4 Free-radical Reactions and Electron Spin Resonance Spectroscopy By A. R. FORRESTER Department of Chemistry University of Aberdeen AB9 2UE In this year's Report the sections on free-radical chemistry and e.s.r. spectroscopy have been combined. In the main emphasis has been placed on new applications of the latter technique to the solution of chemical problems as wider coverage of the spectroscopic results will be given in a forthcoming Specialist Periodical Report.'" 1971 has also seen the retirement of Professors Hey and Waters whose enormous contribution to the field of free-radical chemistry has been commemo- rated by the publication of a most useful volumelb containing essays on a wide variety of aspects of free-radical chemistry. 1 Carbon Radicals 2-Methylpropenyl(l) exhibits a selectivity similar to that of phenyl in its reactions with benzylic hydrogen atoms (primary :secondary :tertiary ; 1.0:3.9 :8.6).In this respect its behaviour is intermediate between that of non-discriminating atomic chlorine and more-stable carbon radicals such as trichloromethyl.2" Decomposition of trans-and cis-a-phenylcinnamoyl peroxides and t-butyl tram-and cis-a-phenylperoxycinnamatesin carbon tetrachloride and other solvents has led to a reappraisal of the stereochemistry of the 1,2-diphenylvinyl radical thus formed.2b Although the results do not completely exclude the linear structure (2),previously advocated for this radical they can be accounted for best by an equilibrating pair of radicals [(3) (4)]which even at 80 "C,react with (a) 'Electron Spin Resonance,' ed.R. 0. C. Norman (Specialist Periodical Reports) The Chemical Society London vol. I in the press; (b) 'Essays on Free Radical Chemistry,' ed. R. 0. C. Norman Special Publication No. 24 The Chemical Society London 1970. (a) P. G. Webb and J. A. Kampmeier J. Amer. Chern. Soc. 1971,93 3730; (b)N. Wada K. Tokumaru and 0. Simamura Bull. Chem. SOC.Japan 1971,44 11 12. 187 A. R.Forrester solvent sufficiently rapidly to ensure a preponderance of that vinyl chloride which has the same stereochemistry as its peroxide precursor. Interconversion of cis-and trails-methoxypropenyl radicals is also slow relative to the rate at which they are trapped by reaction with cumene and it would seem that both the size and the electronegativity of substituents affect the rate of inversion of vinyl radical^.^ Stereochemistry of the products in the above reactions depends upon the rate at which the radicals are trapped and it is significant that addition of toluene-p-sulphonyl iodide to a series of alkynes gave,4 in almost all cases exclusively the 1:1 trans-adduct (9,which implies that the sulphonyl iodide is R SO,C,H,Me-p \/ I /c=c\* an extremely efficient chain-transfer reagent.4 Consideration5 of the known rate (k 2 3 xlo7s-') of inversion of vinyl radicals at -180 "C (extrapolated to ca.10' s-'for propenyl at 25 "C)appears to exclude the mediation of propenyl radicals in the production of 2,4-hexadienes (with retention of configuration at the olefinic bond) from cis- and trans- 1-propenylcopper and related organo- metallics.It was estimated that the rate constant for dimerization would have to be about 10l2mol-' s-' to produce this result and this value is much greater than known values for rate constants of radical-radical coupling reactions in solution. The reactions of vinyl radicals have been critically reviewed.6 Geometrical isomerization of substituted allyl radicals [(6a) S(6b)l in the gas phase has been demonstrated by the isolation of six isomeric dienes in R3 (64 (6b) proportions which suggest that a 'nearly equilibrated' mixture of epimeric allyl radicals was the so~rce.~ 1,l-Dichloroallyl radicals (7) also dimerize to give a mixture of tetrachlorohexadienes in which (8) predominates.' '(a) M.S. Liu S. Soloway D. K. Wedegaertner and J. A. Kampmeier J. Amer. Chem. Soc. 1971,93 3809; (6) R. C. Neuman and G. D. Holmes J. Amer. Chem. Soc. 1971 93 4242. W. E. Truce and G. C. Wolf J. Org. Chem. 1971 36 1727. 'G. M. Whitesides C. P. Casey and J. K. Krieger J. Amer. Chem. Soc. 1971 93 1379. 'L. A. Singer in 'Selective Organic Transformations,' ed. B. S. Thyagarajan Wiley New York vol. 11 p. 239. 'R.J. Crawford J. Hamelin and B. Strehlke J. Amer. Chem. Soc. 1971 93 3810. *W. R. Dolbier and C. A. Harmon Chem. Comm. 1971 150. 189 Free-radical Reactions and Electron Spin Resonance Spectroscopy (7) Further evidence that cyclopropyl radicals are not planar comes from the isolation of small quantities of optically active l-methyl-2,2-diphenylcyclopro-pane on decomposition of the optically active peroxide precursor in carbon tetrachloride.' The product arises by rapid disproportionation within the solvent cage.Additionally a radical-pair mechanism has been advanced to account for the formation of a series of 2,2-diphenyl- 1-substituted cyclopropanes (in which there is 83-86 % retention of configuration) by decarboxylation of the corresponding aldehydes with tris(tripheny1phosphine)rhodium chloride.' Although interconversion of the exo- and endo-isomers [(9a) and (9b)I was more rapid than their reactions with either toluene or di-isopropylbenzene and the same ratio of exo- to endo-chlorohydrocarbon was produced irrespective of the stereochemistry of the radical precursor,' ' complete equilibration of the 7-chloronorcaranyl radicals [( 10a) (lob)] was not achieved before their reduction with triphenyltin hydride.lZ Unlike a-fluorocyclopropyl radicals (1l) a-fluoro-cyclobutyl radicals (12) (if they are indeed non-planar) invert faster than they can be reduced by triphenyltin hydride. CND0/2 calculations predict that the barriers to inversion of the radicals (1 1H14) are 10.5,1.9,4.0 and 0.8 kcal mol-' respectively.' Rearrangement of cyclopropyl radicals to allyl radicals depends upon the stability of the latter. Thus 2,2-and 1,2-diphenylcyclopropyl radicals (15 ; R' = Ph R2 = H; R' = H R2 = Ph) gave the corresponding resonance-stabilized allyl radicals (16) and thence the dimers (17) (30% yield) whereas phenylcyclopropyl radicals (15; R' = R2 = H) under the same conditions ' H.M. Walborsky and J.-C.Chen J. Atner. Chern. Soc. 1971 93 671. lo H. M. Walborsky and L. E. Allen J. Amer. Chem. Soc. 1971 93 5465. L. A. Singer and J. Chen Tetrahedron Letters 1971 939. l2 L. J. Altmann and R. C. Baldwin Tetrahedron Letters 1971 2531. A. R.Forrester merely gave phenylcyclopropane. ' By comparison for rigid cyclopropyl- carbinyls or cyclopropylcarbinyls with a strong conformational preference as in the cholestan-6-yls (18a) and (1 8b) rearrangement may be stereoelectronically controlled. In the rearrangement of (18a) and (18b) to (19) and (20) respectively the bonds which are broken are those which overlap most closely with the orbital of the unpaired electron.l4 Another instance' in which this principal may well apply is the preferred rearrangement of the radical (21) to the primary homoallylic radical (22) rather than to the more stable secondary radical (23). A more complex example is the apparently equilibrating mixture of radicals [(24H28)] derived from bullvalene which gave rise to a series of trienes but l3 J. C. Chen Tetrahedron Letters 1971 3669. l4 A. L. J. Beckwith and G. Phillipou Chem. Comm. 1971 658. l5 E. C. Friedrich and R. L. Holmstead J. Org. Chem. 1971,36 971. Free-radical Reactions and Electron Spin Resonance Spectroscopy 191 here prediction of products is compounded by uncertainties in the stereochemistry of the intermediates. l6 (27) (28) Factors which affect the norbornenyl-nortripticyl radical equilibrium [(29a) (29b)+(29c)] and the nature and stereochemistry of the products that they give rise to has greatly intrigued chemists in the past and last year was no exception.Results obtained by reduction of 3-acetoxynorborn-2-en-5-ylmer-cury(I1) chloride and related acetoxy-mercurials with sodium boro-deuteride (294 (29b) (294 and -hydride strongly suggest that hydrogen (deuterium) abstraction by (29a and 29c; R = OAc) is highly stereoselective but is not so for (29b; R = OAc).17 However reduction of the radical (29b; R = OAc) (in modest yield) by reaction with sodium-naphthalene is faster than skeletal rearrangement to (29a or 29c ; R = OAc) and hence may proceed in a stereospecific manner.'* Products formed by free-radical addition of trimethyltin hydride to norbornadiene have been accounted for" by consideration of the likely modes of approach of the trimethyltin radical (endoor exo)and of the steric effects of the bulky trimethyltin group on the intermediate equilibrating radicals corresponding to (29).Further evidence that 'non-classical' radicals do not intervene in these reductions has been obtained2' from a study of the addition of thiophenol to 3-methylenenor- tricyclene (30). In a series of papers dealing with reactions of polychlorinated H.-P. Loffler Chem. Ber. 1971 104 1981. G. A. Gray W. R. Jackson and V. M. A. Chambers J. Chern. SOC.(C),1971 200. " T. C. Morrill and F. L. Vandemark Terrahedron Letters 1971 181 I. l9 H. G.Kuivila J. D. Kennedy R. Y.Tien I. J. Tyminski F. L. Pelczar and 0.R. Khan J. Org. Chem. 1971 36 2083. 2o S. J. Cristol and R. Kellman J. Org. Chem. 1971 36 1866. A. R.Forrester norbornenyl radicals generated in several ways but generally by standard free- radical procedures Davies and his colleagues2 ’ have shown that these radicals undergo rearrangements as indicated for (29),and that two of the most important A factors which decide the composition of the product mixtures are the tendency for rearrangement from unstabilized to chlorine-stabilized radicals and the effect \ (,tH -+ \kc,) / which the relatively bulky chlorine substituents have on chain-transfer reactions of the equilibrating radicals. The reversibility of intramolecular homolytic alkylation in which the spiro- tetraenyl (32) intermediate is implicated has been elegantly demonstrated22 by isolation of mixtures of (34) and (35) on decomposition of the acyl peroxide precursors of both (31) and (33).For other pairs of radicals as well e.g. (36) and 1 1 (34) (35) ” D. 1. Davies and P. Mason J. Chem. SOC.(C) 1971 288 295; R. Alexander and D. I. Davies J. Chern. SOC.(C),1971 896; R. Alexander D. I. Davies D. H. Hey and J. N. Done J. Chem. Sor. (C) 1971 2367. 22 J. C. Chottard and M. Julia Tetrahedron Letters 1971 2561; see also M. Julia and B. Malassine Tetrahedron Letters 1971 987. Free-radical Reactions and Electron Spin Resonance Spectroscopy (37),23 and (38) and (39),24 there is good reason to believe (from product analysis) that equilibration with an open-chain radical occurs but the evidence at present is less conclusive than that given for (31) (32)C(33).By comparison i'l: \ H CH2I \O 0 / the diphenylcyclohexadienyl radical (40) does not rearrange but fragments as shown only at relatively high temperature^.^ 5a This observation indicates that homolytic arylation is not an appreciably reversible process under the usual conditions,25b and the suggestion has been made that in cases where reversibility has been observed under mild conditions it has been due to special structural or other features of the particular system. Although para-electron-withdrawing substituents facilitate intramolecular alkylation with rearrangement of the radical (41; R = C1) to the phenanthrene (42; R = C1) the effect is relatively slight and consequently the derived relative rates of migration provide little conclusive information on the nature of the transition state leading to rearrangement.26 The nucleophilic character of 23 D.H. Hey Quarf.Rev. 1971 25 483; D. H. Hey G. H. Jones and M. J. Perkins J. Chem. Soc. (C) 1971 116. 24 P. S. Dewar A. R. Forrester and R. H. Thomson J. Chem. SOC.(C) 1971 3950. *' (a) M. J. Perkins and P. Ward Tetrahedron Lerters 1971 2379; D. J. Atkinson M. J. Perkins and P. Ward J. Chem. Sor. (C),1971,3240; (b)J. Saltiel and H. C. Curtis J. Amer. Chem. SOC.,1971 93 2056. 26 P. N. Cote and B. M. Vittimberga J. Amer. Chem. Soc. 1971 93 276. A. R.Forrester alkyl radicals is clearly shown by the high yields of mono- and di-substituted products obtained on methylati~n,~’ of benzylation,’* and cyclohe~ylation~~ protonated pyridines quinolines and related bases3’ Displacement of the nitro-group of a series of nitrofurans by methyl generated as shown from ferrous ions and hydrogen peroxide in dimethyl sulphoxide occurs under unusually mild condition^:^' Fe2+ + H202 + Fe(OH)2++ -OH .OH + MeSMe -+ --+ Me.+ MeS0,H II 0 The phenylfluorenyl radical (43) dimerizes to give an ethane (whose structure was confirmed by 3Cn.m.r. spectro~copy),~~ but not so the bis(t-butylpheny1)- methyl radical (44) [a,(CH) = 1.818,a(o-H) = 0.413,a(m-H) = 0.153 and a(p-H) = 0.366 mT] which like triphenylmethyl gives a 2,5-cyclohexadiene dimer.3 The dimer of triphenylmethyl may be converted into (49,even on treatment with weak acids such as E.s.r.evidence has been presented35 in support of 27 F. Minisci K. Bernardi F. Bertini R. Galli and M. Perchinummo Tetrahedron 1971 3575 H. J.-M. Dou G. Vernin and J. Metzger Bull. SOC. chim. France 1971 1021. 28 H. J.-M. Dou G. Vernin M. Dufour and J. Metzger Bull. SOC. chim. France 1971 111. 29 H. J.-M. Dou G. Vernin and J. Metzger Bull. SOC. chim. France 1971 3553. 30 M. Baule G. Vernin H. J.-M. Dou and J. Metzger Bull. SOC. chim. France 1971 208 3. 31 U. Rudqvist and K. Torssell Acta Chem. Scand. 1971 25 2183. 32 H. A. Staab K. S. Rao and H. Brunner Chem. Ber. 1971 104 2634. ” F. Bolsing and K.-D. Korn Tetrahedron Letters 1971 3865.34 H. Takeuchi T. Nagai and N. Tokura Bull. Chem. SOC. Japan 1971 44 753. 35 D. Braun U. Platzek and H. J. Hefter Chem. Ber. 1971 104 2581. Free-radical Reactions and Electron Spin Resonance Spectroscopy the proposal that the polymer formed on treatment of p-tolyldiphenylmethyl chloride with pyridine arises by addition of p-tolyldiphenylmethyl radicals (47) to the dimethide (46). The growing polymer radical is easily detected by e.s.r. and its spectrum is generally similar to that of (47). Highly conjugated molecules e.g. Tschitschibabin’s hydrocarbon (48) which may be formulated as either singlets or triplets usually have singlet ground states but in solution form mono- radicals by hydrogen abstraction from the solvent by the thermally populated triplet Such is the case with a number of nitrones isatogens and bian- throne.36b However for Schlenk’s hydrocarbon (49) no conventional singlet structure can be formulated and this hydrocarbon has now been shown36c to Ph Ph Ph Ph I I I I Ph-q Ph-q uc-ph exist as the triplet in the solid state (D = 192 MHz E = 17 MHz); in solution in toluene it behaves more like (48),forming two monoradicals one of which is (50).Similar con~lusions~~ have been reached for the diradical (51). Perchlorinated triphenylmethyls e.g. (52) are much more stable than their hydrocarbon ana- logues and not only withstand high temperatures and do not dimerize but are ’‘ (a)G. R. Luckhurst G. F. Pedulli and M. Tiecco J. Chem. SOC.(B) 1971 329 and refs.therein; (6)L. S. Singer I. C. Lewis T. Richerzhagen and G. Vincow J. Phys. Chem. 1971 75 290; (c) H.-D. Brauer Chem. Ber. 1971 104 909. ’’ R. Schmidt and H.-D. Brauer Angew. Chem. internat. Edn. 1971 10 506. A. R. Forrester inert to corrosive reagents such as concentrated nitric or sulphuric acids powdered sodium hydroxide bromine or chlorine.38 The principal stabilizing factor is the protection which the ortho-chlorine atoms afford the central carbon prohibiting bond formation at this centre. However this protection does not prevent electron-transfer reactions of the radical (or the corresponding carbonium ion39) such as that which occurs with hydroxide ion in dimethyl sulphoxide nor does it make the radical insensitive to light.On irradiati~n,~' the very stable CI CI CI clQcl CI CI' CI (54) fluorenyl(53) is formed which can be more conveniently obtained by thermolysis of (52) at 25&320°C [cf triphenylmethyl which gives the fluorenyl (43) on irradiation but this dimerizes rapidly]. The e.s.r. spectra of (52) and related radicals (g z 2.0026) show splittings attributed to a(ci-13C)= 3.0mT and a(Ar-13C)= 1.0mT; the chlorodiaryl radical (54) has g = 2.0055 and a(a-Cl) = 0.22mT. An X-ray analysis of an impure crystal of the latter radical revealed that the rings are twisted by 47" and 43" with respect to the trigonal plane of the central carbon4' There is an increasing awareness aided by three reviews,42 of the importance of hydrocarbon radical-anions as reactive intermediates and of their synthetic 38 M.Ballester J. Riera J. Castaner C. Badia and J. M. Monso J. Amer. Chem. SOC. 1971,93,2215. 3y M.. Ballester J. Riera-Figueras J. Castaner and A. Rodriguez-Siurana Tetrahedron Letters 1971 2079. 40 M. Ballester J. Castaner and J. Pujadas Terraheriron Letters 1971 1699. 41 J. Silverman L. J. Soltzberg N. F. Yannoni and A. P. Krukonis J. Phys. Chem. 1971 75,1246. 42 N. L. Holy and J. D. Marcum Angew. Chetn. Internat. Edn. 1971,10 115; L. L. Miller J. Chem. Ediic. 1971 48 168; L. Eberson and H. Schafer Fortschr. Cheni. Forsch. 1971 21 5. Free-radical Reactions and Electron Spin Resonance Spectroscopy usefulness. Thus 3-nitroperylene may be conveniently obtained (66% yield) by treatment of the radical-cation of perylene with nitrite ion whereas its preparation by direct nitration of perylene is tedious.43 Adaptation of this procedure to the nitration of other hydrocarbon radical-ions is under investigation.The heat of dimerization of the perylene radical-cation (8.8 kcal mol- ') has been evaluated from electronic absorption meas~rements.~~ Reaction of the radical-anion of naphthalene (55) with aryl halides in THF gave mainly benzene and products derived therefrom by reaction with phenyl radicals generated as shown [equation (l)]."' A preparation of aryltrimethyl- and benzyltrimethyl-silanes has been devised46 in which a mixture of the benzyl or aryl halide and trimethylchlorosilane is treated with (55). However triphenylchlorosilane reacted preferentially with (55) to give he~aphenyldisilane."~ The radical-anions of [2,2]cyclophanes e.g.(56),and diarylalkanes e.g. (57),are much less stable than those of their simple mononuclear aromatic counterparts disproportionating at ca. -60 "C to the parent hydrocarbon and dianions ; the latter subsequently fragment to two monoanions [equation (2)] easily trapped by pr~tonation.~~ The electron- exchange rates measured by e.s.r. between the two n-systems of the diarylalkanes PhCH2CH,Ph2--* 2PhCH,--* 2PhCH (2) [lo6-lo8 s-for (57 ;Ar = 1-naphthyl)] are predictably much smaller than those of the [2,2]paracyclophanes (> lo8 s-') in which there is considerable n-n overlap because of the constrained geometry of these molecules.49 Similarity in the e.s.r.spectra of the radical-anions of t-butylbenzene and the bicyclo[2,2,2]- '' C. V. Ristagno and H. J. Shine J. Amer. Chem. Soc. 197 1 93 18 11. 44 K. Kimura T. Yamazaki and S. Katsumata J. Phys. Chem. 1971 75 1768. 45 T. C. Cheng L. Headley and A. F. Halasa J. Amer. Chem. Sac. 1971 93 1502. 46 S. Bank and J. F. Bank Tetrahedron Letters 1971 458 1 see also H. Sakurai A. Okada M. Kira and K. Yonezawa Tetrahedron Letters 197 1 15 1 1. " F. W. G. Fearon and J. C. Young J. Chem. Soc. (B),1971,272. '' J. M. Pearson D. J. Williams and M. Levy J. Amer. Chetn. Sac. 1971 93 5478. '' D. J. Williams J. M. Pearson and M. Levy J. Amer. Chem. Soc. 1971 93 5483. A. R.Forrester octane (58) obviates the possibility that there is electron transfer between the two n-systems via the bicyclo-octane skeleton.However this non-conjugated bridg- ing group does permit intramolecular transmission of singlet excitation in (59) COPh from naphthyl to benzoyl and triplet excitation from benzoyl to na~hthyl.~' The hyperfine splitting uH= 0.23 mT in the seven-line e.s.r. spectrum of the triptycene radical-cation has been attributed to interaction with the six protons indicated in (60),these being rendered magnetically equivalent by electronic 11 t I I H transannular interaction. Among several other interesting e.s.r. studies of aromatic radical-anions reported this year is that in which the coupling constants of the annulene (61)are compared with those of the radical-anion of naphthalene (55) over a wide temperature range.52 The a(&-H) values (0.24mT) for (61) are much smaller than those of (55) (0.49mT) because of deviations from planarity and are [unlike those of (55)] very sensitive to the small changes in geometry that a change in temperature inevitably produces in a non-planar molecule.The e.s.r. spectrum of the radical-anion produced from the atropisomers of the novel hexaphenylene (62) shows53 two different types of coupling a(1-H) = 0.082 and 42-H) = 0.033 mT. 5" H. E. Zimmerman and R. D. McKelvey J. Amer. Chem. Soc. 1971,93 3638. " R. M. Dessau J. Chem. Phys. 1971 54 5430. 52 F. Gerson K. Miillen and E. Vogel Helv. Chim. Acra 1971 54 2731. '' G. Wittig and D. Riimpler Annalen 1971 751 1. Free-radical Reactions and Electron Spin Resonance Spectroscopy 199 The most obvious mechanism for the radical rearrangement [(63) j (65) via (64)] has been discounted54 because the intermediate radical (64) could not be detected by e.s.r.in a flow-system whereas the initial and final radicals [(63) and 0 Me II I C-Me '5 0' -+ I 0I -+ MeCHCH,OCOMe Me AH-kH MeCH-CH (65) (65)] could. The alternative routes (cage process or three-membered-ring transi- tion state) seem unattractive possibilities but experiments with *O-labelled acetate are clearly required. The e.s.r. spectra of (64) (generated by hydrogen abstraction from the corresponding cyclic acetal) and those of a large number of related cyclic and acyclic radicals in which the tervalent carbon is linked to one or two oxygen atoms have been reported at some length and the considerable amount of information which may be derived from such spectra about the ge- ometry of the radicals has been ernphasi~ed.~~9'~ The conclusion of these reports is that the introduction of oxygen substituents adjacent to the tervalent carbon and/or a reduction in the size of the ring containing the tervalent carbon tend to increase the degree of bending and hence the s-character of the half-filled orbital at the radical centre.However use of absolute values of u(a-l3C)and u(a-H)to gauge the degree of bending can be misleading because the signs of these values may change as the mechanism of spin transmission varies with the geometry. Oxyalkyl radicals also participate in the reaction of enols and enol ethers with hydroxyl radicals.56 Thus the adduct radicals (66) [a(cc-H)= 1.7 a(P-H) = 0.86 ..q.n-30 + + H,O MeOCH=CH + -OH * MeO-CH-CH .-MeO=CHkH * MeO-CH-CH, I jb OH H+ and adOMe) = 0.18 mT] and (68) [a(a-H)= 2.24 and a(P-H) = 1.89 mT] were detected when hydroxyl radicals were allowed to react with methyl vinyl ether in a flow system and are considered to undergo acid-catalysed intercon- version via (67).Attempts' to detect aminoalkyl radicals directly during the reaction of amines with hydroxyl radicals (from H202-Ti1") in a flow system were unsuccessful but their presence was established by 'spin trapping' experiments with nitroso-t- butane. Ease of oxidation by hydrogen peroxide [equation (3)] was suggested 54 A. L. J.Beckwith and P. K. Tindal Austral. J. Chrm. 1971 24 2099. " A. J. Dobbs B. C. Gilbert and R. 0. C. Norman J. Chem. SOC.(A) 1971 124. D. J. Edge B. C. Gilbert R. 0. C. Norman and P. R. West J. Chem. SOC.(B),1971 189. 57 N. H. Anderson and R. 0. C. Norman J. Chrm. SOC.(B),1971,993. 200 A. R.Forrester to account for this failure. Such radicals have been detected after irradiation of for example ethylamine trapped in a crystal of adamantane and are considered to be intermediates in the formation of iminyls (R'R2C=N.) in this way (vide infr~).~~" For the radical (69; R' = Me R2 = H) aN= 0.48 a,(NH) = 0.50 H202 + R'R'CNH -R'R2C=&H +.OH + -OH (3) (69) and a(a-H) = a(,$-H) = 1.48 mT. They have also been detected after irradiation of aqueous solutions of amines and amino-acids with high-energy electrons in the cavity of the ~pectrorneter.~~' Phenylthioalkyl radicals generated in a number of way^,^^,^' gave an array of products most of which may be accounted for by the usual dimerization disproportionation and radical-coupling processes.However formation of diphenyl disulphide as a principal product ( -30 % yield) of the electrolysis of PhSCH,CO,-Na' PhSCHzC03Bu' (70) (71) the salt (70) and thermolysis of the ester (71) requires an explanation. Fragmenta- tion of the thioalkyl radical (72) to methylene and thiophenoxyl has been pro- posed but convincing evidence for the mediation of substantial amounts of PhSCH -+ PhS. + :CH2 -+ PhSSPh (72) methylene is lacking at present." Rate constants for the thermal decomposition of t-butyl a-arylthioperacetates and a-aryloxyperacetates in ethylbenzene were similar ( -4.0-5.0 x lo4s-') electron-donating puru-substituents increasing the rate in each case.Formation of diphenyl disulphide in the decomposition of the former esters was not recorded.60b Thermolysis of the ethane (73)in xylene at cu. 100 "Cgave a radical which was easily detected by e.s.r. but whose spectrum showed no fine structure. Identification of this species as tris(pheny1thio)methyl (74) has now been achieved by '3C isotopic labelling.61 The large a(a-' 3C) value observed (4.37 mT) suggests that (74) is considerably bent. (PhS),C-C(SPh) --* 2 (PhS),C* (73) (74) 58 (a)D. E. Wood and R. V. Lloyd J. Chem. Phys. 1970,52 3840; T.Richerzhagen and D. H. Volman J. Amer. Chem. SOC.,1971 93 2062; (b)P. Neta and R. W. Fessenden J. Phys. Chem. 1971 75 738. 5y K. Uneyama S. Torii and S.Oae Bull. Chem. SOC.Japan 1971 44 815. 6o (a) A. Ohno N. Kito and Y. Ohnishi Bull. Chem. SOC.Japan 1971 44 463 467; A. Ohno and Y. Ohnishi Tetrahedron Letters 1969,4405; (6)C. Ruchardt and H. Bock Chem. Ber. 1971 104 577 CJ C. Riichardt and I. Mayer-Ruthardt Chem. Ber. 1971 104 593. 61 A. K. Beck D. Seebach and H. B. Stegmann Angew. Chenz. Internat. Edn. 1971 10. 500. Free-radical Reactions and Electron Spin Resonance Spectroscopy 20 1 Although the relative stabilities (based on their ease of formation by hydrogen abstraction from the parent compound62 and on other chemical evidence6') of the radicals (75H78) is that shown the lifetimes of the e.s.r.signals of a number of related radicals generated by thermolysis of the corresponding azo-compounds Me ArCH < ArOkH < ArSeH < Ar?!4CH2 (75) (76) (77) (78) did not bear out this order.63 However in such comparisons it is important that a clear distinction be made between thermodynamic and kinetic stability due account being taken of the circumstances under which the radicals are generated (see for example ref. 64). 2 Nitrogen Radicals Esr. spectra of the aziridinyl and azetidinyl radicals have been obtained by measuring irradiated solutions of the corresponding amines in cyclopropane containing di-t-butyl peroxide.6s The coupling parameters uN = 1.25 and 1.4 and a(a-H)= 3.07 and 3.83mT and g-values 2.0043 and 2.0045 respectively confirm that in these radicals the unpaired electron is in the 2p orbital of the nitrogen as depicted in (79).A kinetic study of the decomposition of dimethyl- aminyl and di-isopropropylaminyl has led to the prediction (supported by pre- liminary results) that di-t-alkylaminyls should be 'stable' in the absence of solvents with readily abstractable hydrogen atoms.66 N-Alkoxy-N-alkylaminyls (81) are formed on irradiation of t-butyl peroxycarbamates [equation (4)] or directly from (80) by photolysis in the presence of di-t-butyl peroxide. The alkoxy-group (79) is effective in removing spin density from the nitrogen atom since for (81; R = Me) aN= 1.45 and a(a-H) = 2.15 mT (cj dimethylaminyl which has a(a-H) = 2.74 mT).67 N-Alkoxy-N-arylaminyls (83) generated68 by addition of relatively bulky alkyl groups (R ') to 2,4,6-tri-t-butylnitrosobenzene (82) predict-ably have smaller aNvalues (-1.1 mT).These aminyls react further with bulky RNHCOOBU~--+ RNH + coz+ .OBU~+ RNHOBU' 3 RNOB~' (4) II 0 (80) (81) '" K. Uneyama H. Namba and S. Oae Bull. Chem. SOC.Japan 1968 41 1928. O3 A. Ohno N. Kito and Y. Ohnishi Bull. Chem. SOC.Japan 1971,44,470. " R. S. Davidson and P. R. Steiner J. Chem. SOC.(C) 1971 1682. h5 W. C. Danen and T. T. Kensler Tetrahedron Letters 1971 2247. " J. R. Roberts and K. U. Ingold J. Amer. Chem. SOC.,1971,93 6687. " W. C. Danen and C. T. West J. Amer. Chem. SOC..1971 93 5582. '* S. Terabe and R. Konaka J.Amer. Chem. SOC.,1971,93,4306. A. R. Forrester t alkyl groups to give the cyclohexadienes (84).6g Since relatively small alkyl radi- cals (R2)add to the nitrogen of (82) giving a nitroxide (85) and (83) and (85) are easily distinguished the nitroso-compound (82) is a novel 'spin trap' for dif- ferentiating between 'large' and 'small' alkyl radicals. A well-resolved e.s.r. spectrum of diphenylaminyl has now been re~orded.~' This shows aN= 0.89 a(o-H) = 0.37 a(m-H) = 0.15 and a(p-H) = 0.43mT from which it may be deduced that the unpaired electron is delocalized to the extent of about 60 on to the phenyl groups. Protonated diphenylaminyl gives a remarkably similar ~pectrum.~ 'Diphenylaminyl radicals generated by thermo- lysis of diphenylnitrosamine in the absence of air coupled N to ortho-C and N to para-C to give semidines and structurally related polymeric material.In the presence of air p-nitrodiphenylamine was a main pr~duct.~ para-Substituted diphenylaminyls (but not di-p-anisylaminyl) gave mainly the parent amines and diaryldihydrophenazine~.~~ Crystalline diarylaminyls containing germanium e.g. (86) have been prepared by treatment of the corresponding tin or lead com-pounds with diarylgermanium halides. Hyperfine structure (a = 0.3 mT) due to coupling with 73Ge (I = 9/2) was detected in the spectra of these radicals at high spectrometer gain.73 The phenoxazinyl (87) exists to the extent of 85 % as the radical but its e.s.r. spectrum has defied interpretati~n.~~ Position of the equilibrium [2 (88) (89)l is determined to some extent by substituents in the Ar' Ar2 and Ar3 rings.ortho-Substituents in the Ar' ring have most effect 69 T. Hosogai N. Inamoto and R. Okazaki J. Chem. SOC.(0,1971 3399. 'O F. A. Neugebauer and S. Bamberger Angew. Chem. Znternat. Edn. 1971 10 71. " P. Welzel Chem. Ber. 1971 104 808. 72 F. A. Neugebauer and H. Fischer Chem. Ber. 1971 104 886. 73 H. B. Stegmann K. Schemer and F. Stocker Angew. Chem. Internat. Edn. 1971 10 499. 74 J. Brandt G. Fauth W. H. Franke and M. Zander Chem. Ber. 1971 104 519. Free-radical Reactions and Electron Spin Resonance Spectroscopy Ph// Ph (86) (87) greatly destabilizing the monomer relative to the dimer.7 Triarylimidazyls (88) are relatively mild oxidizing agents the most reactive being those with bulky electron-withdrawing groups in the ortho-positions of the Ar' ring.76 Oxidation may proceed either by hydrogen (e.g.from phenols and thiols) or by electron ArZ 2 Ar3 Ar' (89) (e.g.from t-amines) removal. 77 Combination of improved e.s.r. measurements on partially and completely deuteriated triphenylimidazyl (88 ; Ar' = Ar2 = Ar3 = Ph) and simple MO calculations has led to the assignment of all the splittings in its spectrum aN= 0.144 a(o-H) = 0.137 a(m-H) = 0.053 and a(p-H) = 0.151 mT (for Ar2 and Ar3 rings) and a(o-H) = 0.24 a(m-H)= 0.089 and a(p-H) = 0.288 mT (for Ar' ring).78 The assignment79 of the spectrum obtained on y-irradiation of malonamide at 77 K (aN,,= 3.6 aNL= 0.7 aH= 8.0mT) to the o-radical (89a) has been ques- tioned" because of the relatively small aNvalue and the alternative suggestion was made that NH is also present in such systems and that it gives rise to some of the lines in the spectrum attributed to (89a).In support of this view irradiation (89a) (90) 75 L. A. Cescon G. R. Coraor R. Dessauer E. F. Silversmith and E. J. Urban J. Org. Chem. 1971,36 2262. 76 R. L. Cohen J. Org. Chem. 1971 36 2280. " L. A. Cescon G. R. Coraor R. Dessauer A. S. Deutsch H. L. Jackson A. MacLachlan K. Marcali E. M. Potraflce R. E. Read E. F. Silversmith and E. J. Urban J. Org. Chem. 1971 36 2267; A. MacLachlan and R. H. Riem J. Org. Chem. 1971 36 2275. N. Cyr M. A. J. Wilks and M. R. Willis J. Chem. SOC.(B) 1971 404.79 N. Cyr and W. C. Lin J. Chem. Phys. 1969 50 3701. 8o M. C. R. Symons J. Chem. Phys. 1971 55 1493; J. Chem. Soc. (A) 1971 3205. A. R.Forrester of urea at 77 K with 'OCo y-rays has been shown to give a radical with spectral parameters deemed8' to be those expected of the n-type amidyl (90). Chemically amidyls show a strong but not always exclusive preference for reaction on nitrogen e.g. the NO-diacylhydroxamic acid (91) on photolysis gave mainly the amide (92) (15%) but also a trace of the ester (93).82 However photolysis of N-chloro- and N-bromo-acetanilides in the presence of an alkene gave a mixture of cis-and trans-1,2-adducts (both coupled tlia N) in good yield.83 When carbon PhCONHCHMeEt 0 IIPhCONHOCCHMeEt -+ PhCONH -t C02 + CHMeEt /" (92) (91) \i NHII PhCOCHMeEt 1 0 I1 Ph COCH MeEt tetrachloride was used as solvent in the reaction of N-bromoacetamide with cyclohexene the main product was the bromoacetimidate (94) but this is formed by an ionic process.84 A long-held misconception that N-bromoacetamides I Me behave like N-bromosuccinimide under Wohl-Ziegler conditions has thus been removed.In the Reporter's view the chemical and spectroscopic information available at present does not allow an unequivocal classification of amidyls as 0-or n-radicals. Synthetic applications of the re?ctions of aminyls amidyls (including sulphonamidyls) and ammoniumyls (R,NH) have been reviewed.85 Addition of alkyl acyl or aryl radicals to nitrile N-oxides with formation of stable iminoxyls provides another method by which reactive radicals may be 'spin trapped'.86" 'Spin adducts' obtained with acyl radicals give e.s.r.spectra in which splittings due to !he protons in the acyl group are distinguishable. An imminoxyl [(Bu'),C=NO] which is sufficiently stable to be isolated has been reported.86b Iminyls (RRC=N-) previously generateds8 by irradiation of amines H. Bower J. McRae and M. C. R. Symons J. Chein. Soc. (A) 1971 2400. B. Danieli P. Manitto and G. Russo Chem. and Ind. 1971 203. D. Touchard and J. Lessard Tetrahedron Letters 1971 4425. 84 S. Wolfe and D. V. C. Awang Cunad. J. Chetn. 1971 49 1384. R. S. Neale Synthesis 1971 1. 86 (a) B. C. Gilbert V. Malatesta and R. 0.C. Norman J. Amer. Chem. Soc. 1971 93 3290; (b)J.L. Brokenshire G. D. Mendenhall and K. U. Ingold J. Amer. Chem. SOC. I97 I 93 5278. Free-radical Reactions and Electron Spin Resonance Spectroscopy 20 5 are thought to be the intermediates in the phot~lysis~~ and thermolysis8* of azines and in the rearrangement of oxime thionocarbamates to thioxime carba-mates [equation (5)].89 R R The rapid growth in the number of publications on nitroxides makes the reviews dealing with spin trapping,” stereochemistry of nitroxides,’ electron resonance of bisnitroxides in anisotropic reactions of N-nitroso-a~etaniIide,~~’and a general review93most welcome. A claim94that the aziridinyl nitroxide (95) is formed on decomposition of the dihydroxylamine (96) has been refuted95 and the product (which is difficult to separate from small quantities of paramagnetic impurities) assigned structure (97) mainly on the basis of its reactions with bifunctional acylating agents.However in a very recent report96 the product has been shown to be identical with acetoxime. Over-reliance on mass spectral data seems to have provided the root for this confusion. Crystallo-graphic analysis of the bicyclic nitroxide (98) has revealed” that like Fremy’s Me Me Me Me NHOH k-i Me N Me NHOH Me (96) \ (97) Me,C=NOH \/ N--O EtN-0 I ONEt I I +- 1.289A CH .:.‘ O--N 7/ / Me \ H’ 2.278k ‘ (99) ” M. Kamachi K. Kuwata and S. Murahashi J. Phys. Chem. 1971 75,164. ” W. J. Middleton J. Amer. Chem. SOC.,1971 93 423.‘9 R. F. Hudson A. J. Lawson and E. A. C. Lucken Chem. Comm. 1971 807. 90 M. J. Perkins ref. 1(6) p. 97; E. G. Janzen Accounts Chem. Res. 1971 4 31. 91 E. G. Janzen in ‘Topics in Stereochemistry,’ ed. N. L. Allinger and E. L. Eliel Wiley New York 1971 vol. 6 p. 177; A. Rassat Pure Appl. Chem. 1971 25,623. 92 (a)G. R. Luckhurst Roy. Znst. Chem. Rec. 1970 3 61 ;(6) J. I. G. Cadogan ref. I(h) p. 71 ; Accounts Chem. Res. 1971 4 186. 43 E. G. Rozantsev and V. D. Sholle Russ. Chem. Reti. 1971,40 233. ’‘ G. R. Luckhurst and F. Sundholm Tetrahedron Letters 1971 675. 9s P. Singh D. G. B. Boocock and E. F. Ullman Tetrahedron Letters 1971 3935. 9h J. F. W. Keana R. J. Dinerstein and D. P. Dolata Tetrahedron Letters 1972 119. ’’ A. Capiomont B. Chion and J. Lajzerowicz Arta Cr-vst.1971 B27,322. A. R. Forrester salt [(KSO,),NO.] 'dimers'98 are formed in the solid state in which the nitrogen and oxygen atoms of molecular pairs are 2.28A apart (99). This structure contrasts with the dimeric transition state (100) which has been proposed9' for the disproportionation of diethyl nitroxide. Reaction of bistrifluoromethyl nitroxide with alkanes gave the corresponding alkenes as intermediates which reacted further with the radical to give 2 1 (nitroxide :alkene) ad duct^.^* With alkynes the initial 2 1 adduct gave rise to a number of products including the corresponding 1,2-diketones. loo Of the numerous reports on the use of spin traps perhaps the most significant is that which describes the trapping of the penta- cyanocobaltate(I1) anion [-Co(CN),l3 -,with nitrosobenzene.The organometallic nitroxide thus formed [aN= 1.38 a(o-,p-H) = 0.32 a(m-H) = O,ll and aco = 1.06 mT] has been compared both with alkyl aryl nitroxides and nitroaromatic radical-anions. lo' Several new spin labels have been synthesized the steroidal nitroxide (101 ;R' = P-OH R2 = R3 = CH,CH=CH,) in which the nitroxide group is incorporated into the molecular skeleton being especially novel.'02 0-I (101) Much publicitylo3 has been given to a new nitroxide spin label (structure un- specified) which may be utilized in the assay of morphine and other opiates in urine. The method seems to depend upon the drug releasing nitroxide spin labels bound to antibodies the concentration of the released spin labels then being measured by e.s.r.The sensitivity of the method is claimed to be 1000 times greater than that of comparable methods using thin layer chromatography. 3 Oxygen Radicals Reviews surveying reactions of oxygen radicals with metal ions,' O4 hydrogen abstraction from O-H bonds lo5 reactions of peroxides with phosphates sul- phides and amines,'06 synthetic applications of the hypoiodite reaction lo7 and 98 K.E. Banks R. N. Haszeldine and B. Justin J. Chem. Soc. (C),1971 2777. y9 K. Adamic D. F. Bowman T. Gillan and K. U. Ingold J. Amer. Chetn. Soc. 1971 93 902. loo R. E. Banks R. N. Haszeldine and T. Myerscough J. Chem. SOC.(C),1971 1951. lo' M. G. Swanwick and W. A. Waters J. Chem. SOC.(B) 1971 1059. lo' R. Ramasseul and A.Rassat Tetrahedron Letters 1971 4623. Varian Instrument Applications 1971 vol. 5 no. 3. Io4 E. T. Denisov Russ. Chem. Rev. 1971,40 24. '05 M. Simonyi and F. Tudos Ado. Phys. Org. Chem. 1971 9 127. lob D. G. Pobedimskii Russ. Chem. Ret;. 1971 40 142. lo' J. Kalvoda and K. Heusler Synthesis 1971 501. Free-radical Reactions and Electron Spin Resonance Spectroscopy a further volume"' in the series entitled 'Organic Peroxides' have been pub- lished. The rates and activation parameters for decomposition of di-t-butyl peroxide in cyclohexane and acetonitrile over a temperature range wider than that previously employed have been re-evaluated and confirmation obtained that they are dependent on the nature of the solvent. However the magnitude of the solvent effect is much smaller than that previously recorded."' The effect which the stability of the incipient radicals have on the rate of thermal decomposition of peresters is thought to be slight because of the surprising thermal stability of the esters (102).Formation of the relatively stable iminoxyl (103) appears to provide little driving force for this decomposition and it was concluded that the R' R' polarity of the R group (in RCOOBu') was the critical factor.' lo Pyrolysis or photolysis of the perfluoro-peroxide (104),which is more stable than dimethyl peroxide in the gas phase gave mainly carbon dioxide and carbonyl fluoride. CF30-OCF3 -+CF300. + CF + CF + F-// (104) 0-0 (1051 J\ COF + F. CO + 2F. These are considered to arise t.ia the novel intermediate (105)(detected by its i.r.spectrum) formed by initial 0-C fission.' ' Generation of benzyl radicals (106) by reaction of the parent peroxide with N-bromosuccinimide did not initiate an intramolecular-induced decomposition of the peroxide group the final product being the corresponding brominated peroxide. 'l2 This result lends support to the mechanism of Cadogan et al.' l3 [Scheme 1 route (a)]rather than to that of Walling et ~1."~[route (b) represented in this case by (107)] for the radical- induced decomposition of dibenzoyl peroxide with formation of a bond between the attacking radical and the ortho or para carbon atom. The mode of thermal decomposition of rn-nitrobenzenesulphonyl peroxide (108) depends critically I"* 'Organic Peroxides' ed.D. Swern Wiky New York 1971 vol. 11. C. Walling and D. Bristol J. Org. Chem. 1971 36 733. C. Ruchardt and R. Pantke Chem. Ber. 1971 104 3456. ''I K. 0. Christe and D. Pilipovich J. Amer. Chem. Soc. 1971 93 51. M. M. Schwartz and J. E. Leffler J. Amrr. Chem. SOC.,1971 93 919. J. I. G. Cadogan D. Hey and P. G. Hibbert J. Chem. Soc. 1965 3939; C. Walling and Z. Cekovic J. Amer. Chem. SOC.,1967 89 6681. A. R. Forrester R-OJ" C=O + PhCOO-+ PhCOO. Scheme 1 upon the solvent. In aromatic solvents an ionic process prevails and electro- philic sulphonyloxylation occurs but in chloroform products derived from rn-nitrobenzenesulphonyloxylradicals were obtained. l4 The mixed peroxide (109) on decomposition in benzene gave products mainly derived from radicals PMe \ only when a base (MgO) was present to remove the toluene-p-sulphonic acid formed during the reaction.' Details of the oxidation of phenols with silver carbonate supported on celite have now been disclosed. This oxidant which is claimed to be highly specific gives products derived exclusively from C-C coupling of the intermediate phenoxyls.'' A new reagent hydrogen[hexacyanoferrate(~~~)] (from K,FeCN and HCl) which is soluble in methanol oxidizes phenols with low oxidation potentials but not phenol itself nor p-cresol. l7 With 2,6-di-t-butyl-4-methyl-phenol the dienone (110) (28 %) was the main product formed it is thought by further oxidation of the initial phenoxyl to a phenoxonium ion which is then solvolysed.Evidence for the intermediacy of phenoxonium ions in anodic 'I4 Y. Yokoyama H. Wada M. Kobayashi and H. Minato Bull. Chetn. SOC.Japan 1971,44 2479. 'I5 R. Hisada H. Minato and M. Kobayashi Bull. Chem. Soc. Japan 1971 44 2541. 'Ih V. Balogh M. Fetizon and M. Golfier J. Org. Chem. 1971 36 1339. G. Biggi F. Del Cima and F. Pietra Tetrahedron Li.trers 1971 281 1. Free-radical Reactions and Electron Spin Resonance Spectroscopy 209 0 HO,Il,O" 1-0-0 0' 0 00'0 tPH2 (1 10) oxidations of 2,4,6-di-t-butylphenol,' ' autoxidations of 2,4-di-t-butylphenols catalysed by aminesopper salt catalysts l9 oxidations of certain hydroxy- diphenyl ether and diphenylmethane derivatives with lead dioxide and tetra- acetate,120 oxidations of bisphenols related to a-tocopherol,12 ' and oxidation of salicyl alcohols and similar phenols by periodate'22 has been presented.Oxidation of salicyl alcohol is particularly interesting since it gives a Diels-Alder dimer of the dienone (112). However the possibility that the monomer arises not from the cation (1 11) but from the periodic ester (1 13) cannot be discounted. Disproportionation of aryloxyl radicals to aryloxy-anions and -cations [equation (6)] at low pH values (by analogy with semiquinone radical-anions) has been 2ArO. S ArO' + ArO-(6) proposed by Waters.'23 Consideration of the extensive data available on phenol oxidations led him to the further proposal that oxidations in acidic solution especially with reagents of high potential gave phenoxonium ions which prefer to couple 0-C whereas reagents such as alkaline ferricyanide gave phenoxyls which prefer to couple C-C.It is difficult to make generalizations in this field because of the complexity of the product mixtures frequently encountered and the lack of detailed information on iizter alia the role of ligand molecules in many oxidations. Nevertheless the above rationalization could serve as a useful guide although it undoubtedly requires qualification in many instances. Con- trary to Waters' proposals Scheme 2 has been devised to account for the produc- tion of aryloxyls in the autoxidation of certain phenols in the presence of copper salt-pyridine catalyst the radicals coupling 0-C to yield commercially sig- nificant polymers.' 24 This scheme accommodates observations such as the coupling process is first order in oxygen pressure and copper catalyst and is independent of phenol concentration but not phenol structure.t-Butylperoxyl radicals bound to cobalt and formed by treatment of acetylacetonatocobalt(I1) with t-butyl hydroperoxide oxidized 2,4,6-tri-t-butylphenol to the corresponding '" A. Ronlan and V. D. Parker J. Chem. Soc. (C) 1971 3214. '" D. G. Hewitt J. Chem. Soc. (0,1971,2967. ''O D. G. Hewitt J. Chem. Soc. (C) 1971 1750. ''I M. Chauhan F. M. Dean K. Hindley and M. Robinson Chem. Comm. 1971 1141. '*' E. Adler S. Brasen and H. Miyake Acta Chem. Scand. 1971 25 2055. W. A. Waters J. Chem. SOC.(B) 1971 2026. lZ4 C. C. Price and K. Nakaoka Macromolecules 1971 4 363.A. R. Forrester 2 ArOH 1-2A~O. 1-2ArO. HO CI (PY)2 2ArOH \culI &/ 'cu,I k/ -2HzO 'OH Scheme 2 'unbound' and 'bound' phenoxyls. Both species (g = 2.006 and 1.996 respec- tively) were detected by e.s.r. the latter showing in addition to an octet splitting due to coupling with the cobalt a doublet splitting which could not be assigned.125 The exact nature of these 'bound' radicals has not yet been clearly defined but comparison of their reactions with those of the corresponding 'unbound' radicals should be revealing. A review describing mechanisms of hydroxylation of aro- matic compounds has been published in which phenol oxidation is given con- siderable attention.' 26 Oxidation of a number of 2- and 4-halogenphenols and 2-nitrophenol with hydroxyl radicals (Ti"'-H202) resulted in a displacement of the substituent with formation of the 1,2- or 1,4-benzosemiquinone.127 The instability of benzo- semiquinones in aqueous alkaline solution has been attributed to their autoxida- tion to more highly oxygenated radical-anions.28 Thus 2-methylquinol gave a mixture of three isomeric radical-anions one of which [( 114); 43-H) = 0.055 a,(Me) = 0.095 and 45-H) = 0.415 mT] is shown. The semiquinone of naphtha- zarin (115) (aH= 0.238 and a,(OH) = 0.051 mT) generated in pyridine-tri- ethanolamine coupled C-C to give a dimer (5%) and a cyclotrimer (1 1%) ;'29 0. HO 0-0-HO 0. (1 14) (1 15) (1 16) some related radical-anions behaved similarly. The semiquinone (116) derived from [2,2]paracyclophane showed an interesting splitting in its e.s.r.spectrum attributed to coupling with the nitrogen (aN= 0.215 rnT).I3' '" A. TkaC K. Vesely and L. Omelka J. Phys. Chem. 75 2580 2575. lZ6 D. I. Metelitsa Russ. Chem. Rer. 1971 40 563. 12' K. Gunther W. G. Filby and K. Eiben Teiruhedroti Leiiers 1971 251. P. Ashworth and W. T. Dixon Chern. Cornm. 1971 1150. H. Brockmann H. Greve and K. Hoyermann Tetrahedron Letters 1971 1493. A. R. Forrester and R. Ramasseul J. Chem. SOC.(B) 1971 1638. Free-radical Reactions and Electron Spin Resonance Spectroscopy 21 1 Spectra of the semidiones of a varied collection of bicyclo[n,l,O]-and bicyclo-[n,l,l]-alkanes have been described from which a wealth of information about the dynamic stereochemistry and other structural features of the carbon skeleton has been gained.Of particular interest is the conversion of the semidione (117) (coupling constants in mT) to the ortho-benzosemiquinone (118) on treatment with an excess of oxygen. An unusual carbonyl insertion has been uncovered during an attempt to make the ketyl (119) from the corresponding ketone by alkali-metal reduction. The spectrum observed was identical with that of the semidione (120) produced from the diketone (121) and had uH = 0.269 uHb= nH = 0.044 mT. The mechanism of this change has not been elucidated. 132 Spectra of the semidiones (122)derived from ferrocene showed no coupling with protons of the interannular ring and only a small splitting (0.05mT) attributable to the ‘ortho’ protons.’33 Evidence134has been presented that reduction of 0-0.90-C=C-Me Fe t R ,Q Hc Hh Ha perfluorobiacetyl with lithium gives in addition to the monomeric semidione (tight and loosely bound ion pairs detected) a triplet species (aF= 1.067 mT-six equivalent fluorines) formulated as (123).Dialkoxy-semidiones (124) formed by electrolytic reduction of dialkyl oxalates have coupling constants for their 131 G. A. Russell J. J. McDonnell P. R. Whittle R. S. Givens. and R. G. Ke5ke J. Amer. Chem. Soc. 1971 93 1452; G. A. Russell P. R. Whittle and R. G. Keske J. Amer. Chem. SOC., 1971 93 1467. 132 J. P. Dirlam and S. Winstein J. Org. Chem. 1971 36 1559. I33 J. J. McDonnell and D. J. Pochopien J.Org. Chem. 1971 36 2092. I34 G. A. Russell. J. L. Gerlock and D. F. Lawson J. Amer. Chem. SOC.,1971 93 4088. 2 12 A. R.Forrester a-protons which depend upon the size of the /?-substituent this controlling the time-averaged dihedral angle. ' For example (124 ;R = Et) has a(a-H) = 0.134 mT and (124; R = Pr') has a(a-H) = 0.71 mT. A number of relatively stable ketyl radical-anions of cyclic a/?-unsaturated ketones in which both hydrogens a to the x-system have been replaced by alkyl or aryl groups e.g. (125) have been produced by electrolytic reduction of the ketones in DMF.'36 For (125) the values (in mT) shown for the coupling con- stants and the absence of a detectable splitting for the protons in the 6-methyl 0-0-(1.18) (0.034) A Me Me (0.08) groups indicate that spin density at the fi-position is greater than not only that at the a-position but also that at the carbon of the carbonyl group.Spectral interpretation was assisted by both Huckel and McLachlan MO calculations. Cyclic dienones were surprisingly easily reduced to the corresponding ketyl radical-anions e.g.(126) (coupling in mT) by reaction with potassium t-butoxide in dimethyl sulphoxide in a flow ~ystem.'~' Good agreement between experi- ment [a(/?-F)= 5.85 and 8.44mT respectively for the ketyl radical-anions of perfluorodiethyl ketone and perfluorocyclobutanone] and theory (INDO cal- culations) supports the notion that spin delocalization into the perfluorinated alkyl groups of the above radicals occurs uia overlap of fluorine p-orbitals and 2p orbitals of the x-system..' 38 Fluorine hyperfine coupling constants for the ketyl radical-anion [(ArF)2CO-] of perfluorobenzophenone [a(o-F) = 0.480 a(m-F) = 0.106 and a(p-F) = 0.836mTl and the ketyl [(AT~)~COH] are unex- pectedly larger than those of their partially fluorinated analogues (in 4,4'-difluoro-benzophenone ketyl radical-anion aFis 11 smaller).' 39 The structure of ion 135 J.Voss Tetrahedron 1971 27 3753. IJ6 G. A. Russell and G. R. Stevenson J. Amer. Chem. Soc. 1971,93 2432. 137 G. A. Russell and R. L. Blankespoor Tetrahedron Letters 1971 4573. 138 W. R. Knolle and J. R. Bolton J. Amer. Chem. Soc. 1971 93 3337. "9 F. P. Sargent and M. G. Bailey Canad. J. Chem. 1971 49 2351. Free-radical Reactions and Electron Spin Resonance Spectroscopy 213 pairs of fluorenone and xanthone radical-anions have been described in terms of a dynamic model in which the cation can 'jump' between two positions of minimum energy the relative populations of these states depending inter ah on the ~olvent.'~' Several simple methyl ketones gave (e.s.r.) complex mixtures of radicals on U.V.irradiation at low temperature in a flow system. The product radicals included ketyls (by photoreduction) alkyl radicals and the protonated semidione MeCOC(0H)Me (by Norrish Type I cleavage of the excited ketone) acylmethyl radicals (by hydrogen abstraction) and solvent-derived radicals (by hydrogen abstraction or induced decomposition of solvent). The radicals detected from a particular ketone depended upon the structure of the ketone and the solvent.141 4 Sulphur Radicals Homolytic bimolecular substitution (S,2) of alkylboranes by alkylthiyl radicals has been established by the detection (esr.) of the displaced radical (R2-)[equation (7)].'42 Similar reactions with Grignard reagent^'^^,'^^ and with organo-bismuth and -antimony compounds have also been observed and it appears that this is R'S.+ R2,B -+ R'SBR' + R2-(7) a common reaction of thiyl radicals and organometallic reagents. The reaction with trialkylboranes has been utilized'44 in a new preparation of unsymmetrical sulphides in which the trialk ylborane and a symmetrical disulphide are heated in THF in the presence of air or U.V. light. Yields of sulphide formed in this chain process [equations (7) and (8)] are -90%.The extensive and frequently contro- versial literature on S,2 reactions has been sifted in a recent m~nograph.'~~ R2. + R'SSR' + R2SR1 + R'S. (8) The reversibility of the intramolecular addition of thiyl radicals has been convincingly dem~nstrated'~~ by the isolation of the same ratio of (134 133; R' = Me) (95 :3) on irradiation of either of the thiol precursors of the radicals (127 and 128 ; R' = Me) at 80 "C. The high proportion of (134) implies that the radical (131 ; R' = Me) is thermodynamically more stable than either (129 or 130; R' = Me) since at -65°C the ratio (134 133; R' = Me) from (127; R' = Me) is 22 :76 and from (128; R' = Me) is 50 50. Significantly the principal product (67 %) from (128 ;R' = H) under conditions of thermodynamic 14" K.S. Chen S. W. Mao K. Nakamura and N. Hirota J. Amer. Chern. SOC.,1971 93 6004; CJ B. J. Tabner and J. R. Zdysiewicz J. Chem. Soc. (B) 1971 1659. 14' H. Paul and H. Fischer Chem. Comm. 1971 1038. 142 A. G. Davies and B. P. Roberts J. Chem. SOC.(B) 1971 1830. 143 A. W. P. Jarvie and D. Skelton J. Organometallic Chem. 1971 30 145. 144 H. C. Brown and M. M. Midland J. Amer. Chem. Soc. 1971,93 3291. 145 K. U. Ingold. and B. P. Roberts 'Free-Radical Substitution Reactions Bimolecular Homolytic Substitutions (S,2 reactions) at Substituted Multivalent Atoms,' Wiley New York 1971. '46 J.-M. Surzur M.-P. Crozet and C. Dupuy Tetrahedron Letters 1971 2025 2031 ; J.-M. Surzur R. Nouguier M.-P.Crozet and C. Dupuy Tetrahedron Letters 1971 2035. A. R. Forrester R' control (80 "C)was (134; R' = H) [which arises viu the primary radical (131 ; R' = H)] and not (133; R' = H) (which arises via a secondary radical). Product (132) was only formed (13%) from (127) when R' = H and then only under conditions of kinetic control (-65 "C). Spontaneous decomposition of arenediazothiolates on heating [equation (9)] is assisted by a radical-induced decomposition in which probably both aryl and ArN=NSR + AP + N + RS. (9) arylcyclohexadienyl radicals (from Ar. and aromatic solvent) participate. 47 The large difference between the coupling constants of the a-and P-protons of the adducts of alkylthiyl radicals and alkenes e.g. (135) (coupling constants in mT) CH ,SCH,CH S,CH3 ?f I.\ (1.489)(2.16) H,C-CH (135) (136) excludes the symmetrically bridged structures (136) previously proposed for such species.14' However the P-coupling constants are smaller than those expected for the favoured conformation (137) and a model (138) has been designed in which the sulphur is distorted somewhat towards the half-filled p-orbital (with a corresponding movement of the B-hydrogens from their tetrahedral positions) not enough to create the bridged species (136) but sufficiently to control the stereochemistry of the subsequent reactions of radical.14' H. Van Zwet J. Reiding and E. C. Kooyman Rec. Trac. chirn. 1971 89 21. 148 P. J. Krusic and J. K. Kochi J. Amer. Chem. Soc. 1971 93 846; T. Kawamura M.Ushio T. T. Fujimoto and T. Yonezawa J. Amer. Chem. Soc. 1971 93 908. Free-radical Reactions and Electron Spin Resonance Spectroscopy 21 5 H Electrochemical reduction of aromatic thiocarbonyls yielded thioketyl radical- anions whose aromatic proton coupling constants are much smaller than those of the corresponding ketyls thus reflecting the less-efficient transfer of free spin from the CS group compared with the CO group to the ring position^.'^^ Spectroscopic features of a wide variety of radicals and radical-ions derived from thiophen have been summarized.' 50 Similarity in the e.s.r. and electronic spectra of the radical-cations (139) and (140)has been attributed to some degree of S-S bonding in (140). ' (139) lJ9 L. Lunazzi G.Maccagnani G. Mazzanti and G. Placucci J. Chem. Soc. (B) 1971 162. I5O L. Lunazzi A. Mangini G. F. Pedulli and M. Tiecco Gazzetta 1971 101 1. Is' B. I. Stepanov W. Ya. Rodionov A. Ya. Zheltov and V. V. Orlov Terruhedron Letters 197 1 1079.