4 Reaction Mechanisms Part (iii) Free-Radical Reactions By R. A. JACKSON School of Molecular Sciences University of Sussex Brighton BN 1 9QJ 1 General After a gap of five years Volume 6 of ‘Advances in Free Radical Chemistry’ has appeared,’ with a new publisher and articles on radical rearrangements polar effects halogenation of cycloalkanes acyl-aryl-nitrosamines and phosphoranyl radicals. Volume 4 of ‘Free Radicals in Biology’ provides further reviews2 in this important interface area. Useful accounts of ipw-s~bstitution,~ electron spin resonance and physical-organic ~hemistry,~” ‘free radical locks',^' and the decomposition of azoalkanes have appeared,’ and further reviews are found in the report OR a Free Radical Symposium held at Baton Rouge6 in 1979.A compilation of e.s.r. para- meters for non-carbon-centred radicals has appeared in the Landolt-Bornstein ~eries.~ Onthe whole 1980has seen steady advances rather than spectacular discoveries but of more than usual interest are some guidelines for chemical reactivity which emphasize stereo-electronic effects suggested by Beckwith Easton and Serelis.8 A number of reactions of potential synthetic value are considered in the section on fragmentation. 2 Structural Studies Electron spin resonance studies dominate this area. Hydrogen abstraction by photolytically produced t-butoxyl radicals is commonly used to produce radicals for structural studies. In some systems stronger signals are obtained if p-methyl- acetophenone is used as a sensitizer for the photolysis of the t-butyl peroxide.’ There has been a continuing interest in persistent radicals.The sterically shielded alkoxyaminyl radical Bu‘ONBU‘ can be prepared by photolysis of the parent hydrox- ‘Advances in Free Radical Chemistry’ Vol. 6 ed. G. H. Williams Heyden London 1980. ‘Free Radicals in Biology Vol. 4’ ed. W. A. Pryor Academic Press New York 1980. M. Tiecco Acc. Chem. Res. 1980,13 51. D. Griller and K. U. Ingold Acc. Chem. Res. 1980,13 (a)p. 193; (b)p. 317. P. S. Engel Chem. Rev. 1980,80,99. ‘Frontiers in Free Radical Chemistry’ ed. W. A. Pryor Academic Press New York 1980. ‘Landolt-Bornstein. New Series Group 11 Vol. 9 Part c2. Organic 0-,P- S- Se- Si- Ge- Sn-,Pb- As- Sb-centered Radicals’ by A. G. Davies J. A. Howard M. Lehnig B.P. Roberts H. B. Stegman and W. Huber Springer Verlag Berlin 1979. A. L. J. Beckwith C. J. Easton and A. K. Serelis J. Chem. SOC.,Chem. Commun. 1980,482. D. Griller K.U. Ingold and J. C. Scaiano J. Mugn. Reson. 1980 38 169. 53 54 R. A. Jackson ylamine in sufficient concentration for the hyperfine splittings due to the "0 that is present in natural abundance to be determined." By use of the very bulky groups R=CH(SiMe3)2 or N(SiMe3I2 radicals of the type *PR2 and -AsR2 have been obtained." The radical *P[CH(SiMe3)2]2 has a half-life >1 yr at room temperature in toluene solution. At the other extreme spin trapping continues to be used for radicals which cannot be observed directly. The azidyl (Ng a) cyanatyl (OCN-) and cyanyl CN) radicals and chlorine atoms have been trapped by a-pheny1-N-t- (a butylnitrone.12 The most recent contribution to the controversial question of the shape of the t-butyl radical is an ab initio calc~lation'~ which suggests a pyramidal geometry with an out-of-plane bending angle of 10".In another controversial area that of radical cations derived from alkanes strong evidence for the formation of [Me3CCMe3]t has been obtained by y-irradiation of the alkane in a glassy Freon solution. A septet of a =29.0 G was interpreted in terms of a structure with staggered conformations about all the C-C bonds and a strong interaction with one (axial) proton from each methyl group.14 The conformations of substituted cycloalkyl radicals (1; n = 3-6) have been studied by adding triethylsilyl radicals to the appropriate methylenecycloalkane.l5 For the substituted cyclohexyl radical (1; n = 4) at -120"C couplings to four P-protons were observed i.e. two axial (38.5G) and two equatorial (6.25 G). At +120 "C these protons became equivalent and a value of AG' of 22 kJ mol-' was obtained from a computer simulation of the lineshape. For the smaller ring sizes the P-protons were all equivalent. It was suggested that the substituted cyclopentyl radical is probably non-planar whilst the substituted cyclobutyl radical is essentially planar. (CH2)" (CH,) SiEt +*SiEt3 -* (1) Triarylsilyl radicals pose the problem of the extent of conjugation of the free electron on the silicon atom with the aromatic n-system but such radicals have not hitherto been observed probably because they appear to be very reactive in aromatic substitution reactions.This year two groups have solved the problem. Tris-(3,5-di-t-butylphenyl)-silyland -germyl radicals have been prepared16 by abstraction of hydrogen from the corresponding silane or germane and trimesityl- silyl -germyl and -stannyl radicals (2,4,6-Me3C6H2)3M have been made17 by U.V. irradiation of the triarylmetal chloride with an electron-rich olefin [RNCH2CH2NRC=I2. The coupling constants of the hydrogen atoms of the ring are smaller than in the analogous triphenylmethyl radical indicating reduced but lo H. Woynar and K. U. Ingold J. Am. Chem. SOC., 1980,102,3813. I' M. J. S. Gynane A. Hudson M. F. Lappert P. P. Power and H.Goldwhite J. Chem. Sac. Dalton Trans. 1980 2428. '' W. Kremers and A. Singh Can. J. Chem. 1980 58 1592; E. G. Janzen H. J. Stronks D. E. Nutter Jr. E. R. Davis H. N. Blount J. L. Power and P. B. McCay ibid. p. 1596. l3 R. E. Overill and M. F. Guest Mol. Phys. 1980.41 119. l4 J. T. Wang and F. Williams J. Phys. Chem. 1980,84 3156. Is L. Lunazzi G. Placucci and L. Grossi J. Chern. Soc. Perkin Trans. 2 1980 1761. l6 H. Sakurai H. Umino and H. Sugiyama J. Am. Chem. SOC.,1980,102,6837. 17 M. J. S. Gynane M. F. Lappert P. I. Riley P. Rivikre and M. Riviere-Baudet J. Organornet. Chem. 1980 202 5. Reaction Mechanisms -Part (iii) Free-Radical Reactions 55 significant delocalization. The value of ~(~'si) in (2,4,6-Me3C6H2),Si* is 135 G; the reduction from the value of -182 G that is found in Me&* indicates some delocalization of the unpaired electron or flattening of the pyramidal structure or both.Photolysis of the phenylazo aryl sulphide PhN=NSC6H4Bu'-p gave an e.s.r. spectrum attributed to the phenyldiazenyl radical PhN=N* with couplings a(N)' = 23 G a(N)* = 9.4 G and aH = 1.1G (two meta protons) and g = 2.0006. The data" support a a-radical structure (as in the benzoyl radical) with free rotation round the Ar-N bond. A study of the e.s.r. spectra of amidyl radicals at different temperatures indicates a wstructure for these radicals;'9a in a contribution to the question of the possibility of two different structures for succinimidyl radicals,lgb the ground state is predicted to be the v-form by a UHF calculation.Sulphur-centred radicals show increasing variety." Sulphinyl (RSO *) and sul- phony1 (RS02 *) radicals are implicated in the anti-oxidant activity of organic sulphides and have been prepared for e.s.r. structural studies by abstraction of chlorine from sulphonyl chlorides by triethylsilyl radicals or by thermolysis or photolysis of sulphoxides and thiolsulphonates ArS02SAr. Addition of CF3S radicals to dialkyl sulphides R,S gives sulphuranyl radicals R2SASCF3 in which the electron is thought to be in the a*-orbital:*l analogous selenuranyl radicals,22 e.g. R2Se'OBu' can be prepared by photolytic production of Bu'O. in the presence of R2Se. The cyclic sulphur radical (3) can be made by photolysis of the peroxide (2) a new system of nomenclature for radicals of this type has been by which the radical (3) would be 9-S-3 (nine electrons centred on sulphur with three bonds).Me 0 + Bu'O* 0-OBu' 0 0 3 Formation Destruction and Radical Stability Thermolysis of suitable organic compounds provides information about stabilization in the radicals formed on the assumption that E = AH for the homolysis. Ther- molysis of hexa- 1,3-diene gives methyl and 2,4-pentadienyl radicals a resonance energy of 77 kJ mol-' for the latter radical was derived.24 In a very-low-pressure pyrolysis (VLPP) study the effects of ring-methyl substituents on the rate of ArCH2-CH3 bond scission were determined.2s The activation energy was lowered T. Suehiro T. Tashiro and R. Nakausa Chem.Lett. 1980 1339. l9 (a)J. Lessard and K. U. Ingold J. Am. Chem. SOC.,1980,102 3262; (b)Y. Apeloig and R. Schreiber ibid. p. 6144. 20 C. Chatgilialoglu B. C. Gilbert B. Gill and M. D. Sexton J. Chem. SOC.,Perkin Trans. 2 1980 1141; C. Chatgilialoglu B. C. Gilbert and R. 0.C. Norman ibid. p. 1429. 21 J. R. M.Giles and B. P. Roberts J. Chem. SOC., Perkin Trans. 2 1980 1497. J. R. M. Giles B. P.Roberts M. J. Perkins and E. S. Turner J. Chem. SOC.,Chem. Commun. 1980,504. 23 C. W. Perkins J. C. Martin A. J. Arduengo W. Lau A. Alegria and J. K. Kochi J. Am. Chem. SOC. 1980,102,7753. 24 A. B. Trenwith. J. Chem. SOC.,Faraday Trans. 1 1980,76,266. 25 B. D. Barton and S. E. Stein J. Phys. Chem. 1980,84 2141. 56 R. A. Jackson by 1.3-1.7 kJmol-' for m-and p-methyl groups and by 5.0-6.3 kJ mol-' for o-methyl groups the larger value being ascribed to a partial relaxation of steric strain in the transition state.The same technique when applied to l-ethylnaph- thalene and 9-ethylanthracene gives values for the extra delocalization energy of the 1-naphthylmethyl and 9-anthrylmethyl radicals over that of the benzyl radical as 19 and 35 kJ mol-' respectively.26 The P-keto-diazene (PhCOCMe2N=)2 has been synthesized and found to decompose 1.1~10~ times faster at 100 "C than the corresponding compound in which the benzoyl groups have been replaced by methyl gro~ps.~' The effect was attributed to the stabilizing effect of the benzoyl group on the incipient radical the order of effectiveness of substituents in speeding up decompositions in compounds (RCMe2N=)2 is Me < OMe < C1 < CN < COPh < Ph < CH=CHZ.In studies of the thermolysis of highly substituted alkanes (R'R2R3C-)2 [R # HI it was concluded that homolysis of the central bond was rate-determining and from a comparison of AG'(300 "C) with strain enthalpies for the compounds obtained from force-field calculations it was concluded that about 40% of residual strain is still present in the transition state of these C-C cleavage reactions.28 The thermal decomposition of azoalkanes shows certain anomalies when com- pared for example with decomposition of peresters. In particular the relative rate of decomposition of RN=NR for R = 1-adamantyl compared with R = t-butyl is only 0.0004 even though other evidence points to the 'normal' nature of the 1-adamantyl radical.To take account of this and other evidence it has previously been suggested that azo-compounds do not decompose in a 'least motion' process. Firestone2' suggests that this process involves a Linnett-type transition state in which the electrons are removed from the breaking bond one at a time with the temporary development of charge during the reaction the transition state being of the type 66+ a+ 6-R -N=N--R'. This fits in with the speeding-up effect of electron-withdrawing substituents on such decompositions and thus one should be cautious about attribut- ing differences in rates of decomposition of azo-compounds mainly to radical stabilization. Pressure effects have been used as a probe for thermal decompositions.Azocumene decomposes with A V+of +5 ml mol-' with pressure favouring dispro- portionation over radical co~pling.~' The decomposition of liquid bibenzyl to toluene and stilbene shows A V+= +31mi mol-' indicating that there is a transi- tion state involving a very wide separation of the two benzyl radicals and the value of the activation energy is in line with other estimates of the stability of the benzyl radi~al.~' A variety of reaction types may be involved in induced decompositions of organic peroxides. For ally1 t-butyl peroxide induced decomposition is responsible for about half the observed rate and studies of the products have shown that both radical 26 D. F. McMillen P. L. Trevor and D.M. Golden J. Am. Chem. Soc. 1980,102 7400. '' R. C. Zawalski M. Lisiak P. Kovacic A. Luedtke and J. W. Timberlake TetrahedronLeft.,1980 21 425. 28 H.-D. Beckhaus G. Kratt K. Lay J. Geiselmann C. Riichardt B. Kitschke and H. J. Lindner Chem. Ber. 1980,113 3441; R. Winiker H.-D. Beckhaus and C. Riichardt ibid. p. 3456. 29 R. A. Firestone J. Org. Chem. 1980 45 3604. 30 R. C. Neumann Jr. and M. J. Amrich Jr. J. Org. Chem. 1980,45,4629. 31 K. R. Brower J. Org. Chem. 1980 45 1004. Reaction Mechanisms -Part (iii) Free-Radical Reactions 57 addition to C=C and abstraction of H from the allylic CHz group take Dibenzenesulphenimide (PhS)2NH induces the decomposition of benzoyl peroxide at room temperature. An intermediate adduct PhSNHS(0Bz)zPh was postu-lated but the precise mechanism of its formation and breakdown (in particular the question of whether electron-transfer reactions are involved) is still not clear.32b Homosolvolysis (radical transfer to a solvent which is itself a stable free radical usually di-t-butyl nitroxide) has been used in an elegant demonstration of capto- dative stabilization of radicals.33 For example BrCH2COZEt (with an acceptor group) is quite inert towards abstraction of the bromine atom by t-butyl nitroxide and BrCH20Me (which provides a donor group only) reacts in about seven hours at room temperature whereas BrCH(OMe)C02Me which contains both a donor and an acceptor group reacts immediately.Although alkyl radicals with @-hydrogen atoms react together in pairs by dispro- portionation as well as by combination it had become accepted that silicon-centred radicals do not undergo disproportionation.That trimethylsilyl radicals do indeed disproportionate has now been established by three independent groups.34 Trimethylsilyl radicals were generated in the gas phase and in solution by photolysis (at 147 nm) of Me6Si2 by photolysis of bis(trimethylsilyl)mercury by mercury- photosensitized decomposition of trimethylsilane or by abstraction of hydrogen from trimethylsilane by t-butoxyl radicals. The MezSi=CHz that was produced in the disproportionation was trapped as Me3SiOBu' or Me2Si(CH2D)OCD3 in the presence of Bu'OH or CD30D respectively (D=2H). Values of kdisp/kcomb of 0.31 (in the gas phase) and 0.19 (in solution) were obtained.A novel technique to determine self- and cross-termination rate constants invol- ves the photolytic generation of two different radicals by a harmonically modulated light source and then following the phase shifts of the two radicals compared with the incident light.3S This general approach should prove very useful in unravelling complex radical kinetics. 4 Radical Transfer SH2Reactions at multivalent centres continue to attract attention. Trifluoromethyl radicals react with neopentane at 300 "C to give isobutane as a primary product and more CF3CH3 than would be expected by cross-combination of methyl and trifluoromethyl radicals.36 The results were interpreted in terms of reaction (1). CF3. + H3C-C(CH3)3 -* CF3CH3 + .C(CH3)3 (1) Even in this relatively favourable case (unhindered position of attack; stable radical formed; high temperature) the SH2reaction only plays a minor role in the overall reaction which suggests that other examples will be difficult to find in the absence of special features.One such feature involves attack by or displacement of 32 (a)R. Hiatt and V. G. K. Nair Can.J. Chem. 1980,58,450; (b)D. F. Church and W. A. Pryor J. Org. Chem. 1980,45,2866. 33 H. Singh and J. M. Tedder J. Chem. SOC.,Chem. Commun. 1980,1095. 34 S. K. Tokach and R. D. Koob J. Phys. Chem. 1980 84 1; J. Am. Chem. SOC. 1980,102 376; L. Gammie I. Safarik 0.P. Strausz R. Roberge and C. Sandorfy ibid. p. 378; B. J. Cornett. K. Y. Choo and P. P. Gaspar ibid. p. 377. 35 H. Paul and C. Segaud Int.J. Chem. Kinet. 1980,12,637. 36 R. A. Jackson and M. Townson J Chem. SOC.,Perkin Trans. 2,1980,1452; Tetrahedron Lett. 1973,193. 58 R. A. Jackson organometallic radicals (see e.g. ref. 37a). For some organometallic radicals the ratio (bond strength to C)/(bond strength to H) is higher than for organic radicals (see e.g. ref. 37b) which is a factor that will favour SH2 reactions by these organometallic radicals compared with -competing hydrogen-transfer reactions. Intramolecular SH2 reactions at carbon are better established. In a recent example,38 bromotrichloromethane reacts with (but-3-enyl)cobaloximes e.g. [(H,C=CHCH,CHM~)CO(~~~H)~~~], as shown in Scheme 1. [( H2C=CHCH2CHMe)Co(dmgH),py] + cc13 + [(C13CCH2CHCH2CHMe)Co(drngH)2py] [BrCo(dmgH)2PYl k[Co"(dmgH)zpy] +14 CI,C Scheme 1 SH2 Reactions at the other Group IVB elements are well established and a smooth SN2-like process is often assumed.Ko~hi~~ argues for a charge-transfer mechanism (Scheme 2) in the reaction of tetra-alkyl-tins with iodine atoms based on selectivity studies and the effect of solvent on the rates of reaction. &Sn + I*-+ [&Sn+.I-]' -+ R3SnI + R* Scheme 2 An excess of tributyltin hydride has been used to reduce ethylene thioketals or thioacetals to the corresponding hydrocarbons in synthetically significant yields (Scheme 3). A 1 1molar ratio gives the ,!?-(alky1thio)ethyltributyltinsulphide SH2 attack at sulphur is thought to be inv01ved.~' "x;] "'x), Bu,SnH+ AIBN R2 R2 H SSnBu3 Scheme 3 The reaction of methylperoxyl radicals with alkenes gives the epoxide and a methoxyl radical.41 It has been suggested that the /3 -methylperoxyalkyl radical (4) undergoes an internal SH2 attack at the peroxide oxygen atom to close the ethylene oxide ring and displace methoxyl radical (Scheme 4).Turning to radical transfer of univalent atoms the contribution of polar influences to such reactions continues to receive attention. Reduction of substituted benzyl CH302*+ H2C=CH2 + CH3-O-O-CH2CH2-+ CH30. + H,C-CH, \/ (4) 0 Scheme 4 37 (a)T. Funabiki B. D. Gupta and M.D. Johnson J. Chem. SOC.,Chem. Commun. 1977,653; (b)R. A. Jackson J. Organomet. Chem. 1979,166 17. M.R. Ashcroft A. Bury,C. J. Cooksey A. G. Davies B. D. Gupta M. D. Johnson and H.Morris J. Organomet. Chem. 1980 195 89. 39 S.Fukuzumi and J. K. Kochi J. Org. Chem. 1980,45,2654. 40 C. G. Gutierrez R. A. Stringham T. Nitasaka and K. G. Glasscock J. Org. Chem. 1980 45 3393. 41 K. Selby and D. J. Waddington J. Chem. SOC. Perkin Trans. 2 1980 65; D. A. Osborne and D. J. Waddington ibid.,p. 925. Reaction Mechanisms -Part (iii) Free-Radical Reactions 59 halides by tributyltin gave a Hammett plot that was best fitted by u-values and values of p of +0.34 (Cl) +0.17 (Br) but +0.81 (I). The high positive value for benzyl iodides suggests that electron transfer may be important for this series. Kinetic isotope effect studies using 13C also support a significant change in the transition state between Br and I in the reactions of triphenyltin radicals with the methyl halides.42b For CH3Cl and CH3Br the kinetic isotope effect is in excess of the equilibrium value signifying a late transition state but the value for CH31is significantly less than the equilibrium value.When irradiated in the presence of t-butyl hydroperoxide the 8-phenylthio-5'- deoxyadenosine derivative (5)was converted into the cyclized product (6).43Pre-sumably the radical that is originally formed having a free electron at the 8-position abstracts a hydrogen atom from the 5'-methyl group and this intermediate methyl- ene radical then adds back at the 8-position followed by loss of a hydrogen atom to another radical in a normal homolytic aromatic substitution sequence to give the observed cyclized product (6).This provides a parallel in vitro to the enzymic functionalization of the methyl group in 5'-deoxyhdenosine in vivo. N 4Nfj ++ ".;Y$ H& H Ox0 Me Me Me Me CHzZ (5) (6) (7) Finally chlorination of the triptycene derivative (7; X = Y = 2 = H) by sul-phuryl chloride initiated by benzoyl peroxide gives a mixture of the three rotamers (7; X or Y or Z = C1; other two atoms = H).44Chromatography on alumina destroys the two synclinal forms allowing isolation of the anti-periplanar rotamer (7; X = C1 Y = 2 = H). Kinetic measurements on the equilibration of the rotamers at 185 and 208 "C indicate a rotational barrier of 126 kJ mol-'. 5 Addition and Homolytic Aromatic Substitution Tedder and Walton have reviewed fifteen years of work on the addition of radicals to C=C double They concluded that for monosubstituted ethylenes steric hindrance to approach is the most important feature favouring attack at the methylene group.Polar influences are also important; methyl radicals are nucleophilic and their addition is facilitated by electron-withdrawing substituents in the alkene while trifluoromethyl radicals are electrophilic with reaction facili- tated by electron-releasing substituents. In view of the exothermic nature of most 42 (a)E. V. Blackburn and D. D. Tanner J. Am. Chem. Soc. 1980,102 692; (b) W. H. Tamblyn E. A. Vogler and J. K. Kochi J. Org. Chem. 1980 45 3912. 43 D. Gani A. W. Johnson and M. F. Lappert J. Chem. SOC.,Chem. Commun. 1980 1244. 44 T. Morinaga S. Seki H. Kikuchi G.Yamamoto and M. bki J.Am. Chem. Soc. 1980 102,1173. 45 J. M. Tedder and J. C. Walton Tetrahedron 1980 36 701. 60 R. A. Jackson additions there is an early transition state; thus stabilization of the incipient radical adduct which is much favoured by textbooks has only minor significance. In cyclizations of hex-5-enyl radicals to cyclopentylmethyl radicals (again not giving the thermodynamically favoured radical) 1-and 3-substituents favour cis -disub-stituted products whereas for 2-and 4-substituents the trans product is The chair-like transition state (8) was suggested with 2- 3- and 4-substituents taking up preferentially the 'equatorial' positions shown and thus giving the observed cis/ trans selectivity. These rules apply to ring-closure of peroxy- radical^:^' the reaction shown in Scheme 5 was reported to occur with complete stereospecificity.H2C H R' R2 R' R2 R;T$ 3R PhS -+ PhS H 0-0' 0-0 H Scheme 5 (8) Asymmetric induction has been observed in the reaction of thiols with (-)-menthyl cr~tonate.~~ For example an 18% enantiomeric excess of the (+)-enan- tiomer of the appropriate alcohol was obtained as shown in Scheme 6. i,ii I7 CH3COSH + CH3CH=CHC02Menth -CH3CO-*C-CH2CH20H Reagents i,azoisobutyronitrile (AIBN); ii LiAIH Scheme 6 E.s.r. evidence for the intermediacy of a cyclohexadienyl radical in the reaction of triphenylsilyl radicals with cyanobenzene has been The -CN group stabilizes the adduct radical (9)sufficiently to make the radical detectable and also accounts for the decrease in coupling constants compared with the unsubstituted cyclohexadienyl radical.Homolytic aromatic arylation of 4-methyipyridine and the 4-methylpyridinium ion by substituted aryl radicals p-XC6H4*shows that the isomer distribution is insensitive to X though the ratios of 2-:3-isomers change from 0.82 :1 to 9 :1 as the nitrogen atom is protonated.'' The total rate factors for arylation of the 4-methylpyridinium ion vary from 5.0 (X=Me) to 0.19 (X= NOz) but all values are in the region of 1 for 4-methylpyridine. Frontier-orbital calculations indicate that p -XC6H4 radicals behave as nucleophiles towards 4-methylpyridinium ion and as electrophiles towards 4-methylpyridine whatever the nature of X. (9) (10) 46 A. L.J. Beckwith T. Lawrence and A. K. Serelis J. Chem. SOC.,Chem. Commun. 1980,484. 47 A. L. J. Beckwith and R. D. Wagner J. Chem. SOC., Chem. Commun. 1980,485. 48 M. Yoshihara H. Fujihara and T. Maeshima Chem. Lett. 1980 195. 49 A. Alberti and G. F. Pedulli Gazz. Chim. Ital. 1979 109 395. R. Arnaud J. Court J. M. Bonnier and J. Fossey Noun J. Chim. 1980 4,299. Reaction Mechanisms -Part (iii) Free-Radical Reactions 61 Nitro-substituted aromatic compounds undergo homolytic aromatic substitution at unsubstituted positions but ipso-attack also occurs facilitated by the stability of the *NO2leaving radical. The thiophen derivative (10)reacts with methyl radicals by homolytic substitution at the 4-position whilst adamantyl radicals displace the nitro-group by ipso-attack.’l It has been suggested that polar influences stabilize the transition state for the addition of the relatively nucleophilic adamantyl radical.6 Fragmentation There is a growing interest in utilizing radical chain reactions that involve fragmenta- tion steps for synthetic purposes. Primary and secondary alcohols can be converted into the corresponding hydrocarbon in good yield by the radical-initiated reaction of the corresponding chloroformate with tripropylsilane (Scheme 7).52 Pr3Si-Pr3SiH ROH -B ROCOCl -Pr3SiC1+R-0-C=O -+ COz + R. -RH + Pr3Si. Scheme 7 A number of bridgehead iodides have been made from the corresponding carboxylic acids in reasonable yields by photolysis with t-butyl hypoiodite (Scheme 8y3 hv RC02H + Bu‘OI + Bu‘OH + RC02-I -+ R-COZ.-* COZ + R* -+ RI Scheme 8 A novel process involving a double fragmentation has been used in the decar- boxylation of carboxylic acids RC02H -+ RH (Scheme 9) via the dihydrophenan- threne derivative (11;X = C1 or PhS).54 RC0,H RH 1 t OCOR C02 + R’ 1 T + RC02‘ q \ Scheme 9 Deamination of aromatic amines can be carried out by their conversion into the diazonium salt with n-pentyl nitrite followed by photolysis to give the aromatic radical which gives the arene if THF is used as the solvent or the corresponding chloride or iodide if CCl or CH212 is ernpl~yed.~’ 51 P. Cogolli,F. Maiolo L. Testaferri M. Tiecco and M. Tingoli,J. Chem. SOC.,Perkin Trans.2,1980,1331. 52 R. A. Jackson and F. Malek J.Chem. Soc. Perkin Trans. 1 1980 1207. ’’ R. S. Abeywickrema and E. W. Della J. Org. Chem. 1980,45,4226. 54 D. H. R. Barton H. A. Dowlatsahi W. B. Motherwell and D. Villemin J. Chem. Soc. Chem. Commun. 1980,732. 5s V. Nair and S.G. Richardson J. Org. Chem. 1980,45 3969. 62 R. A. Jackson Amino-acids and amino-glycosides have been deaminated'6 by their conversion into the corresponding isocyanide and treatment of this with tributyltin hydride in the presence of AIBN at 80 "C (Scheme 10). I I1 iii RNHz +RNHCHO +RNC --+ RH Reagents i MeC0,CHO; ii POCl, Et,N; iii Bu,SnH AIBN at 80 "C Scheme 10 In the mechanistic area a number of stereochemical studies have been made. The reaction of trichloromethyl radicals with 2-(trimethylstanny1)butanegives cis-and trans-2-butene among the products with 75% of anti-periplanar eliminati~n:~~ the elimination may be concerted with anti and syn pathways or an initial anti abstraction of hydrogen may give an intermediate radical which can either immedi- ately collapse to the alkene (see Scheme 11)or undergo rotation before it produces the butene.CI,C* + /ySnMe -"Xiel + +.ke Scheme 11 Phosphoranyl radicals (12) can undergo p-scission5* to give the alkyl radical R* and the phosphate esters (RO),P=O. Fragmentation was thought previously to be preferentially from apical sites but kinetic e.s.r. studies now indicate that fragmenta- tion takes place predominantly from the equatorial positions. A number of e.s.r. kinetic studies of ring-opening reactions of compounds with a radical centre that is a to a cyclopropane or a cyclobutane ring have been made.Cyclopropylmethyl radicals undergo ring-opening much more readily than do cyclobutylme thy1 radicals. 59 7 Rearrangements Rearrangements are not as common in radical as in cationic systems but are still involved in a number of interesting processes. Amongst the products of reaction of recoil tritium atoms with CHF=CHF CHT=CF2 was unexpectedly found. The sequence shown in Scheme 12 was suggested.60 56 D. H. R. Barton G. Bringmann and W. B. Motherwell Synthesis 1980,68;J. Chem.SOC.,Perkin Trans. 1 1980 2665. 57 T. J. Stark N. T. Nelson and F. R. Jensen J. Org. Chem. 1980,45,420. B.P.Roberts and K.Singh J. Chem. Soc. Perkin Trans. 2 1980 1549.59 Y.Maeda and K. U. A. L. J. Beckwith and G. Moad J. Chem. SOC.,Perkin Trans. 2,1980,1083,1473; Ingold J. Am. Chem. SOC. 1980,102,328;K.U.Ingold and J. C. Walton J. Chem. SOC.,Chem. Commun. 1980,604. 6o E. E. Siefert and Y.-N.Tang J. Chem. SOC.,Chem. Commun. 1980 814. Reaction Mechanisms -Part (iii) Free-Radical Reactions H F \/ c=c /\F H-H T 1 T -H* Scheme 12 that a [1,2]hydrogen shift occurs first (to give the less stable n-propyl radical) followed by fragmentation as shown in Scheme 13. A similar 'uphill' rearrangement has been invoked to explain the formation of some 1,l-diphenylethane when bibenzyl is thermolysed62 at 366-400 "C (Scheme 14). (CH3)2CD*+ CH3CHDCH2. + CH3. + CHD=CH2 Scheme 13 PhCHzCH2Ph .(PhCH2)2 -B PhCH2. APhCHCH2Ph -+ Ph2CHCH2. -* Ph2CHCH3 Scheme 14 2-Tetralyl radicals (13) and 1-indanylmethyl radicals (14) prepared by thermoly- sis of their t-butyl peresters were found to interconvert readily equilibration via the intermediate neophyl-type radical (15) was In an attempt to observe a spiro-radical intermediate in reactions of this type hydrogen was abstrac- ted from the spiro-diene (16) but fragmentation of the intermediate radical (17) was fast and only the P-phenylethyl radical was observed by e.~.r.~~~ The acetylenic radical (18)was shown to rearrange to radical (19) at 45-88 "C,presumably via the cyclopropyl radical (20).64 61 I. Szirovicza and I. Sziligyi Znt. J. Chem. Kinet. 1980 12 113. 62 M. L. Poutsma Fuel 1980 59 335.63 (a)J. A. Franz and D. M. Camaioni J. Org. Chem. 1980 45 5247; (b) A. Effio D. Griller K.U. Ingold J. C. Scaiano and S.J. Sheng J. Am. Chem. SOC.,1980 102 6063. 64 K. U. Ingold and J. Warkentin Con. J. Chem.. 1980 58 348. R. A. Jackson ++-+%-+3-I (18) (20) (19) In experiments related to the enzymic conversion of diols into aldehydes and ketones photolysis of 4,5-dihydroxycyclo-octyl(pyridine)cobaloxime gave cyclo-octanone in 40% yield.65 A favourable transannular 13-H shift converts the 4,5-dihydroxycyclo-octylradical into the 1,2-dihydroxycyclo-octyl radical (Scheme 15). Scheme 15 8 Electron Transfer In the past decade there have been various suggestions for involvement of electron transfer in the addition of Grignard reagents to ketones.An intermediate charge- transfer complex (21) might collapse directly to adduct or break up to give Ph2COMgX and R*in a solvent cage. The free radical R.could now combine with Ph,d-O* *. .+ Mg-X R/ (21) Ph,COMgX to give the adduct or escape from the cage. Evidence for the inter- mediacy of a free radical in at least some cases is provided by the formation of the cyclic product (22) in 12% yield in the reaction shown; this cyclic product is likely I Me (22) [12%] I + Ph2C0 + + H2C=CHCH2CH2yCH2MgC1 I Me Me OH I I H2C=CHCHzCHzC-CH2-CPh2 I Me " B. T. Golding C. S. Sell and P. J. Seliars J. Chem. Soc. Perkin Trans. 2 1980,961. E.C. Ashby J. Bowers and R. Depriest Tetrahedron Lett. 1980 21 3541.Reaction Mechanisms -Part (iii) Free-Radical Reactions to have arisen by cyclization of the hex-5-enyl radical H2C=CHCH2CH2CMe2CH2 to give a substituted cyclopentylmethyl radical intermediate.66 When dimesityl ketone is allowed to react with metal hydrides such as AlH3 deeply coloured solutions are formed6' which show intense and complex e.s.r. spectra and which decay slowly as the reduction product is formed in 100% yield. It has been suggested that a radical anion/radical cation pair is formed as an intermediate as shown in Scheme 16. Mes2C0 + MH fast .+ __* Mes2C-0 M-H slow __* Mes2CH-OM Scheme 16 On mixing iodine with organometallic compounds R,M (M = Sn Pb or Hg) the absorption spectrum of a charge-transfer complex is observed immediately.The subsequent disappearance of this charge-transfer absorption is accompanied by iodinolysis of R,M at a rate which strongly depends on solvent polarity and this suggests that the rate-determining step is transfer of an electron from the alkylmetal to the iodine to form the ion pair [R,Mt I:] which subsequently gives the iodinolysis products.68 Polarographic reduction of substituted benzyl chlorides or bromides show Ham- mett plots which correlate best with K,suggesting that the potential-determining process involves C-Hal bond breaking or the formation of a radical ion intermedi- ate. Benzyl iodides react differently and form benzylmercuric iodide as an inter- mediate in the redu~tion.~~ Finally Walling has considered the general question of electron transfer in slow organic reactions exemplified by the reaction of an electron donor with an organic peroxide.It is commonly observed that when the donor is varied a plot of RT In k versus the one-electron oxidation potential of the donor is linear but with a slope of less than unity. This has been interpreted as an electron-transfer process which is neither rate-determining nor complete in the transition state and it has been pointed out that in the reaction of ROOR with a donor D the transition states to the different possible products {D' + RO' + RO-} {ROD * + RO a} and {ROD' + RO-} may be very similar.7o 67 E. C. Ashby A. B. Goel and R. N. DePriest J. Am. Chem. SOC., 1980,102,7779. 68 S. Fukuzumi and J. K. Kochi J. Am. Chem. SOC.,1980,102,2141.69 D. D. Tanner J. A. Plambeck D. W. Reed and T. W. Mojelsky J. Org. Chem. 1980,45 5177. '"C. Walling J. Am. Chem. SOC.,1980 102,6854.