on the force constants summarised below Table 3 with the following additions: k m 716 kcal h o o ~ 130,R o ~ c 144, k O m 100 kcal radian-2, angle strain reduced by a factor of 0.7.For CNC bonds a barrier height of 4.4 kcal mol-1 is as-sumed.Equilibrium bond lengths and bond angles as below Table 3 with the following additions: C-N 1.472 A, CtN 111.5', NcH109.6', C$C 109". 8 From a TB starting conformation with retention of C2 symmetry of the 6-8-6 ring system and allowing loss ofC, symmetry. e From a starting conformation with 45.8 - 30.8, $6.7 - 81.4.conformations may be distinguished on the basis of theirsymmetry. The first of these pairs comprises theconformations €3 and B* (6a and b) which have Cdsymmetry. This symmetrical conformation can under-go changes which involve principally torsion about the6,6- and 6,7-bonds to give eventually the second pair ofsymmetrical conformations TB and TB* (7a and b)which have C, symmetry (the C, axis is indicated by thebroken line).This type of conformational behaviourhas been discussed 6 for cyclo-octa-1,4-diene, and it hasbeen pointed out that some angle deformation must beinvolved in the process analogous to B (6a) + TB (7a).The relative energies of the B and TB conformationaltypes clearly depend upon the nature of the groups Xt Conformational diagrams (4) and (6)-(8) are based uponcomputer-produced perspective drawings of the various energy-minimised Conformations of the thiocin derivative (2a) (Table 3).$ In accord with the nomenclature used in our previous paper 1the description ' Boat a refers to any conformation of the boatfamily. The descriptions B, B*, TB, and TB* are specific.expected, non-bonded interactions between the groupsX and Y which result in considerable strain energy,although the interactions are substantially reduced by aflattening distortion of the eight-membered ring.These&@+Y x Y0 \( 6 a ) B ( + + - - + - Iinteractions are steadily reduced during the processB + TB, although the attainment of C, symmetry 8in the TB and TB* conformations (7a and b) requires theintroduction of some strain in the eight-membered ringThe conformational types TB and TB* (7a and b) do nothave Ca symmetry for Y = NR. This conformation, however,does have Ca symmetry if the substituent on the nitrogen atom isignored1976 921in addition to slight out-of-plane deformation of thearomatic rings (see Tables 3 and 4).The torsional situations about the 5,6-and 6,7-bondsin the TB and TB* conformations (7) are close to ideal(torsion angles $6.6 and $6.7 close to 60').In some casestorsional strain in the flattened B and B* conformations(6) may be small [e.g. for the thiocin derivative (2a)]whereas in other cases it may be appreciable [e.g. forthe azocine derivative (2m)I. Although in both thecases examined by strain energy calculations the TBconformation (7) is less strained than the B conformation(6), the differences in strain energies vary and areclearly dependent upon the nature of the groups X andY. The minimum energy boat conformation [see (S)]for the thiocin derivative (2a) lies fairly close to the TI3conformation [see (7)]. However the situation is lessclear for the azocine derivative (2m) and there isapparently little difference in energy between thevarious Boat conformations that lie on the pathwayTB + B --)- TB* --+ B*.From general consider-ations it is probable that for most of the compounds (Z),the minimum energy Boat conformation lies rathercloser to TB (7) than to B (6). In cases where thegroup X can conjugate with the aromatic rings then theTB conformation will be additionally favoured by therelatively small torsion angles about the lla,12- and12,12a-bonds [see (7)].Having considered possible minimum energy conform-ations for the compounds (2), we must now considerpathways by which these conformations may be inter-converted, and possible transition states for each path-way. These have been examined by strain energycalculations for the thiocin derivative (2a) (Table 3) andthe azocine derivative (2m) (Table 4).The followinggeneralisations are based upon these calculations and anexamination of molecular models.The low energy boat conformations TB and TB* (7)are interconverted, by a process involving principallytorsion about the single bonds of the eight-memberedring analogous to the pseudo-rotation of the boatconformations of six-membered rings, by way of theconformations B and B* (6). On the basis of theresults presented in Tables 3 and 4, this process willgenerally involve relatively low energy barriers : con-formations showing the spectral characteristics of rapidinversion are therefore assigned as Boat.The diastereoisomeric C and Boat conformations areinterconvertible by three definable pathways involvingthree different transition states, TSI, TS2, and TS3.The transition state TS1 (9) lies on the pathway C 4TB and is defined by the coplanarity of atoms 12a, 4a,5, 6, and 7 [TSlA, (9a)I or atoms l l a , 7a, 7, 6, and 5[TSlB, (Sc)].These chiral conformations are relatedas mirror images, but it is convenient when discussingspectral changes to consider additionally the conform-ations TSlA* (9b) and TSlB* (9d), and the distinctionbetween these four conformational types is recognisableby the numbering scheme used in formulae (9) toindicate coplanarity of sets of atoms. Calculationsindicate (Tables 3 and 4) that the principal source ofstrain in these conformations TS1 (9) is angle deform-ation (Eo), and that in addition torsional strain isassociated with the eclipsed 5,6- (TSlA) or 6,?-bond(TSlB). The second type of transition state TS2 (10)has C, symmetry and lies on the pathways B L-.C*and B* -* C; it is defined by the coplanarity of atoms(8al Boat ( + - - + + - )#5,6 - 2 2 . 9f6,7 - 8 4 - 0(8b) B o a t * ( - + + - - + ) 9 5,6 i-22.9 56 6,7 $ 8 4 - 04a, 7a, Ila, 12, and 12a. This transition state involvesconsiderable angIe deformation in addition to torsionalstrain, and is, in the two cases examined by strainenergy calculations , significantly higher in energy thanthe other two transition states. The third transitionstate TS3 (11) also has C, symmetry and lies on thepathways I3-C and B**C*. It is defined bythe coplanarity of atoms 4a, 5, 6, 7, and 7a and althoughit involves angle strain rather similar in magnitude tothe transition state TSl (9), it is higher in energy owingto the additiona1 torsional strain associated witheclipsed 5,6- and 6,7-bonds.The relationships amongst the conf ormational typesC, B, TB, TS1, TS2, and TS3 are summarised in Figure 8.It is now necessary to relate this scheme to the n.m.r.spectra, and their temperature dependence for com-pounds (2a-I).summarised in Tables 1 and 2922 J.C.S. Perkin Iadditional site exchanges involving the Boat conform-ation. The additional temperature dependence of theC-12 methylene protons of the monosulphone (2b) wasincluded in the simulated spectra by using method 11.Type 2.These compounds (2a, c, d, f, and h) showa singlet signal (AB2) in their n.m.r. spectra at lowtemperatures which is assignable to the C-5 and C-7methylene groups of a rapidly-inverting, mobile con-formation which must be of the Boat type. In additionthe spectra show an AB system (A1,Bl) associated withthe C-5 and C-7 methylene groups of a slowly invertingconformation which must be of the Chair type. Theseassignments of signals to Chair and Boat conformationsfollow directly from the discussion of low energy con-formational species and the strain energy calculationslisted in Tables 3 and 4. Spectral line-shapes for thesecompounds can be accurately reproduced (Figures 1, 3,6, and 7) using method I for the line-shape calculationstogether with a fast input rate for the inversion of theBoat conformation (Figure 9; k, + 00).Under theseconditions the spectral line-shapes depend, as expected,only upon the average chemical shift of the A2 and B2sites and the population, p,, of the Boat conformation.The separate chemical shifts of the A2 and B2 sites andthe coupling constant J z , which are not obtainable fromthe low temperature spectra, are not required for line-shape computation. The rate for the process C-Boat (Figure 9; I&) is twice the rate for the trans-formation C -+ C* (see Figure 9) : the rates and freeenergies of activation are given in Table 2 for bothprocesses. As the population of the Boat conformation,p,, tends to zero, the spectral line-shapes tend to those ofOn the basis of the information in Tables 1 and 2compounds (2a-k) can be grouped according to threemain types (1-3) of spectral temperature dependence.The relationships between the rate constants for the p 12 x,, l Z a La 5/( 9 a ) T S l A ( 0 0 + - 4- - 1 ( 9 b ) T S l A * ( 0 0 - + - + )( 9 c 1 T S l B ( + - 0 0 f - 1 ( S d ) T S I B * ( - + O 0 - + )t 10al T S 2 ( - .+ - + O O ) ( l o b ) T S 2 * ( + - + - 0 0 )( l l a ) TS3 ( l l b ) lS3*conformational changes, summarised in Figure 8, and thesite-exchange rates, deduced from spectral line-shapes(see Experimental section and below), are summarisedin Figure 9.TyPe 1. The spectra of the C-5 and C-7 methyleneprotons of compounds of this type (2b, i, j, and k)appear to consist of just a single AB system at lowtemperatures (sites A1 and B1, Figure 9) which coalescesto a singlet a t higher temperatures.The spectra insome cases show some asymmetry for site exchange ratesat which some line-broadening occurs (Figure Z), butbelow the exchange rate at which the A and B signalscoalesce. For compounds (2i and k) this asymmetricalline shape can be reproduced using method I by insert-ing into the calculation a few percent of a rapidlyinverting boat conformation (Figure 9; k,--+ co; seealso Experimental section). On the other hand, for theremaining two compounds of this type (2b and j) areasonably good agreement between computed andobserved spectra could be obtained using methods I1and I11 (Experimental section) without postulatingFIGURE 8 Conformational changes of heterocyclic analogues(2) of 5,6,11,12-tetrahydrodibenzo[a,d]cyc~o-octenek*l k, 4 2 C [Boat Boat*] C*kl a k, k,,k,, kS1 A1 .- A2 B1- B2kl, k,,k ,A2 B2k ,FIGURE 9 Relationships between rate constants for confor-mational changes and rate constants used for computation ofspectral line-shapes.a single coalescing AB system, with sites A1 and B1,coupling constant J1, and site exchange rate k, =0.5 k,, (see Figure 2).This latter situation resemblesthat found most frequently for ring-inversion processes1976 923For the amines (2c, f, and h), the n.m.r. spectra showtwo singlet signals at low temperature assignable to theN-benzyl methylene groups of Chair and Boat conform-ations. These coalesce to a single singlet at highertemperatures and line-shapes for this region of then.m.r. spectrum may be accurately simulated usingmethod IV.The close agreement of rate constants forthese three compounds obtained by methods I and IV(Figures 6 and 7) confirms the general correctness of theapproach used for line-shape simulation by method I.Type 3. These two compounds (2e and g) give n.m.r.spectra for the C-5 and C-7 methylene protons that aresimilar to those of compounds of type 2, but at lowtemperatures the rate of inversion of the Boat conform-ation is sufficiently slow for a second low intensity ABsystem (A2,B2) to be observed. For these cases therate of inversion of the boat conformation (Figure 9;K,) may be obtained from low temperature spectral line-shapes using method I for spectrum simulation [withinput values of K,, = kZ1 ==.0 as in Figures 4 and 5(a)].For spectra recorded at higher temperatures the rateconstants, K,, and k,,, associated with the processChair Boat may be obtained using method I as forcompounds of type 2 [Figure 5(b)].Rates of Conformational Change and Activation Para-metem.-The rates of conformational change for com-pounds (2a-k) were measured by matching observedand computed spectral line-shapes at a single selectedtemperature (Table 2 and Figures 1-7). The dis-cussion that follows is based upon the premise that for aconformational change the entropy of activation, A S ,is near zero. The potential energy difference betweenthe minimum energy conformation and the transitionstate for a conformational change is therefore bestcompared with the free energy of activation for thechange, AGS.The free energies of activation for theprocess C - w Boat for compounds (2a-k) are listedin Table 2, which in addition lists free energies ofactivation for the process C C*, since in some casesthis is the only observable process. If the processC _t Boat involves the chiral transition states TSlA(9a) and TS1B (9c) [and similarly C*+ Boat*involves TS1A* (9b) and TSlB* (9d)J, there are twoequivalent pathways available for the transformation.Therefore the observed free energy of activation mustbe corrected by the addition of RTln2 before beingcompared with the calculated potential energy ofactivation (see scheme in Figure 8).On the other hand,if the transition state for the process C -+ Boat isactually the achiral conformation TS3 (Il), only asingle pathway is available and the correction shouldnot be applied. On the basis of the strain energycalculations reported in Tables 3 and 4, it is doubtfulwhether the pathways C -+ Boat* and C* + Boat,involving the achiral transition states TS2 and TS2*(lo), contribute to any of the observable conform-ational changes of compounds (2a-k).Comparison of the data in Tables 3 and 4 with those inTable 2 for the thiocin derivative (2a) (AGt 16.2 kcalmol-1 for CI Boat, corrected for a single pathwayinvolving TSl, and AEt 17.6 kcal mol-l based upon thestrain energy difference between the conformations Cand TS1) shows that the best agreement is obtainedbetween observed and calculated energies of activationif the conformation TS1 (9) is the rate-determiningtransition state.Similarly an assignment of transitionstate geometry to the conformation TSl (9) also leadsto best agreement between observed and calculatedactivation energies for the azocine derivatives (2 ;X = CH,, Y = NMe or N*CH,Ph) (AGS 17.4 kcal mol-Ifor C -+ Boat, corrected for a single pathway involvingTSl, and A E t 18.2 kcal mol-l based upon the strainenergy difference between conformations C and TSl).Group X in (210 CH2 S SO2-S-19 I 1 8 iGroupY in ( 2 )NCH2Ph S SO2i7-S-CH2--S--0-FIGURE 10 Relationships between AG 4 (C --t B) and groupsX and Y [see ( Z ) ] .The horizontal lines are placed correctlyon a vertical energy scale, each line representing the value ofAG 9 for the single compound having the group Y or X indi-cated on the line. The left hand set refers to the group Xremaining constant and Y changing, and the right hand set togroup Y remaining constant and X changingThese results suggest that similar pathways involvingthe transition state TS1 (9) are followed for the con-formational change, C + Boat, for all the compounds(2a-k). The relative free energies of activation shouldtherefore be related to the relative strain energies of theconformations C (4) and TS1 (9). Unfortunately strainenergy calculations cannot be made with confidence forcompounds (2; X = 0, S, or SO,) in which the aromaticrings are potentially conjugated with a heteroatom inposition 12, owing to the difficulties associated with theselection of potential functions for deformation of theAr-X-Ar fragment.The discussion that follows istheref ore necessarily qualitative.The relative heights of the free energy barrier for theprocess C-Boat are shown in Figure 10, whichillustrates the effects of a change in the groups X and Yupon the energy barrier. From the strain energ924 J.C.S. Perkin Icalculations in Tables 3 and 4, these activationenergies reflect principally angle strain in the eight-membered ring of TSl (9) and an increase in non-bonded interactions in TSl. The increase in anglestrain is associated principally with tQe angles a, p, y,and 6 of the eight-membered ring [see (12)], which areopened up to values in TS1 considerably greater than7 E&$=> 5equilibrium values.The changes in non-bonded inter-actions are principally associated with a marked increasein the interaction between C(12)H2 and C(7)H2 in theconformation TS1 [see (12)] relative to C and a rathersmaller decrease in the interaction between C(12)H2 andC(5)H2. In addition, for azocine derivatives (12;X = NR), non-bonded interactions involving the sub-stituent on the nitrogen atom are also increased in TSl.In view of the dominant part played by angle strainin determining the relative energies of the C and TS1conformations, and hence the relative rates of the processC-Boat for the compounds (2a-k), a series ofapproximate calculations of angle strain in TS1 (9) wascarried out.The results are summarised in Table 5.TABLE 5C ~_t Boat for compounds (2a, c, d, f, g, and h)Bond AGtCom- lengths (A) Ee in TSl (C __t B)pound X Y Ar-X CH,-Y /kcal mol-l /kcal mol-lCalculated and observed values of AGb for the process(2a) CH, S 1.51 1.82 14.2 15.8(2c) CH, NR 1.61 1.47 16.0 17.01.42 1.82 13.2 13.616.3 14.7 NR 1.42 1.471.76 1.82 16.4 17.317.3 13.9 (Zh) S NR 1.76 1.470 Based UPOR the angle strain Ee in TS1 [see (9) and (14)]using the expression Eo = 0.021914 (2A$ + Aaz + A@, +Ap + A8s)kcd mat-', where Ac, etc. are the devrations from theunstrained values: a = (3 = E = 120", y = lll", and 6 =109" [see (14)J. For all compounds the unstrained values for thebond lengths d, and dB [see (14)] are taken as 1.39 A (d,) and1.61 A (d,).(24 s(2f)(2g) s sIn view of the uncertainty regarding relative values offorce constants for angle deformation, the same forceconstant was used for all the angles considered in thecalculations, as in our earlier work1 on heterocyclicanalogues of 5,6,11 I 12- tetrahydrodibenzo [a, e] cyclo-octene (1).The value of approximate calculations ofthis type is strictly limited. However, the results doshow that the differences in the energies of activationfor the process C d Boat are consistent with changesin angle strain in the transition states TS1. The originof these differences lies in the changes in the lengths ofthe Ar-X and C-Y bonds; the activation energiesincrease with increasing Ar-X and decrease withincreasing C-Y bond length (Table 5).The observed higher free energies of activation for thesulphones (2; X = SO,) as compared with the sulphides(2; X = S) are also consistent with the transition stateTSl (9) for the process C-Boat.Thus there areimportant non-bonded interactions between one of thesulphone oxygen atoms and C(7)H, in TSlA (Qa) [orC(5)H2 in TSlB (Sc)} which would be considerablyincreased relative to the corresponding interactions inthe C conformation (4). On the other hand, the sul-phones (2; Y = SO,) show lower free energies ofactivation than the sulphides (2; Y = S); this pre-sumably results from the different force constants andequilibrium values for angles and bond lengths associatedwith the groups CS0,C and CSC, but insufficient isknown about these parameters for detailed comment.The experimental result for compound (2h) [AGt(C + Boat) 13.9 kcal mol-l] does not correlate at allwell with the energy of activation derived from crudeangle strain calculations (Table 5 ; calculated anglestrain in TSl 17.3 kcal mol-l). This lack of agreementmay simply result from the very approximate nature ofthe calculations.However, in view of the rather goodcorrelation for the other compounds listed in Table 5,it may indicate in this case that the geometry of TS1 (9)is not a good model for the transition state for theconformational change C _.t Boat.The relative free energies of C and Boat conform-ations are difficult to discuss in a general way.Quali-tatively it is evident frQm the data in Table 2 thatAG (C _+c Boat) increases as the size of the group Yincreases (Y = SO, > S > N*CH,Ph). This is readilyunderstandable in terms of the non-bonded interactionsbetween X and Y in the conformation B (6), but theseare present to an increasingly smaller extent in Boatconformations lying between B and TB and they dis-appear in the TB conformations (7). It is similarlydifficult to discuss the energy barriers for the processBoat +Boat*, although on the basis of the strainenergy calculations presented in Tables 3 and 4 it isprobable that the transition state for this process isclose to the B or B* conformation (6). We note thatthis process is slow, on the n.m.r.time scale, for onlytwo of the compounds studied (2e and g), even at-110 "C, although in two other cases (2d and h) thesignals assignable to the C-5 and C-7 methylene groupsof the Boat conformation are considerably broadened a t-110 "C and the associated free energies of activationmay not be much less than 8-9 kcal mol-l.The amino-ketone (21) is of particular interest in viewof the suggestion,Z0 made on the basis of indirect evidence,that the )CO,)NR transannular interaction is attractiverather than repulsive in certain medium-sized rings.The n.m.r. spectrum of (21), in contrast with those of(2a-k), shows no temperature dependence down to20 F. A. L. Anet, A. S. Bailey, and R. Robinson, Chem.andIozd., 1963,944; N. J. Leonard, D. F. Morrow, and M. T. Rogers,J. Amer. Chem. Soc., 1967,79, 6476; N. J. Leonard, T. L. Brown,and T. W. Milligan, ibid., 1959.81, 6041976-110 "C. The two singlet signals assignable to the C-5and C-7 methylene groups (T 6.26) and the N-benzylmethylene group (7 6.39) of (21) have chemical shiftswhich are similar to those of the rapidly inverting Boatconformations of (2c) ( T 6.26 and 6.19), (2f) (7 6.43 and6.50), and (2h) ( T 6.29 and 6.49), which are ratherdifferent from the averaged chemical shifts for the C-5and C-7 methylene protons and the N-benzyl methyleneprotons of the Chair conformation of these compounds[(2c): T 5.67, 6.80; (2f): T 5.75, 6.84; (2h): T 5.50,6.781. It is therefore concluded that the amino-ketone(21) adopts the Boat conformation, and from the lack oftemperature dependence in its n.m.r. spectrum thereappears to be less than 2-3% of the Chair conform-ation.The latter might well be undetectable owing toits low concentration, but assuming that the Chair +Boat exchange rate (Al2) would change from slow to fastin the temperature range over which the n.m.r. spectrumremains unchanged [-110 to +lo0 "C, cf. compounds(2c, f, and h)], this lack of observable line-shape changessets a low value for the population of the Chair con-formation.The Boat conformation of compound (21) [with AG(C-Boat) < -2.3 kcal mol-l at 25 "C assuming(2% of the Chair conformation] suggests that the > C0,)NR interaction may be attractive to the extent ofat least 2-3 kcal mol-l [cj.compound (2f), havingAG (C _.t Boat) 0.37 kcal mol-l] as compared with theexpected non-bonded interaction between the nitrogenand carbon atoms. The free energy of activation forthe process B + B* for (21) is evidently less than8 kcal mol-l from the lack of signal separation at lowtemperatures; assuming that the N,CO attractive inter-action would be lost in the transition state for con-formational inversion, and assuming that the free energyof activation for this process would otherwise be zero, anupper limit of ca. 8 kcal mol-l is suggested for anyattractive interaction that may be present. This rangeof 2-8 kcal mol-l implies a weakly bonding interactionwhich may be represented in MO terms as involving fourelectrons in two of the three MOs resulting from linearcombinations of the carbon and oxygen 2fi orbitalstogether with a nitrogen se3 orbital in a homo-amidegrouping with the bonding depicted in (13); the C,Ndistance in the undistorted C, Boat (B) conformationshown in (13) would be ca.2.4 A but we have no06-definitive information regarding the type of boat con-formation actually adopted by the azocine derivative (21).R. N. Renaud, R. B. Layton, and R. R. Fraser, Canad. J .Chew., 1973, 51, 3380.CortcZusions.-Compounds (2a-k) exist in Chair (4)and Boat conformations in solution, and in all the casesexamined the Chair conformation is of lower energy.The kinetics of the process Chair-Boat may bedetermined by n.m.r. line-shape methods, and theassociated activation parameters may be compared withthe results of strain energy calculations. This com-parison shows that the conformation TSl (9) is theminimum energy transition state for this conformationalchange, and from the strain energy calculations (Tables 3and 4) it is also evident that the energy barrier for theprocess Chair --+I Boat results principally from theconsiderable angle strain in the transition state TS1 (9).The conformational changes of compounds (2a-k)therefore provide a useful probe into the magnitude ofangle strain for quite large angle deformations. Inaddition, for systems lacking ArXAr conjugation (X =CH,), the measured activation energies for the processChair-Boat may be used to examine variousprocedures to calculate angle strain.The examination of compounds such as (21) shouldprovide information concerning the magnitudes oftransannular interactions between > C=O and (NRgroups which have previously been discussed zo only inqualitative terms. These interactions have also beendiscussed for 1 ,%bridged naphthalene derivatives.18Pala ef aZ.ll have discussed briefly the n.m.r. spectrumof the azocine derivative (2c) and closely related com-pounds. The results agree essentially with ours al-though exchange rates and activation parameters werenot reported. The dibenzazocine derivative (2m) hasbeen examined in some detail and the results werepublished 21 after the completion of our studies.18 Thecalculated values of AG and AGt for Chair __B. Boat(' crown ' -+ ' flexible in the nomenclature of ref. 21)agree well with those determined in this work for thedibenzazocine derivative (2c), and are also in goodagreement with the strain energy calculations for (2m)reported here [AG (Chair + Boat) 1.8 kcal mol-l,AGzllt (Boat -+ Chair) 15.3 kcal mol-l, AH (Chair +Boat) 3.2 kcal mol-l, A S (Chair Boat) 7 & 3 calmol-l K-l, calculated A E (Chair + Boat) 1.24 kcalmol-l, AE (Boat -w TSl) 17.04 kcal mol-l]. An ex-amination 22 of the crystal structure of this dibenzazocinederivative (2m) has also been reported; the structure ofthe chair conformation found in the crystalline state is inreasonably good agreement with the structure detailsobtained from strain energy calculations. The corre-sponding N-t-butyl derivative (2n) is, however, foundto adopt a boat conformation in the crystalline state,*3and in solution, with torsion angles in the crystalmidway between the conformations described in thispaper as B (6) and TB (7) (45,6 -115.5", &, +44.1" inone enantiomer and 45,6 -64.5, +6,, +135.9 in theother).[4/2699 Received, 30th December, 1974323 A. D. Hardy and F. R. Ahmed, Acta Cryst., 1974, BSO, 1670.*3 A. D. Hardy and F. R. Ahmed, Acta Cryst., 1974, B80, 1674

ISSN:1472-7781    

DOI:10.1039/P19760000913

出版商:RSC    

年代:1976

数据来源: RSC