2574 J.C.S. Perkin ISynthesis of New Macrocycles. Part 5.l Cyclization of 2,2rsquo;-Dithiodi-benzoic Acid DerivativesBy Siegfried E. Drewes and Brian G. Riphagen, Chemistry Department, University of Natal, Pietermaritzburg,Republic of South AfricaCondensation of 2,2rsquo;-dithiodibenzoyl chloride with a series of alkanediols, but-2-yne-l.4-diol. triethylene glycol.and tetraethylene glycol affords cyclic esters in yields ranging from 5 to 95. On the basis of n.m.r. evidence, pre-dictions are made regarding conformations of the macrocycles.THE behaviour in selected cyclization reactions ofderivatives of 2,2rsquo;-dithiodibenzoic acid (1), a difunctionalacid possessing in its skeleton certain elements ofrigidity, was investigated as an extension of our earlierstudies on this theme.2 Initial attempts at cyclizationinvolved treatment of the dipotassium salt with di-bromoalkanes, under conditions similar to those em-ployed in synthesizing phthalic acid macro cycle^.^Only 1,4-dibrornobutane and 1,6-dibromohexane yieldedcyclic products, Since the diacid chloride of (1) isreadily prepared 4 and is also reasonably stable at roomtemperature, this derivative was employed in subsequentreactions with selected diols.Best results were achievedwith benzene as solvent, and a low reaction temperaturemaintained over an extended period, with triethylamineas base. This technique yielded macrocyclic esters(2)-(9) in yields ranging from 5 to 95 and varying inring size from 14 to 21 members, including botheven and odd numbers of atoms (Table).Monomersonly were isolated. With the lower diols (ethane-1,2-diol as well as propane-1,3-diol), solid material wasobtained from the condensation but this proved in-soluble in all organic solvents tried, making furtherinvestigation impossible. Similar insoluble productswere obtained from reactions with hexa-2,4-diyne-l,6-diol and o-phenylenedimethanol, Smolinski and Jam-rozik5 report similar problems in the synthesis ofmacrocycles based on pentaerythritol.In their recent review on the synthesis of macrocyclicsulphides, Bradshaw and Hui6 draw attention to thelack of specific information regarding this type ofmacrocycle in comparison with oxygen and nitrogenmacrocycles.report that 2,2rsquo;-thiodibenzyl bromide forms macro-cycles on treatment with amines.Yields of the 8-membered monomers varied between 15 and 46 andof the corresponding 16-membered dimers between3 and 27. Yields of our cyclic ester disulphides(2)-(9) compare favourably with these dimer yields.Overall, they are also higher than our yields of macro-cyclic phthalic acid ester~.l-~ Compound ( 7 ) , whichHowever, Tanaka and his co-workersPart 4, S. E. Drewes and P. C. Coleman, J.C.S. Perkin I ,S. E. Drewes and B. G. Riphagen, J.C.S. Perkin I , 1974,S. E. Drewes and P. C. Coleman, J.C.S. Perkin I, 1972, 2148.F. iirndt, A. Kirsch, and P. Nachtwey, Ber., 1926, 59, 1077.S. Smolinski and I. Tamrozik, Tetvahedron, 1971. 27. 4977.1974, 2578.1908.bears a rigid bridging group, was not obtained in higheryield than the corresponding 14-membereomd cpoundq + c=oS(101(11) (12)(2) which contains a flexible bridge.Although ouryields tended to be higher for the smaller rings, theJ. S. Bradshaw and J: Y . K. Hui, J . Heterocyclic Chern:, 1974, 7 S. Tanaka, K. Hashimoto, and H. Wanatabe, J . PAarm.11, 649. SOC. Japan, 1973, 93, 9911976 2575opposite effect has been observed for a homologousseries of dithia-heterocycles.*All eight macrocycles were tested for complex form-ation with metal ions by using the picrate m e t h ~ d . ~No positive result was obtained. This was not surprisingfor the macrocycles (2)-(7) since Pedersen9 has shownthat complexing power for alkali and alkaline earthmetals is largely destroyed by substituting -S- for -0-in the ring.Compounds (8) and (9), however, althoughthey contain a poor complexing group in the form of adiester, still contain two sulphur atoms and at least twoether oxygen atoms in the rings. Nevertheless, bothgave a negative result with sodium, potassium, andlithium cations. Even with silver nitrate, which usuallyforms strong complexes with macrocyclic polyethers~lphides,~ no success was achieved. It has to beassumed that these macrocycles are not capable oftwisting in such a way that a favourable co-ordinationgeometry is achieved around the central ion.1deg;Stability of the Macrocyc1es.-The only way in whichthe macrocycles could be prepared satisfactorily in-volved a prolonged (3-5 days) reaction at room tem-perature.If the solvent (benzene) was refluxed, only1,4-dibromobutane and 1,6-dibromohexane gave recog-nizable cyclic products, although in decreased yields.Recrystallization was performed from dioxan through-out, but even during this process considerable poly-merization took place. This was particularly evidentfor the alkyne macrocycle (7). In the crystalline formall the macrocycles were stable for several months atroom temperature.N.M.F. a;tzd Mass S$ectra.-All the macrocyclesexhibited a molecular ion peak in the mass spectrum,and except for (2) and (7) this was of 357; relativeabundance. The only other significant peaks werecommon to all the macrocycles, at mle 136 (base peak),153, and 180, probably due to the ions (10)-(12),respectively .,411 eigli t macrocycles exhibited a complex eight-protonn.m.r.multiplet at 6 7-8.5 due to the aromatic protons.For (2)-(6) the terminal methylenes resonated as atriplet at about 6 4.4 and the central methylenes as amultiplet between 8 1.42 and 2.05. A similar pattern,differing only in chemical shifts was shown by (5) and (9).In ( 7 ) the signal for the two methylenes appeared as asharp singlet at 6 5.07.Possible Conformation of the Macrocyc2es.-The crystalstructure of 2,2'-dithiodianiline l1 suggests that hydrogenbonds between the amino-groups tie the molecules inroughly parallel chains. The angle between the planesof the two rings is 29". We suggest that the diacid (1)takes up a very similar conformation and further thatthe cyclic compounds (2)-(9) also possess a closelysimilar conformation.Accurate molecular models showthat a bridging chain consisting of 4-8 methylenegroups, in the staggered conformation, can be insertedbetween the two oxycarbonyl groups without disturbingD. W. Allen, P. N. Braunton, I. T. Millar, and J. C . Tebby,J . Chem. SOC. (C), 1971, 3456.the conformation. Such a conformation also satisfies then.m.r. data since (i) it allows the two terminal methylenesin each macrocycle to exist in an identical magneticenvironment and (ii) it can accommodate the observ-ation that the complex but characteristic aromaticmultiplet of (2) is identical in form with that exhibitedby the other macrocycles (3)-(6).EXPERIMENTALMass spectra were obtained with a Varian CH7 spectro-meter at 70 eV and n.m.r.spectra with a Varian T60instrument.2,2'-DithiodibenxoyZ Chloride.-Treatment of the acid (1)with thionyl chloride gave the acid chloride, m.p. 153"(60), which could be stored unchanged in a vacuumdesiccator for extended periods.7,8,9,10-Tetvahydrodibenzoc,g l, 10,5,6dioxadithiacyclo-tetradecin-5,12-dione (2) .-The acid chloride (1.5 g, 4 mmol),butane-1,4-diol (0.39 g, 4 mmol), and triethylamine (1.23ml, 8.8 mmol) in benzene (75 ml) were stirred for 72 h a t22 'C. Moisture was rigorously excluded. A precipitateof triethylamine hydrochloride formed during the reaction,and the reaction was terminated when precipitation wasjudged to be complete. After addition of water (100 ml)the benzene layer was separated, dried, and evaporatedin vacuo.The solid product crystallized from dioxan asfine needles (1.5 g, 95), m.p. 157' (Found: C , 60.2; H,4.6. C,,H1,04S, requires C, 60.0; H, 4.5).8,9,10,1 l-Tetrahydro-7H-dibenzoc,g-J 1,10,5, 6dioxadithia-cycZofientadecin-5,13-dione (3) .-This compound was pre-pared as above but was more difficult t o obtain pure.Chromatography on an alumina column (chloroform aseluant) gave white needles (9.2), m.p. 85-87' (fromethyl acetate) (Found: C, 60.7; H, 4.9. C,,H,,O,S,requires C, 61.0; H, 4.85).Analytical and physical data for macrocycles (2)-(9)(IFound Reqd. - bsol;I Ring Yield M.p. ,-*Coinpd. size bsol;yo) {"C) C H Formula C H( 2 ) 14 95.0 157 60.2 4.6 C18H1,04S, 60.0 4.5(3) 15 9.2 85- 60.7 4.9 C,,H,,O4Sp 61.0 4.8587(4) 16 53.1 147 61.5 5.3 C20H,,0,S, 61.8 5.2(5) 17 56.9 117 63.0 5.5 C,,H,,0,S2 62.7 5.5(6) 18 5.5 186 63.4 6.25 C,,H,,O,S, 63.4 5.8(7) 14 51.4 148- 60.5 3.2 C,,H,,O,Sp 60.7 3.4150(8) 18 5.4 193 57.3 4.75 C,,H,oO,S, 57.1 4.8(0) 21 4.9 141- 56.4 4.8 C,,H,,O,S, 56.9 5.2142The remaining macrocycles, 7,8,9,10,11,12-hexahyd~o-dibemoc,g 1, 10,5,6dioxadithia~yclohexadecin-5,14-dione(4), 8,9,10,11,12,13-hexahydro-7H-dibenzo c, g 1, 10,5,6di-oxadithiacycloheptadecin-5,15-dione (5), 7,8,9,10,11,12,13,14-octahydrodibenzo c, g 1 , 10,5,6dioxadithiacyclo-octadecin-5,16-dione (6), 8,9-didehydro-7,10-dihydrodibenzoc,g-l, 10,5,6dioxadithiacycZotetradecin-5,12-dione (7), 7,8,10,-11,13,14-hexahydrodibenzoc,g 1,4,7,10,14,15tetraoxadithia-cycZo-octadecin-5,16-dione (8), and 7,8,10,11,13,14,16,17-octalzydrodibenzoc,gl,4,7,10,13,17,18~entaoxadithiaeyeZo-?zeneicosin-5,19-dione (9), were prepared by essentially the9 C.J, Pedersen, J . Ovg. Chem., 1971, 36, 254.10 L. F. Lindoy, Chern. SOC. Rev., 1975, 4, 421.11 A. H. Gomes de Mesquita, Acta Cryst., 1967, 23, 6712576 J.C.S. Perkin Isame procedure as for (2). Reaction times varied between Analytical and physical data are summarized in the3 and 5 days at 22 "C. All the macrocycles were verysoluble in chloroform, fairly soluble in benzene and dioxan,and insoluble in acetone. The alkanediols were purchased We thank the South African Council for Scientific andcommercially as were but-2-yne-l,4-diol (7)1, tri- Industrial Research for financial assistance and a Bursaryethylene glycol for (S) and tetraethylene glycol for (to(911. 6/1256 Received, 28th June, 19761Table.G. R-)
展开▼
