4 Reaction Mechanisms Part (iii) Free-radical Reactions By D. CRICH Department of Chemistry University College London 20 Gordon Street London WClH OAJ 1 Synthesis Intramolecular Processes.-1988 was an encouraging year for the free-radical chemist with the publication of several exciting innovations. In the first instance and following on from the work of Hanessian and Kametani (see last year's Report) several papers have been published illustrating the possibility of preparing six- membered rings as opposed to the more usual five-membered rings by radical cyclizations' as exemplified in Schemes l2 and 23. / /hBz Bu,SnH. AlBN ( II 82% Scheme 1 Scheme 2 70% A further example (Scheme 3) however draws attention to the fact that caution must be exercised and that rearrangements can cause such cyclizations to take an unexpected co~rse.~ Thus the aryl bromide (1) on treatment with tributyltin hydride (TBTH) under standard conditions led presumably uia radicals (3) and (4) to the phenanthrene (2) and not to the anticipated isomeric product (5).T. Sugawara B. A. Otter and T. Ueda Tetrahedron Lett. 1988 29 75; T.Rajamannar and K. K. Balasubramanian ibid. p. 5789; S. A. Ahmad and D. A. Whiting 1. Chern. SOC.,Chern. Cornrnun.,1988 1160; S. Hatakeyama N. Ochi H. Numato and S. Takano ibid. p. 1202. ' M. Kim R. S. Gross H. Sevestre N. K. Dunlop and D. S. Watt J. Org. Chern. 1988 53 93. T. Ghosh and H. Hart J. Org. Chern. 1988 53 2396. N. S. Narasimhan and J. S. Aidhen Tetrahedron Leu. 1988 29 2987.71 72 D. Crich ro ro 0 OH Bu,SnH AlBN b +-Me0 Me0 80"C Me0 Me0 Me0 OMe OMe OMe (5) (1) (2) 23% I T -Me0 Me0 I OMe OMe (3) (4) Scheme 3 Returning to the formation of the more familiar five-membered rings a high yield example of an apparent 5-endo-trig cyclization (Scheme 4) has been published without ~ornrnent.~ Scheme 4 Several authors have studied amides as linkages for the formation of five-membered nitrogen heterocycles by radical cyclizations with widely varied results. Thus successful cyclizations were reported by Yamaguchi6 (Scheme 5) and by Dittami'" (Scheme 6; R = Me). However by simply substituting benzoyl for acetyl in the latter example Togo7' obtained an alternative mode of cyclization (Scheme 6; R = Ph) demonstrating the susceptibility of such reactions to substituent effects.A. Gepalsamy and K. K. Balasubramanian,J. Chem. SOC.,Chem. Commun. 1988 28. R. Yamaguchi T. Hamasaki and K. Utimoto Chem. Lett. 1988 913. ' (a)J. P. Dittami and H. Ramanathan Tetrahedron Lett. 1988,29,45; (b) H. Togo and 0.Kikuchi ibid. p. 4133. Reaction Mechanisms -Part (iii) Free-radical Reactions 0 0 62Yo Scheme 5 Bu,SnH 80"C Bu SnH 80°C -P R= Me R=Ph I Ac 04 ) 100% 91% Scheme 6 Bowman et aL8 encountered problems in their work on the synthesis of oxindoles by the TBTH-mediated cyclization of o-iodophenylacrylamides owing to the unfavourable (for cyclization) trans-amide configuration adopted by the intermedi- ate radical.Similar problems were faced by Jolly and Livinghouse' when treatment of the iodoamide (6) with TBTH under standard conditions gave only a low yield (27%) of cyclized product (7) with the major product (53%) resulting from reductive deiodination. The problem was circumvented in this case by the use of a combination of hexabutyldistannane and ethyl iodide under photolytic conditions resulting in the cyclized product (8) with a much improved yield (88%). This ingenious solution relies on the ability of ethyl iodide to quench not only the cyclized radical but also the ring-opened trans-amide radical (9) so enabling its recycling to (6). Me (7) I I I Me Me (8) (9) Three groups have independently reported on cyclizations preceded by what Curran has termed 'translocation of radical sites by intramolecular 1,5-hydrogen abstraction'.Thus Cekovic" obtained the cyclopentane (11) in 32% yield from a 'W. R. Bowman H. Heaney and B. M. Jordan Tetrahedron Lett. 1988 29 6657. R. S. Jolly and T. Livinghouse 1.Am. Chem. SOC.,1988 110 7536. Z. Cekovic and D. Ilijev Tetrahedron Lett. 1988 29 1441. 74 D. Crich modified Barton reaction involving photolysis of the nitrite ester (lo) whilst Parsons' entry into the pyrrolizidine precursor (13) was achieved (60-85%) by photolysis of the iodide (12) in the presence of TBTH." The examples (10) -+ (11) and (12)+ (13) are conceptually different in so far as in the former the initial radical provoking the 1,Shydrogen transfer is remote from the double bond required for the cyclization whilst in the latter the double bond intervenes in the reaction both as a hydrogen-abstracting vinyl radical and subsequently as a radical trap.Curran has published several examples of each type but has drawn particular attention to the use of modified protecting groups in the former type for radical generation and protection of the cyclized product as in Scheme 7.12 Bu,SnCl NaBH,CN - crCOZEt 61'/o Scheme 7 The cyclopropylmethyl/3-butenylradical rearrangement has been used as a key step in various tandem radical procedures by several groups. Thus Utim~to'~ reported on the rearrangement of butadienylcyclopropanes to alkenylcyclopentenes as illus- trated in Scheme 8. The reaction is initiated by addition of a phenylthio radical to the terminal position of the diene followed sequentially by cyclopropylmethyl ring opening 5-hexenyl ring closure and eventual p -elimination of the phenylthio group.Two groups of workers have published on an alternative procedure (Scheme 9) in which a phenylthio radical adds to a vinylcyclopropane generating after ring " D. C. Lathbury P. J. Parsons and I. Pinto J. Chem. Soc. Chem. Commun. 1988 81. D. P. Curran D. Kim H. T. Liu and W. Shen J. Am. Chem. SOC.,1988 110 5900. l3 T. Miura K. Fugami K. Oshima and K. Utimoto Tetrahedron Lett. 1988 29 1543. Reaction Mechanisms -Part (iii) Free-radical Reactions PhSH,AIBN 60 ' u L-J"c 0 Me 73O/O Scheme 8 COzMe PhSH,AIBN - VCOzMe OEt 60°C EtOv =(om OEt 79yo Scheme 9 opening a 3-butenyl radical that was quenched by addition to an alkene to give a 5-hexenyl radical which then underwent closure with expulsion of the phenylthio radi~a1.l~ The approach adopted by Motherwell" (Scheme 10) is somewhat different in so far as the fragmentation/ cyclization sequence was initiated by application of the Barton-McCombie reaction and also as the alkene introduced on cyclopropylmethyl ring opening took no further part in the sequence.Bu,SnH AlBN 80 "C 'i 71% SiMe Scheme 10 In each of the three examples listed above regioselective cleavage of the cyclopro- pyl group was crucial to the success of the reaction and whilst in the first two cases this may be explained in terms of radical stability this is not so in the third example where stereoelectronic control was invoked.Further examples of regioselective opening of cyclopropyl radicals were reported by Griller16 and Japanese workers." Turning to the related opening of oxiranylmethyl radicals Murphy'* has reported further cases in which the C-0 bond is cleaved preferentially to the C-C bond. 14 T. Miura K. Fugami K. Oshima and K. Utimoto Tetrahedron Lett. 1988 29 5135; K. S. Feldman A. L. Romanelli R. E. Ruckle and R. F. Miller J. Am. Chem. SOC.,1988 53 3300. J. D. Harling and W. B. Motherwell J. Chem. Soc. Chem. Commun. 1988 1380. 16 M. Campredon J. M. Kanabus-Kaminska and D. Griller J. Org. Chem. 1988 53 5393. 17 T. Morikawa M. Uejima and Y. Kobayashi Chem Lett. 1988 1407.J. A. Murphy C. W. Patterson and N. F. Wooster Tetrahedron Lett. 1988,29,955; J. Chem. SOC.,Chem. Commun. 1988 294; A. Johns and J. A. Murphy Tetrahedron Lerr. 1988 29 837. 76 D. Crich It is apparent that unless C-C cleavage leads to a stabilized radical (see last year’s Report) oxiranylmethyl radical cleavage takes place with scission of the C-0 bond to give 2-propenyloxyl radicals. Stork” has studied the use of ally1 radicals in cyclization reactions and found that good yields of five-membered rings can be formed (Scheme 11). f Br I 5 Jij Scheme 11 Several groups have investigated the use of acyl radicals in cyclization reactions leading to cycloalkanones. The Reporter and Boger preferred selenoesters as the precursor and noted the efficient formation of cyclohexanones by the 6-em-trig mode (Scheme 12).20 Zard on the other hand preferred S-acylxanthate esters as photolabile acyl radical precursors.21 Kagan22 has reported the isolation of hydroxy-cyclopropanes on treatment of 0-allylsalicylyl chloride with samarium( 11) iodide and suggested inter alia that the reaction might proceed by acyl radical cyclization closure of the cyclopropane and trapping of the alkoxy radical by some samarium species (Scheme 13).Blanco and Ma~s0ux-i~~ have drawn attention to the possibility of the reversibility of acyl radical cyclizations by isolating ring-opened products (17) (18) and (19) in 18,33 and 22% yield respectively from the reaction of (14) with ferric chloride. The presumed intermediates were the radical (15) the ring- opened acyl radical and the acyl chloride (16).72Yo + Scheme 12 19 G. Stork and M. E. Reynolds J. Am. Chem. SOC.,1988 110 6911. 2o D. Crich and S. M. Fortt Tetrahedron Lett. 1988 29 2585; D. L. Boger and R. J. Mathvink J. Org. Chem 1988,53 3379. 21 P. Delduc C. Tailhan and S. Z. Zard J. Cbem. SOC.,Cbem. Commun. 1988 308. 22 M. Sasaki J. Collin and H. B. Kagan Tetrahedron Lett. 1988 29 6105. 23 L. Blanco and A. Massouri Tetrahedron Lett. 1988 29 3239. Reaction Mechanisms -Part (iii) Free-radical Reactions 51% Scheme 13 Ph (16) X=C1 (17) X=OH (18) X=OEt (19) X=O c1Q Ph The use of monothioacetals or ketals in conjunction with TBTH as precursors to a-alkoxyl radicals and eventually to cyclopentanols following radical cyclization has been exploited independently by two groups (Scheme 14).24 n Bu,SnH AIBN 80 "C 60% Scheme 14 Nugent and RaJanBab~~~ have reported the generation of p-oxy radicals from epoxides on treatment with bis(cyclopentadieny1)chlorotitanium.Following cycliz- ation on to an appropriately placed double bond the organometallic reagent further served as an alkyl radical trap. Work-up with water provided alcohols (Scheme 15) and with iodine iodoalcohols. c1 c1 I OF ____ cp2i&cp2 H,O *Cp,TiC -$) THF 82Yo Scheme 15 24 V. K. Yadav and A. G. Fallis Tetrahedron Lett. 1988 29 897; T. L. Fevig R. L. Elliott and D. P. Curran J. Am. Chem Soc.1988 110 5064. 25 W. A. Nugent and T. V. RajanBabu. J. Am. Chem. SOC.,1988 110 8561. 78 D. Crich Finally in this section the cyclization of alkyl radicals on to oxime ethers was found to give good yields of N-cyclopentyl- and in some cases N-cyclohexylhy- droxylamines.26 Intermolecular Processes.-The report by Hart27 of the ability of the couple bis(trimethylstanny1) benzopinacolate (20)/ 0-benzylformaldoxime to generate alkyl radicals from halides and pseudohalides and to occasion their one-carbon homologa- tion (Scheme 16) represents a significant advance in the field of radical-mediated intermolecular carbon-carbon bond formation. In a similar vein one-carbon homologation of radicals generated by the 0-acyl thiohydroxamate route was achieved with isonitriles substituted with electron-withdrawing groups e.g.4-nitrophenylisonitrile.** OCHzPh Me,SnO OSnMe I Ph,C-CPh (20) CH,=NOCH,Ph 80°C OAc OAc Scheme 16 80% Diethyl azodicarboxylate has been reported to trap alkyl radicals generated from alkyl halides with TBTH leading after chain transfer to N-alkylhydrazinedicarboxy-late^.^^ The modest yields observed in this radical C-N bond forming process could probably be improved with the replacement of TBTH by (20). The use of organocobaloximes as precursors to alkyl radicals was again popular in 1988 the main exponents being the groups of Branchaud and Pattenden.30 The Pattenden variation of hydrocobaltation-radical coupling-dehydrocobaltation is especially interesting as very minor modifications to the experimental procedure allow the coupling of two alkenes with the controlled formation of either of two regioisomers (Scheme 17).I +ph 1” Ph Ph- CN 55% Ph Scheme 17 26 P. A. Bartlett K. L. McLaren and P. C. Ting J. Am. Chem. SOC.,1988 110 1633; K. A. Parker D. .M. Spero and J. Van Epp 1. Org. Chem 1988 53 4628. 27 D. J. Hart and F. L. Seely J. Am. Chem. SOC.,1988 110 1631. D. H. R. Barton N. Ozbalik and B. Vacher Tetrahedron 1988 44 3501. 29 M. Ohno K. Ishizaki and S. Eguchi J. Org. Chem. 1988 53 1285. 30 B. P. Branchaud M. S. Meier and Y. Choi Tetrahedron Lett. 1988 29 167; B. P. Branchaud and M. S. Meier ibid. p. 3191; H. Bhandal and G. Pattenden J. Chem. SOC.,Chem Commun. 1988 1110; J. E. Baldwin and C.S. Li ibid. p. 261. Reaction Mechanisms -Part (iii) Free-radical Reactions Other novel uses of cobaloximes have included their use in the generation of glycosyl radicals31 and their use as radical precursors for the addition of radicals to nitronate anions and pyridinium Utim~to~~ has extended his work on the assistance of various TBTH-mediated radical reactions by triethylborane (see last year's Report). Of particular note is the reaction of a-bromoketones with TBTH/Et,B resulting in the formation of a boron enolate by a radical mechanism and its in situ quenching with benzaldehyde (Scheme 18). But perhaps the observation of greatest potential importance was that the addition of Et3B enables the TBTH reduction of secondary alkyl thioesters (Barton-McCombie reaction) to be carried out efficiently at room temperature rather than the previously required minimum of 80 "C.0 OBEt2 I ~ph 74% Scheme 18 New Radical Sources.-Tris( trimethylsily1)silane (21) has been advocated34 as a convenient replacement for TBTH given the similar strength of its Si-H bond and the Bu3Sn-H bond and the comparable reactivities of the (Me3Si)3Si' and Bu3Sn' radicals towards alkyl halides. Three new sources of alkoxyl radicals -nitrate e~ters/TBTH,~' diphenylbis(tri-fluoroacetoxy)selurane (22)/iodine and an and 0-alkyl thiohydroxamates (23)37-have been reported in the course of the year. The ease of preparation of compounds (23) render them very attractive for this purpose. (Me3Si),SiH Ph,Se( OCOCF,) (21) (22) The reaction of phosphonyl chlorides with 2-mercaptopyridine N-oxide has been reported to provide the thiohydroxamic acid derivatives (24) which on photolysis with a thiol or tetrachloromethane underwent cleavage of the P-C bond by a radical 31 A.Ghosez T. Gobel and B. Giese Chem. Ber. 1988 121 1807. 32 B. P. Branchaud and G. X. Yu,Tetrahedron Lett. 1988 29 6545; B. P. Branchaud and Y. L. Choi J. Org. Chem. 1988 53 4638. 33 K. Nozaki K. Oshima and K. Utimoto Tetrahedron Lett. 1988 29 1041 6125 6127; Y. Ichinose K. Oshima and K. Utimoto Chem. Lett. 1988 1437. 34 C. Chatgilialoglu D. Griller and M. Lesage J. Org. Chem. 1988 53 3641. 35 B. Fraser-Reid G. D. Vite B. W. A. Yeung and R. Tsang Tetrahedron Lett. 1988 29 1645; G.D. Vite and B. Fraser-Reid Synth. Commun. 1988 18 1339. 36 R. L. Dorta C. G. Francisco R. Freire and E. Suarez Tetrahedron Lett. 1988 29 5429. 37 A. L. J. Beckwith and B. P. Hay J. Am. Chem. SOC.,1988 110 4415. 80 D. Crich chain mechanism; yields however were only modest even for ally1 and benzyl radicals.38 Stereochemical Aspects.-Tanne9’ has reported the use of chiral dihy-dronicotinamides for the reduction of unsymmetrical ketones. Thus the homochiral NADPH model (25) brought about reduction of phenyl trifluoromethyl ketone to the corresponding alcohol in good yield (82% ) and interesting enantiomeric excess (68%) by a chain mechanism involving SET and enantioselective hydrogen atom transfer. The synthesis of P-glycosidic linkages by stereoselective radical reactions at the anomeric centre of pyranosides has been demonstrated by two groups.40 This new chemistry differs from the previously known glycosyl radical chemistry in so far as it is a hydrogen atom that is introduced into the axial position in the stereoselective radical step rather than the more usual trapping by addition of the glycosyl radical to an alkene.The requisite 1-alkoxyglycosyl radicals may be generated by decarboxy- lation of ulosonic acids in the presence of a thiol (Scheme 19) or by the action of TBTH on a monothio-orthoester. PhCH~O~ofl} PhCHzO PhCH20 H -phCHzOh0q COZH H QReagents i b4S C1-; ii RSH hv 51%; p:ff=11:1 0 Scheme 19 2 Mechanism and Physical Rearrangements.-Radicals of the type (26) behave in substantially different ways according to the nature of the S-alkyl group.Thus in a model study for the methylmalonylSCoA to succinylSCoA rearrangment Halpern41 treated the bromide (27) with TBTH under standard conditions and obtained the expected product (28) 38 L. Z. Avila and J. W. Frost J. Am. Chem. SOC.,1988 110 7904. 39 D. D. Tanner and A. Kharrat J. Am. Chem. SOC.,1988 110 2468. 40 D. Crich and T. J. Ritchie Tetrahedron 1988,44 2319; J. Chem. Soc. Chem. Commun. 1988 1461; D. Kahne D. Yang J. J. Lim R. Miller and E. Paguaga J. Am. Chem. Soc. 1988 110 8716. 41 S. Wollowitz and J. Halpern J. Am. Chem. SOC.,1988 110 3112. Reaction Mechanisms -Part (iii) Free-radical Reactions of 1,2-thioester migration.On the other hand Tada42 isolated p -thiolactones (30) in high yield (92%) on photolysis of cobaloximes (29). Evidently 1,2-thioester migration and SH2 at sulphur are competing processes with the balance being tipped in favour of the latter by good radical leaving groups on sulphur. 0 C0,Et COlEt 0 II I I II EtSC-C-CH2Br HC-CH2-CSEt I I Me Me (27) (28) The high yield (80%) formation of the tetrahydrothiophene (32) observed on treatment of the mannitol-derived dixanthate (3 1) with TBTH has been explained in terms of the migration of a thiocarbonate moiety on to an alkyl radical followed a homolytic substitution at sulphur.43 Ph Ph The 1986 report (see Annual Reports B 1986 Volume 83) that the TBTH-promoted reductive rearrangement of the cholestanol derivative (33) to (34) proceeded without inversion of the acetoxy group oxygens has prompted further studies into 1,2-acyloxy migrations.Thus Sustmann and GieseU have looked at benzoyloxy migration in the carbohydrate field and concluded from an ‘*O labelling experiment that the migration takes place as had been previously thought with complete inversion of the ether and carbonyl oxygens via a five-membered transition state. However in a more complete study Beck~ith~~ noted that inversion of the two oxygens was not complete in the case of (33) -+ (34) and also that the rate of this particular migration was between 3 and 4 orders of magnitude faster than the analogous acyclic rearrange- ment. It was concluded that two mechanisms operate in 1,2-acyloxy migrations one 42 M.Tada M. Matsumoto and T. Nakamura Chem. Lett. 1988 199; J. Am. Chem. SOC.,1988 110,4647. 43 A. V. R. Rao K. A. Reddy M. L. Gurjar and A. C. Kunwar J. Chem. SOC.,Chem. Commun. 1988 1273; see also H. Sano T. Takeda and T. Migita Chem. Lett. 1988 119. 44 H. G. Korth R. Sustmann K. S. Groninger M. Leisung and B. Giese J. Org. Chem. 1988 53 4364. 45 A. L. J. Beckwith and P. J. Duggan J. Chem. SOC. Chem. Commun. 1988. 1000. 82 D. Crich Y (33) X=OAC; Y=p-Br (34) X=H;Y=cz-OAC (36) X = a -0OH (37) X=p-OOH involving a five-membered transition state and a second more rapid mechanism involving either a three-membered transition state or a tight ion pair. In the related area of the rearrangement of the allylic hydroperoxide (35) to (36) and (37) the Davies group conducted experiments under an atmosphere of '*OZ and found no incorporation into (36) but 80% incorporation into (37).46 It was therefore concluded that the initial rearrangement of the peroxyl radical derived from (35) to that derived from (36) takes place uia a five-membered transition state but that rearrangement of (36) to (37) requires dissociation of the intermediate peroxyl radicals.Radical Clocks and Probes.-Rate constants for the rearrangement of (38) +(39)47 and (40) +(41)48 were measured and found to be extremely rapid (1.4 x 10" s-' unspecified temperature and 1.1 x lo9s-' at 80 "C,respectively). Ph h phA I o=s---;r-o=s II 0 0 (41) 46 A. L. J.Beckwith A. G. Davies I. G. E. Davison A. Maccoll and M. H. Mruzek J. Chem. SOC.,Gem. Commun. 1988 475; D. V. Avila A. G. Davies and I. G. E. Davison J. Chem. SOC.,Perkin Trans. 2 1988 1847. 47 L. Mathew and J. Warkentin Can. J. Chem. 1988 66 11. 48 B. Vacher A. Samat A. Allouche A. Laknifli A. Baldy and M. Chanon .Tetrahedron 1988 44,2925. Reaction Mechanisms -Part (iii) Free-radical Reactions With the aid of the 0-alkyl thiohydroxamate (23; R = 4-pentenyl) Beckwith was able to estimate the rate of ring closure of the 4-pentenyloxyl radical to be 16 x lo8s-' at 80 0C.37 The controversy over the validity of cyclizable probes of the 5-hexenyl iodide variety for the elucidation of SET reaction steps continues and is summarized in two reviews published by the main protagonist^.^^ Although there are still many points of contention particularly regarding the presence or absence of impurities capable of initiating SET chain reactions and also reasons for the lack of cyclized products observed on reaction of capto-dative probe (42) with metal hydride reducing agents it is now widely agreed and appreciated by all that the simple observation of cyclic products from cyclizable radical probes especially iodides cannot be taken as evidence of an SET step.M. Newcomb and D. P. Curran Acc. Chem. Res, 1988 21 206; E. C. Ashby ibid. p. 414.