4 Electron Spin Resonance Spectroscopy and Free Radical Reactions By W. T. DIXON Dept. of Chemistry Bedford College. Regent’s Park Inner Circle 1ondon NW7 4NS 1 Jntroduction Once again this year has been one of further consolidation rather than of innovation in the field of free radical chemistry. Leaving aside the explosion of interest in CIDNP and related topics which give information about what happens immediately before there is a change of spin multiplicity and more particularly about the species involved in such a step most studies of free radicals and their reactions have utilized electron spin resonance. Classical ‘kinetic’ studies of the type long-used in electrophilic aromatic substitution in which different molecules or sites compete for some active radical such as hydroxyl aryl alkyl or even hydrogen seem to be dwindling in the absence of any clearly defined aims or interpretation.This might be ex-pected since the emphasis tends to be on substituent effects in the molecules ‘attacked’ and not so much on the free radicals themselves. Any branch of chemistry which pertains to a class of compounds will consist mainly of their synthesis analysis and their chemical and physical properties. E.s.r. is an analytical technique par excellence for free radicals and is also useful in following their reactions i.e. by means of their rates of disappearance appearance of new radicals etc. Not only do e.s.r. spectra provide useful finger- prints for radicals the exact values of the coupling constants (and also the g-factors) depend rather critically on the geometry of the radical in question and this aspect is one that is being very actively pursued.The ‘synthesis’ of free radicals is achieved by relatively few methods. One usually starts with organic molecules in which all the spins are paired. Unpaired spins can only be introduced either by photons inorganic radicals or by thermal or mechanical (the same thing really) fission of bonds. A survey of the literature shows relatively few main groups of active workers each of which has developed expertise in a particular technique of getting free radicals. One might regard these techniques as ways of initiating the synthesis of radicals and they consist in the main of radiolysis (i.e. irradiation with prays) irradiation or reduction of peroxides nitro-compounds ketones etc.auto-oxidation and oxidation by metal ions such as Ce”’. Usually the radicals of interest arise from some secondary 151 I52 W. T. Dixon or tertiary process. For example hydroxyl radicals formed from the reduction of N20 by solvated electrons which are in turn formed in the y-radiolysis of water attack aromatic substrates to give in general adduct radicals. Apart from the predictable investigations of stable radical anions or nitroxides of more and more intriguing shapes and the filling in of 'gaps' in the array of simple radicals which have been studied the most interesting developments have been in stereochemistry. The conformations and configurations of radicals have been very much emphasized and some stereospecific radical reactions have been thoroughly investigated.As was the case last year it seems that the best way of classifying the material is in terms of the types of radicals. 2 Simple Alkyl and Aryl Radicals Remarkably strong steric interactions of fluorine atoms exist in fluorinated alkyl radicals so much so that fluorinated ethyl radicals invariably give e.s.r. spectra which are temperature dependent implying hindered rotation. For example CF2CH has an internal rotation barrier of about 2 kcal mol-' and exists in an equilibrium between the three equivalent forms (1)+3).' It is unusual 'a ,Hb Hc ,Ha \I \I \\I L C -C //\ /;\ F:F F:F Hc Hb ,Hc '\ /' to be able actually to observe directly restricted rotation of a methyl group and the effect here may be enhanced by the fact that the geometry at the a-carbon atom is pyramidal and not planar as in ethyl itself since the arrangement about the a-carbon deviates more and more from planarity as hydrogen alkyl or even CF groups2 are replaced by fluorine atoms.This trend is deduced from the changes in 13C or 19F coupling constants; for example a(a-F) increases with the number of fluorine atoms attached directly to the a-carbon atom. Rather unexpectedly perhaps symmetrical conformations are the most stable for P-fluoroethyl radicals whether there be one two or even three p-fluorine atois. This has been deduced from the size of the P-proton coupling constants in (4) and (5) though the e.s.r.spectra* of these radicals are temperature depen- dent showing hindered rotation ' 'K. Schen and J. K. Kochi Chem. Phys. Letters. 1973 23 233. R. V. Lloyd and M. T. Rogers J. Amer. Chem. SOC.,1973 95 1512. I. Biddles J. Cooper A. Hudson R. A. Jackson and J. T. Wiffen Mof. Phys. 1973 25 225. *When not specifically stated splittings are given in units of G (=lo-* T). Electron Spin Resonance Spectroscopy and Free Radical Reactions The e.s.r. spectra of chloro- and bromo-alkyls also generally show some tem- perature dependence as one might expect and there is a certain amount of rather flimsy theoretical and empirical evidence that there are unsymmetrical halogen bridges in molecules such as CH,CH,Cl and CH,CH,Br. This idea of halogen bridging is deduced from low values of P-proton splitting~,~.' trans addition reactions,6 and also from CIDNP observed in the n.m.r.spectra of the products of addition reactions.' Perfluoroalkyl radicals formed by the U.V. irradiation of perfluorinated fatty-acid salts of Pb'" abstract protons but not fluorine atoms from -CHF or -CH,F groups,8 and methyl radicals abstract hydrogen from the a-positions of side-chains of alkylbenzenes in preference to adding to the aromatic ~keleton.~ This contrasts somewhat with the behaviour of phenyl," thiazolyl,' or even hydrogen atoms,' which tend to add to aromatic rings in a rather unselective way. In such studies the isomer distribution of the products may not be simply related to the relative reactivities of the original positions.lo A dramatic decrease in attack ortho to a t-butyl group'.' shows the importance of steric factors in such reactions. Phenyl radicals will not only abstract aliphatic hydrogen atoms but also iodine and since the rate of this process with (6)is the same as that with (7),there can be no 'anchimeric assistance' from sulphur in the latter case.13 This may be attributed to a large dihedral angle making such assistance via S-bridging unfavourable. 3 Benzyl-type Radicals The technique of y-irradiation of aqueous solutions containing nitrous oxide has proved to be an efficient way of generating organic radicals uia the hydroxyl I. H. Elson K. S. Chen and J. K. Kochi Chem. Phys. Letters 1973 21 72. K. S. Chen I. H. Elson and J. K. Kochi J.Amer. Chem. SOC.,1973 95 5341. P. S. Skell R. P. Pavlis D. C. Lewis and K. J. Shea J. Amer. Chem. SOC.,1973,95,6735. ' J. H. Hargis and P. B. Sherkin J.C.S. Chem. Comm. 1973 179. P. B. Ayscough J. Machora and K. Mach J.C.S. Faraday II 1973,69 750. S. J. Hammond and G. H. Williams J.C.S. Perkin II 1973 484. lo J. T. Hepinstall jun. and J. A. Kempmeier J. Amer. Chem. SOC.,1973 95 1904. ' I G. Vernin H. J. M. Dou and J. Metzer J.C.S. Perkin II 1973 1093. W. A. Pryor T. H. Liu J. P. Stanley and R. W. Henderson J. Amer. Chem. Soc. 1973 95 6993. l3 W. C. Danen D. G. Saunders and K. A. Rose J. Amer. Chem. Soc. 1973 95 1612. 154 W. T. Dixon radical or its anion (Scheme 1). The 0' ion abstracts hydrogen readily from alkyl groups and so generates for example benzyl radicals from alkylbenzenes.The electronic spectra of these radicals have been observed in pulse radiolysis y-rays e Scheme 1 experiment^'^ and it has been shown that the mechanism in strongly alkaline solution is direct hydrogen abstraction whereas in more acidic solutions (below pH -10)a more complex process occurs (Scheme 2). In both extreme cases the C6H,CH3 + -OH -.+ [C,H,CH,(OH)] % C,H,CH2-etc. observed Scheme 2 main product is bibenzyl. The reaction in alkaline solution has been utilized to generate many substituted benzyl radicals,' which are remarkable in that the coupling constants depend only very slightly on the nature of the substituents e.g. (8) and (9). s0,-0-( 8) (9) The signs of the fluorine coupling constants in 0-,m-,and p-fluorobenzyl have been determined using CIDNP,l6 since the polarizations induced depend on the signs of the coupling constants.The reaction used was the photolysis of the appropriate benzylphenone in carbon tetrachloride (Scheme 3). The 0-and l4 H. C. Christensen K. Sehested and E. J. Hart J. Phys. Chem. 1973,77 983. '' P.Neta and R. H. Schuler J. Phys. Chem. 1973 77 1368. I6 (a) J. Bargon and K. G. Seigert J. Phys. Chem. 1973 77 2877; (b) D. Bethell M. R. Brinkman and J. Hayes J.C.S. Chem. Comm. 1972 1324. Electron Spin Resonance Spectroscopy and Free Radical Reactions ArCH,COPh ‘2‘ ArCH,. + SCOPh * CCI , 1 [ArCH,. *CCI,] I pola;:;tion ArCH,-CCI (polarized n.m.r. spectrum) Scheme 3 p-fluorine coupling constants were positive but the m-F was negative as expected.Similar empirical conclusions have been deduced from studies of aryl semi- quinones.” The conformation of adifluorobenzyl radicals is a planar one in the absence of ortho substituents.’* This has been deduced from the fluorine splittings which are about 50 G compared with 56 G in CFHCONH, known to be planar and with 72 G in CF,CONH, which is pyramidal. 4 Cyclic Adduct Radicals The y-radiolysis of water gives solvated electrons which can be effectively re- garded as hydrogen atoms these may add on to suitably active species such as furan to give adducts analogous to cyclohexadienyl. As we have already said the presence of N,O leads to hydroxyl radicals which can add on similarly or abstract active hydrogen atoms.Thus the parent radicals of the furan series” can be obtained in concentrations sufficiently large for detection by e.s.r. (Scheme 4). H Scheme 4 P. Ashworth and W. T. Dixon J.C.S. Perkin II 1973 1533. L. D. Kispert. H. Liu. and C. U. Pittman jun. J. Amer. Chem. Suc. 1973 95 1657. l9 R. H. Schuler G. P. Laroff and R. W. Fessendem J. Phys. Chem. 1973,77,456. 156 W. T. Dixon In base when the active species in the presence of N20is O' 2,5-dimethylfuran yields an analogue of benzyl (Scheme 5). 2-Nitrofurans lose the nitro-group 13.2 Scheme 5 when subject to attack by *OHin a similar way,20 but in the case of l-bromo- furanoic acid the ring opens (Scheme 6). Similar types of adducts are formed * 02C0Br 0 -HO Br 02'no] Scheme 6 with pyrole derivatives,21 although the parent molecule itself reacts furtner to eliminate water (Scheme 7).02cQI .OH) 02cUH3.4 \ OH IH H0.28 Scheme 7 C. C. Greenstock I. Dunlop and P. Neta J. Phys. Chem. 1973,77 1187. 21 A. Samuri and P. Neta J. Phys. Chem. 1973,77 1629. Electron Spin Resonance Spectroscopy and Free Radical Reactions The tendency of OH to add to unsaturated systems wherever that is possible is again illustrated by the radiolysis of aqueous ally1 alcohol22 in the presence of N,O (Scheme 8). CH,=CHCHOH Y Scheme 8 Renewed interest in Fenton's reagent has led to further exhaustive studies of the isomer distributions of products from the hydroxylation of nitrobenzene chlorobenzene toluene etc.,' and a large number of hydroxycyclohexadienyl- type radicals have been identified by means of e.s.r.spectroscopy in the oxidation of benzene- and pyridine-carboxylic acids23 by Fenton's reagent (Scheme 9). Fe" + H,O -+ Fe"' + -OH + -OH Scheme 9 5 Further Radicals from Peroxides Fenton's reagent has also been used to oxidize unsaturated aliphatic com- pound~~~,~~ and some interesting reactions have developed. From alkynes a number of possibilities can occur depending on whether substrate-reducing or -oxidizing agents are present in excess (Scheme 10). Similar types of situation can occur with Ti"'-H,02 and e.s.r. kinetic studies indicate that the acetonyl radical is reduced by Ti"' quite effectively in neutral solution.26 Limitations in the kinetic approach not only to the Ti1"-H,02 system but also to steady-state situations in general have been pointed out again.,' During the photolysis of maleoyl phenyl peroxide,28 fission of the peroxy bond leads to the formation of two carboxyl radicals in the same molecule and CO is lost stepwise so that there is appreciable probability of again forming 22 M.K. Eberhart and M. Yoshida J. Phys. Chem. 1973 77 589; M. Simic P. Neta and E. Hayon ihid. p. 2662. *' T. Shiga T. Kishimoto and E. Tomita J. Phys. Chem. 1973,77 330. 24 C. Walling and G. M. El Taliawi J. Amer. Chem. Soc. 1973,95 844. 25 C. Walling and G. M. El Taliawi J. Amer. Chem. SOC.,1973 95 848. 26 B. C. Gilbert R. 0.C. Norman and R. C. Sealy J.C.S.Perkin If 1973 21 74. 27 D. Meisel G. Czapski and A. Samuni J.C.S. Perkin fI 1973 1702. 28 M. M. Martin and J. M. King J. Org. Chem.. 1973. 38. 1588. 158 W. T. Dixon CH,CHO HO CH-CH .OH ~ (Fell-H 20z) \ / C=kH H HOCH,CHO Me \ MeCGCMe 'OH b C=eMe Ho\ /c,'c/ Me Me \ Me CMe .4 MeC Scheme 10 a bond between the two radical fragments (Scheme 11). In contrast to alkyl- carboxyl radicals alkoxy-adducts with CO are comparatively long-lived and 0 It PhC -C I1 \ 0 HC. p.0' phgo Ph PhCZCH Polymers 0 PhC. 0. tI / HC-C \ phq 0 0 Scheme 11 add on to double bonds2' to form carbonates (Scheme 12). However in the absence of alkenes the e.s.r. spectra of the alkoxyl radicals only are observed.6 Radical Reactions of Cyclopropyl Derivatives It is not difficult it seems to form cyclopropane by a free radical path. For example 1,3-di-i0dopropane reacts with benzoyl peroxide in carbon tetra- 29 D. J. Edge and J. K. Kochi J. Amer. Chem. Soc. 1973,95 2635. Electron Spin Resonance Spectroscopy and Free Radical Reactions 0 / RO-C 0 \ / 0 RO-C l-\ 7 #>i< 0 hv 0-/ 0 RO-C I C \\ + 0 (RO. + CO2) RO’ ‘\o Scheme 12 chloride to form amongst other things cyclopropane (Scheme 13).30 However the three-membered ring is easily opened by radical attack and some interesting Bz202+2Bz0.~2Ph. + 2C02 ICH,CH,CH,CI ICH2CH2CH21~ICH2CH2CH2 -A Scheme 13 stereochemistry results. The radical bromination of cis,cis-trimethylcyclo-propane31 gives only three of the possible products which may be formed by the sequence shown in Scheme 14.The observed products show that the initial bromine attack was from an equatorial direction and this has been confirmed in a number of other cases.32 Generally speaking cyclopropyl itself can only be observed at low temperatures since it tends to open to form an allyl-type radical (Scheme 15).33 Thus ring- opening occurs by fission of the bond opposite to the site of the odd electron. Facile ring-expansion can occur if there is an odd electron on a side chain34 attached to the cyclopropyl ring (Scheme 16) and this is really the same type of reaction as the ring-opening by a bromine atom. A similar type of reaction takes place in cyclopropoxyl radicals but there the formation of a ketonic group is preferred to the formation of a four-membered ring (Scheme 17).35 ’O A.F. Drury and L. Kaplan J. Amer. Chem. SOC.,1973,95 2217. ” G. G. Maynes and D. E. Appleguist J. Amer. Chem. SOC.,1973,95 856. ’‘ K. J. Shea and P.S. Skell J. Amer. Chem. SOC.,1973 95 6728. ’’ K. S. Chen D. J. Edge and J. K. Kochi J. Amer. Chem. SOC.,1973,95 7036. )‘M. P.Doyle R. W. Raynolds R. A. Barents T. R.Bade W. C. Danen and C. T. West J. Amer. Chem. SOC.,1973 95 5988. ” C. H. De Puy H. L. Jones and W. M. Moore J. Amer. Chem. Sor. 1973 95 477. 160 W. T. Dixon Me M&e -+ Br< (equatorial approach) Br (threo) \ TBrMe dl (500/,) Scheme 14 (seen at low temperatures) Scheme 15 Scheme 16 Scheme 17 161 Electron Spin Resonance Spectroscopy and Free Radical Reactions 7 Further Aliphatic Radical Reactions Alcohols may undergo hydrogen abstraction from the hydroxy-group during the decomposition of persulphate ion3' catalysed by photons or reducing metal ions such as Ag'.The reaction is supposed to proceed by simple electron transfer R'R'CHOH '0,;[RIR'CHOH]~-~ R'R'CHO-The alkoxyl radical may be trapped by the N-oxide PhCH=N which \ Bu' adds on alkyl and alkoxyl radicals to give nitroxides; the latter may be identified from their e.s.r. spectra. In a particular case the reaction proceeds as shown in Scheme 18. The mechanism has been elucidated by means of the ingenious PhCH,CH,OH so -* PhCH2CH20* -+ PhCH,.+ CH,O PhCH OH PhCHZCH2 Ph PhCH =NOBu' 1' PhCHO 0 But / \ PhCH -N N-0 I\ / PhCH,CH,O But PhCH Scheme 18 device of conducting it first in the presence of the above radical trap which evidently stops the reaction before the break-up of the alkoxyl radical and then separately in the presence of t-butyl pitrone which traps alkyl radicals but not alkoxyl radicals. The adduct of PhCHCH,OH is a very minor constituent of the nitroxide product of which the mqjor component is benzyl t-butyl nitroxide. An interesting radical rearrangement occurs3' when the halogen atom is abstracted from a b-bromoalkyl ester by trialkyltin hydride plus t-butoxyl radicals (Scheme 19). The most plausible mechanism seems to be that the ester group 'walks' to its new position (Scheme 20).It is interesting that in the case when R2 = phenyl it migrates perhaps via formation of a cyclohexadienyl-type of intermediate (Scheme 21). Intramolecular radical cyclization also takes place in the addition of perfluoroalkyl iodides to hepta-1,6-diene in preference to simple addition to one of the double bonds (Scheme 22).38 36 A. Ledwith P. J. Russell and L. H. Sutcliffe J.C.S. Perkin /I 1973 630. 37 A. L. J. Beckwith and C. B. Thomas J.C.S. Perkin I/ 1973 861. 3* N. 0.Brace J. Org. Chem.. 1973. 38 3167. 162 W. T. Dixon R’ R’ / I 0-c=o PC=O / Me-C-CH,Br Bu3Sn; MeC I I\ R’ R’ CH,. Bu,SnH 1 MeCHRZCH,OCOR1 Scheme 19 1 Product Scheme 20 -$ Me MeC-kH Met-CH, I OAc I OAc OAc Scheme 21 Scheme 22 The radical addition of t-butyl hypochlorite to ~inylacetylene~~ starts with addition of t-butoxyl to one end of the molecule and thereafter chlorine goes on at one of the others (Scheme 23).’’ M. L. Poutsma and P. A. Ibarbia. J. Amer. Chem. SOC.,1973. 95 6000. Electron Spin Resonance Spectroscopy and Free Radical Reactions BuOCl \ CH,=CH~=CHOBU BuOCH2CHC1CECH BuOCH,CH=C=CHCl CH,=CHCCl=CHOBu 1 (CH,CICH=CClCHO) Scheme 23 8 RadicaIIons A particularly interesting anion which has been generated with great difficulty is that of hexaflu~robenzene,~' i.e. C6F6- which although it would seem super- ficially to resemble C6H6-is evidently completely different since the six fluorine nuclei have spIittings of 137G.The reason for this enormous coupling constant is that the odd electron is in a Q rather than a n molecular orbital. This is itself a result of the large number of electrons in the n molecular orbitals owing to the fact that each fluorine atom contributes two n-electrons. Because of this the odd electron goes into the first empty Q anti-bonding orbital instead of the first n* orbital as is the case with the benzene negative ion. Dinegative ions of benzen-oid carboxylic acids have been produced in basic media by reduction with solvated electrons4' and the carboxylate group has a similar effect to oxygen in phenoxyl or semiquinone ions so that the coupling constants rather resemble those in those radicals. These splitting patterns can be understood in terms of the symmetry of the benzene MOs as shown in Scheme 24 the substituents co; i0 c02-.coz-0 0 0.05 0.75 Scheme 24 having the effect of putting the odd electron into the orbital with the greatest value at the adjacent carbon atoms. *O L. F. Williams M. B. Yin and D. E. Wood J. Arner. Chem. Soc. 1973 95 6475. ** P. Neta and R. W. Fessenden J. Phys. Chem. 1973,77,620. 164 W. T. Dixon The heptafulvalene trinegative ion4 has been prepared and there is coupling only with the protons of one ring. Since this is large and in view of the strong repulsions between the two ring systems arising from three negative charges it seems that the two rings could be perpendicular to each other as shown in (10).Some naphthalene negative ions have been generated usually electrolytically and it is found that t-butyl or trimethylsilyl groups attached to the naphthalene skeleton do not have very much effect on the spin distrib~tion.~~ However bridges across the peri-positions can have much greater as well as providing rigid frameworks through which spin density can be transmitted relatively efficiently e.g. (1 1) and (12).45 HH I19 014 1.1 (12) The negative ions of halogenoquinolines have been studied p~larographically~~ and from the half-wave potential loss of chloride ions could be deduced. Thus both quinoline itself and 6-fluoroquinoline gave simple half-wave potentials corresponding to the formation of their negative ions. 6-Chloroquinoline on the other hand gave two half-wave potentials one corresponding to that of quinoline.This is attributed to the sequence shown in Scheme 25. It is doubtful however that the sigma 6-quinolinyl radical' would ever actually have a discrete existence. A disproportionation step in the presence of protons seems more likely. Steric strain is apparent from the e.s.r. spectra of 4,4'-polymethylene- biphenyl radical anions (13),47 but when the methylene chain is longer than 15 CH units the coupling constants indicate that strain is virtually absent. 42 N. L. Bauld C. S. Chang and J. H. Eilert Tetrahedron Letters 1973 153. " A. G. Evans B. Jerome and N. H. Rees J.C.S. Perkin II 1973 2091. 44 S. F. Nelsen and J. P. Gillespie. J. Amer. Chem. Soc..1973 95 1874. 45 S. F. Nelsen and J. P. Gillespie J. Amer. Chem. SOC.,1973 95 2940. 46 K. Alwain and J. Grimshaw J.C.S. Perkin II 1973 1811. 47 K. Ishizu F. Nimoto H. Hasegawa K. Yamunoto and N. Nakazaki Bull. Chem. Soc. Japan 1973 46 140. 165 Electron Spin Resonance Spectroscopy and Free Radical Reactions . ,..-. % ,.__. first A,+ N le- Scheme 25 Incidentally aromatic radical anions can be produced advantageously by using trimethylsilyl~odium~~ rather than sodium metal dispersed in a suitable solvent. Further e.s.r. work has been done on the auto-oxidation of hydroquinones and quin~nes’’~~~~~~ and dimeric intermediates e.g. (14) and (13,positively -? 0 identified.49b It is interesting that whereas spin density is transmitted around the ring of an aryl substituent attached to semiquinone the carbonyl groups of a quinonoid substituent seem to present a barrier to the odd electron.The e.s.r. spectra of some of these dimeric species were originally attributed to radicals of different and this illustrates the dangers of trying to identify radicals only on the basis of the number and type of coupling constants in an e.s.r. spectrum. A remarkable feature about some ‘heterocyclic’ ~emiquinones~’ is that the second rings containing nitrogen atoms e.g. (16) and (17) seem to have but little effect on the spin density which remains mainly on the semi- quinone part. 40 H. Sakumi. A. Okada H. Umino and M. Kira J. Amer. Chem. SOC.,1973,95,955. 49 (a)J. A. Pedersen J.C.S. Perkin II 1973,424; (6) P.Ashworth and W. T. Dixon J.C.S. Perkin II 1973 2128. M. K. V. Nair K. S. V. Santhanaan and B. Venkataraman J. Magn. Resonance 1973 9 229. 166 W. T. Dixon 0' 0' A number of naphthoxyl radicals e.g. (18) and (19) have been identified by e.~.r.~l and it seems that substituents have little effect on the spin distribution which is slightly different in the case of fl-naphthoxyl(l9) from that expected by comparison with phenoxyl and naive MO theory. 2.5 10.75 1.45 0.0 A ring-closure reaction has been observed when suitably hydroxylated benzo- phenones are oxidized by alkaline ferricyanide to xanthones (Scheme 26).52 It would be interesting to see whether it is the dihydroxylated ring which under- 0-major product minor product Scheme 26 goes the initial attack because although the radical cyclization uia the oxygen shown in Scheme 26 looks feasible the spin density on the oxygen atom must be small whereas in the para-position of phenoxyl it is large.'' W. T. Dixon W. E. J. Foster and D. Murphy J.C.S. Perkin 11 1973 2124. 52 P. D. McDonald and G. A. Hamilton J. Amer. Chem. SOC.,1973.95 7752. Electron Spin Resonance Spectroscopy and Free Radical Reactions 9 Radicals containing Group IV Elements Several more examples of sH2 reactions of tin and lead compounds have been reported and these can be grouped into those which occur at a halogen atom :53*54 RCI + SnMe -P R-+ ClSnMe which are of course already well-known and which are used extensively in the case of alkylsilyl radicals for the production of alkyl radicals and those which take place at the Group IV atom:” BuO.+ R,SnX,- + BuOSnR,-,X,- + R-(X = halogen) This last type of reaction is to be contrasted with the reaction with tin tetra-alkyl-tin where hydrogen abstraction takes place in preference to substitution :” R’CH2SnR2 + BuO. -P BuOH + R’t?HSnR2 The rates of addition of Group IV radicals to carbonyl compoundss6 are in the order R,Si-> R,Ge. -R,Sn. > R,Pb. and the rate of this reaction with a given radical R3M* is in the order diketone > oxalate > ketone > trifluoroacetate > acetate The reactions of the silicon analogue of the t-butoxyl radical have been studied via photolysis of the peroxide:” Me,SiOOBu‘ !% Me,SiO* + Bu’O* This radical reacts with the starting material to give a silicone ether by an sH2 mechanism Me,SiO.+ Me,SiO,Bu + Me,SiOSiMe + *02Bu In the presence of alkenes both primary radicals may add Me,SiO* + CH,=CH2 -P Me,SiOCH,CH,. and with buta-1,3-diene there are several possibilities such as straightforward addition or more complex reactions e.g. Scheme 27. In the case of organosilyl ” J. Cooper A. Hudson and R. A. Jackson J.C.S. Perkin II 1973 1056. ’* D. A. Coates and J. M. Tedder J.C.S. Perkin II 1973 1570. 55 A. G. Davies and J. C. Scaiano J.C.S. Perkin 11 1973 1777. 56 J. Cooper A. Hudson and R. A. Jackson J.C.S. Perkin If 1973 1933. 57 D. J. Edge and J. K. Kochi J.C.S. Perkin 11 1973 182. 168 W. T.Dixon Me,SiO,Bu I SiMe SiMe3 SiMe (cis-trimethylsilylallyl) Scheme 27 derivatives of hydroxylamine rearrangement to the nitroxide can occur :'* Bu,O 2Bu0. R,SiNHOSiR3 BU0.r R,SiNa-SiR 1 (R,Si),N'O 10 Radicals containing Nitrogen Nitroxides are still probably the most widely studied radicals because of their ease of formation use of nitroso-compounds as radical traps and their stability. They are so stable that large enough concentrations can be achieved to enable one to obtain their n.m.r. spectra in some cases. From the n.m.r. spectrum one can obtain the signs of the coupling constants since they determine the direction of the 'contact' chemical shift. This exercise has been completed for a-and for P-naphthyl t-butyl nitroxides (20) and (2l)." The spin distributions bear 0' I 13.5 NBu -0.218 NBu -0.064 -0.101 -0.43 0.39 0.865 almost no relationship to those in a- and P-naphthoxyi and in the case of the a-naphthyl derivative the coupling constants suggest that there is some degree of twisting about the carbon-nitrogen bond.A variety of cyclic nitroxides have 58 R. West and P. Boudjouk,J. Amer. Chem. SOC.,1973,95 3983. 59 J. L. Duncan A. R. Forrester G. McConnachie and P. D. Mallinson J.C.S. Perkin If 1973. 718. Electron Spin Resonance Spectroscopy and Free Radical Reactions 169 been prepared and the spin label allows investigation of the flapping of rings6' as well as their conformations6' to be determined from the e.s.r. parameters. An interesting class of nitroxide radicals has been studied using the spin trap (22).Addition occurs at the carbon atom and the radical formed may be strongly Bu' / R.=C=N -RCH,NB~~ -\ I Scheme 28 asymmetric i.e. Scheme 28.62 Looking along the methylene-carbon-nitrogen bond there will be two stable conformations (23) and (24). If R is an asymmetric HA \\ \ -0-N -But I group such as CH,CHOH then even when there is fast interconversion between these two conformations the two protons will have different average splittings. The disparity increases with the size of R.. If there is sufficient steric hindrance alkyl groups may be forced to add to the oxygen of a nitroso group rather than to the nitrogen atom (Scheme 29).63a From the coupling constant of the nitrogen atom (ca.10G)the N-alkoxyanilino- radical is probably a n-type and so are the aliphatic analogues e.g. Scheme 30.63b On the other hand iminoxyl radicals are of a-t~pe,~~ as are iminyl radicals (Scheme 31). An interesting rearrangement takes place after radical attack on aryl hydra- zonates (Scheme 32).65 To finish this section on nitrogen radicals it seems appropriate to mention the reduction of nitroxides to the corresponding secondary amines with iron carbonyls (Scheme 33).66 'O R. E. Rolfe K. D. Sales and J. H. P. Utley J.C.S. Perkin 11 1973 1171. " V. Malatesta and K. U. Ingold J. Amer. Chem. SOC.,1973 95 6390 6395; J.C.S. Perkin 11 1973 2134; S. F. Nelsen and R. T. Landis jun. J. Amer. Chem. SOC.,1973 95 6454. 62 B. C. Gilbert and M.Trenwith J.C.S. Perkin 11 1973 1834. 63 (a)S. Terabe and R. Konaka J.C.S. Perkin 11 1973 369; (b)W. C. Danen C. T. West and T. T. Kensler J. Amer. Chem. SOC.,1973,95 5716. "G. D. Mendenhall and K. U. Ingold J. Amer. Chem. SOC.,1973,95,2963 3422 65 A. F. Hegarty J. A. Kearney and F. L. Scott J.C.S. Perkin 11 1973 1422. 66 H. Alper J. Org. Chem. 1973 38 1417. 170 W.T. Dixon (nitroxide) M = C Si or Sn (N-alkoxyanilino-radical) scheme 29 0 4 RNHC hv RNH (amino-radical) \ OOBu' (proxy-carbamate) *"". I RNOBU' RNHOBu' Scheme 30 Bu,C=NH Bu'o. Bu,C=N* -% BuC=N-o* 2.8 28.4 (sigma radical) \ INo Bu,C=NNO Scheme 31 11 Radicals containingGroup V Elements The work on phosphoranyl radical^^',^^ and the corresponding arsenic com- pounds has continued along the same lines as last year.The structure of these radicals is one of a trigonal bipyramid with alkoxyl or halogen ligands filling 6' D.Griller and B. P. Roberts J.C.S. Perkin 11 1973 1339. 68 D. Griller and B. P. Roberts J.C.S. Perkin 11 1973 1416. Electron Spin Resonance Spectroscopy and Free Radical Reactions 0 I Ph1’ PhC=N-NHPh =N-NPh X 9‘ 0-Ph&=N-NPh 3 0 Ph‘ II Ph’ PhC-N-L-Ph ;;-; PhCONHN’ \ Ph Scbeme 32 Fe,(CO) I Q-Q I I 0 H Scheme 33 the axial positions preferentially e.g. (25) and (26). The phosphorus and arsenic compounds undergo similar types of reaction (Scheme 34). Some chelating ligands have been attached to phosphorus yielding ‘spiro- phosphoranyl’ radicals,68 and exchange of halogen between equatorial and axial positions has been observed by e.s.r.(Scheme 35). 12 Radicals containingGroup VI Elements The pulse radiolysis of thiols or ‘sulphydryl’ compounds leads to sulphur radicals which readily form links with further sulphur to give a negative ion in which 172 W. T. Dixon BuOPX +X. 7 PX,WBuOPX L Bus +OPX BuOAsPh +Ph-2 Ph,As%BuOAsPh L Bu. +OAsPh Scheme 34 CI OR 1 ,OR I,c1 CCP EP I\OR I \OR OR OR Scheme 35 there are two sulphur atoms (Scheme 36).70371This well-known oxidative coupling of sulphur is so strong a tendency that it appears that it can even happen with dialkyl sulphides (Scheme 37).72 A great variety of radicals may be generated from such aliphatic sulphur cation radicals for example when there is aa-hydroxy- group (Scheme 38).73 te-RSSR H *idisproportionat ion /SH s R +R’I \ \I SH S Scheme 36 Some interesting thiophen derivatives (27H29) have been prepared in which different isomers can be distinguished in the e.s.r.spe~tra.’~ 69 E. Furirnsky J. A. Howard and J. R.Morton J. Amer. Chem. SOC.,1973 95 6574. ’’M. Z. Hoffman and E. Hayon J. Phys. Chem. 1973,77,990. P. C. Chan and B. H. J. Bielski J. Amer. Chem. Sor.. 1973 95 5504. ’’B. C. Gilbert D. K.C. Hodgeman and R.0.C. Norman J.C.S. Perkin 11 1973 1748. 73 B. C. Gilbert J. P.Larkin and R.0.C. Norman J.C.S. Perkin If 1973 272. 74 C. M. Camaggi L. Lunazzi and G. Placucci J.C.S.Perkin 11 1973 1491. Electron Spin Resonance Spectroscopy and Free Radical Reactions +. +. Me,S .OH Me,S b Me,SSMe, (Ti"l-H,O, 10H 6.8 (12 protons) CH,SMe 16-5 3.6 Scheme 37 +. RCHOHCH,SR -+RCHO +CH,SR + H+ L R \ .+ C=O + H+ + SR / H3C (trapped but not observed directly) Scheme 38 R\S Qs--0 s .- s I I I I R I R R (27) &cis (28) cis,trans (29) trans,trans Sulphonyl radicals (30)and (31) have been prepared by abstraction of chlorine from sulphonyl chlorides 75 with the triethylsilyl radical. The assignments were made on the assumption of twisting of the -SO,. and that the configuration CH,CH,CH,SO,CI b CH ,CH2CH,SO5 0.7 2.1 7 (30) at the sulphur is pyramidal the preferred conformation leading to asymmetry.This is rather different from the case of the primary alkyl sulphonyls (32)-(34) where rotation is relatively unrestricted it seems. 13 Spin Delocalization through a-Bonds The existence of substantial P-proton coupling constants in alkyl radicals and in alkyl-substituted aromatic radicals proves in itself that spin density 7s A. G. Davies B. P. Roberts and B. R.Sanderson J.C.S. Perkin If 1973,626. 174 W. T. Dixon can be transmitted through o-bonds. This process which arises from the inter- action of a n-type atomic orbital and the orbitals of a neighbouring a-bond is called hyperconjugation and the idea can be extended to account for some very long-range coupling constants e.g. those in (35)and (36).'6 10.8 = 1.5 In the bridged naphthalene anion already mentioned there is even an observable S-coupling constant.These long-range splittings generally are observed in systems in which a rigid conformation exists which is favourable for each hyper- conjugative interaction. There are some empirical rules which have some foundation in simple MO theory which amount to saying that transmission through a a-system only occurs when the bands are lined up with each other. More sophisticated the~ries'~ are able to account for negative spin densities on y-protons as well as positive densities on p-and &protons when the conforma- tion is favourable (see the Figure). Particular cases of o-delocalization occur when the steric forces cause twisting in aromatic systems.' 7g78*79 The coupling constants of protons attached to aryl substituents generally arise from a mixture of two components one from transmissions of spin density through the n-system and the other through the a-system.Twisting becomes apparent when the ortho and para proton splittings decrease in relation to the meta proton coupling constants. When the twisting becomes much more pronounced a situation is found in which the meta splitting is much larger than the others e.g. in the rubrene positive or negative radical ions (37).79 A particularly clear example of this effect appears in the series of aryl semiquinones (38)--(40).'7 The large meta coupling which is apparent when the aryl ring is approximately perpendicular to the nodal plane of the odd ele~tron,'~ is simply a case of S-coupling described before.76 J. K. Kochi P. Bakuzis and P. J. Krusic J. Amer. Chem. Soc. 1973 95 1516. 7' Y. Ellinger A. Rassat R. Subra and G. Berthier J. Amer. Chem. SOC.,1973,% 2372. H. G. Aurich H. Forster A. Lotz and W. Weiss Chem. Ber. 1973 103 2832. l9 R. Biehl K. P. Dinse K. Mobius M. Plato and H. Kurreck Tetrahedron 1973,29,363. Electron Spin Resonance Spectroscopy and Free Radical Reactions ab > 0 Figure Diagram showing interactions responsible for transmission of spin density through a-bonds. Orbitals in which there is relatively high positive spin density are shaded. Ph Ph Ph Ph (37) O @27 P (0.05).13 e0.0 \ 0.16 / 0.13 / / 0.15 0.27 0. I 3 \ / Me \ (0.05) 0-0-0-(38) (39) (40) increasing dihedral angle + 14 Some Miscellanies There has been some interest in the structure of methyl-lithium which is a tetramer consisting of a tetrahedron of lithium atoms with a methyl group by 176 W.T. Dixon the centre of each face. The e.s.r. spectrum of the corresponding alkyl radical,80 in which one hydrogen atom has been removed has been obtained and shows as expected hyperfine coupling with three equivalent 7Li nuclei [1(7Li) = $1. This is consistent with the structure (41). There must be rapid rotation of the methylene group to account for the equivalence of the three lithium nuclei. Further examples of emission lines in e.s.r. spectra have been observed8' which can give information in particular about the situation when a radical is actually being formed or possibly destroyed.A report of a reaction which has a measurable dependence on magnetic field82 strength might well be related to this for though magnetic energies are much too small under usual conditions to have any visible effect on radical reactions when the singlet-triplet transition probability depends on the field the latter could have pronounced effects on that step of a reaction. If this result is confirmed it could have important theoreti- cal and practical implications. Finally it should be noted that the following texts have appeared a simple introduction to free radical chemistry83 and a comprehensive monograph.84 K. S. Chen F. Bertini and J. K. Kochi J. Amer. Chem.Sor. 1973,95 1340. P. W. Atkins A. J. Dobbs and K. A. McLauchlan Chem. Phys. Letters 1973,23,204. R. Z. Sagdeev K. V. Salikhov T. V. Leshina M. A. Kamkha S. M. Skein and Yu. N. Molin Pis'zma Zhur. eksp. teor. Fiz. 1972 16 599. 83 J. I. G. Cadogan 'Principles of Free Radical Chemistry,' Monographs for Teachers The Chemical Society. London. 1973 No. 24. 84 J. K. Kochi 'Free Radicals Vol. I Dynamics of Elementary Processes' Wiley New York. 1973.