4 Reaction Mechanisms Part (iii) Free Radical Reactions By A. G. DAVIES Department of Chemistry University College 1ondon 20 Gordon Street London WClH OAJ 1 General For organic chemists who are concerned with e.s.r. spectroscopy the latest volumes of Landolt-Bornstein are invaluable. In 1965 Volume II/1 listed the magnetic properties of all the radicals which were then known. Data on the radicals which have been studied in the succeeding decade are now being compiled in five supplementary volumes of which the first three have already appeared. Volume 9/a lists the inorganic radicals volume 9/b the organic carbon-centered radicals and volume 9/cl the organic N-centred and NO radicals up to the end of 1975.’ Volumes 9/c2 (organic radicals with 0 P Si Ge Sn Sb As or Se as the central atom) and 9/d (radical ions polyradicals and spin-labelled biomolecules) should be published during the coming year.Other monographs include Symons’ ‘Chemical and Biological Aspects of E.s.r. Spectroscopy’,* Box’s ‘Radiation Effects E.s.r. and Endor Analysi~’,~ and Davies and Parrott’s ‘Free Radicals in Organic Synthe~is’.~ The papers delivered at the symposia on organic free radicals in Aix-en-Provence’ and in Chicago6 in 1977 have been published in bookform and an issue of Journal of Physical Chemistry contains a state-of-the-art collection of papers on radical ions7 2 Kinetics Most values for the rate constants of elementary homolytic processess have been measured by comparison against a small number of absolute standards which must be determined as accurately as possible.One such standard is the rate constant for abstraction of hydrogen by t-butoxyl radicals and this has now been measured more accurately by the use of laser flash Landolt-Bornstein’s ‘Numerical Data and Functional Relationships in Science and Technology’ New Series II/1 (1965) II/9a (1977) I1/9b (1977) I1/9c 1 (1978) ‘Magnetic Properties of Free Radicals’ Springer Berlin. M. C. R. Symons ‘Chemical and Biological Aspects of E.s.r. Spectroscopy’. Van Nostrand Reinhold New York 1978. H. C. Box ‘Radiation Effects E.s.r. and Endor Analysis’. Academic Press New York,1978. D. I. Davies and M. J. Parrott ‘Free Radicals in Organic Synthesis’. Springer-Verlag Berlin 1978. ’ ‘Radicaux Libres Organiques’ ed.J.-M. Suzur C.N.R.S. Paris 1978. ‘‘Organic Free Radicals’ ed. W. A. Pryor A.C.S. Symposium Series Washington 1958 Vol. 69. ’J. Phys. Chem. 1978,82 [lo]. 70 Reaction Mechanisms 71 photolysis.* Di-t-butyl peroxide is photolysed by a laser pulse of a few nanoseconds to generate t-butoxyl radicals in the presence of diphenylmethanol or cumene (RH). The build-up of the concentration of the diphenylhydroxymethyl or cumyl radicals over a few microseconds is monitored by U.V. spectroscopy giving the rate constant for the reaction B~O.+RHk B~OH+R. For other substrates R'H where the radical R'* does not absorb in the u.v. t-butoxyl radicals are generated in the presence of a mixture of R'H and cumene and the effect of R'H on the rate of formation of the cumyl radical is again followed by U.V.absorption. Typical values which have been obtained for k at room temperature with 1:2 benzene-di-t-butyl peroxide as solvent are toluene 2.3 X lo5,cyclopentane 8.8 X lo5,methanol 2.9 x lo' t-butyl hydroperoxide 2.5 x lo8,di-t-butyl peroxide ~1.6 x lo51mol-1 s-l. The relative values of these rate constants are similar to those which have been determined by other methods but the absolute values are substantially higher than those which have been used as reference standards in the past with the result that many rate constants quoted in the literature are too low. The absolute rate constants for the reactions of aryl radicals have been determined by time-resolved U.V. or e.s.r. ~pectroscopy.~ Pulse radiolysis of the p-bromoben- zoate ion ArC02- gave the Ar*radical and the rate of the formation of the adduct [ArArC02-].was monitored by optical absorption or of the decay of the radical Ar. by e.s.r. spectroscopy. It is concluded that the rate constants for the reaction of phenyl radicals with most aromatic systems are in the range 5-10 X lo6 1mol-' s-l and that the rate of hydrogen abstraction from isopropyl alcohol is 5.2X lo61 mol-' s-'; with such a high reactivity the life-time of the phenyl radical in most organic media will be very short. In an alternative approach,1° phenyl radicals formed by thermolysis of phenyl- azotriphenylmethane (1)are caused to abstract hydrogen from mineral oil (RH) in competition with a scavenging reaction with Iz or CBr4 the rate of which in the viscous medium is diffusion controlled and can be calculated from diffusion theory.R/phH Ph3CN=NPh -* Ph3C. + N2 +Ph. (1) CRr4 \tt.. PhX The absolute rate constants for other reactions of phenyl radicals can then be determined from appropriate competition reactions based on the rate constant for the reaction of mineral oil as the primary standard. For example the rate constant for the reaction Ph. + 0,+PhOZ* has been found to be 5 x lo91mol-' s-l and not abnormally low as had been suggested. R. D. Small and J. C. Scaiano J. Amer. Chem. Suc. 1978 100 296; H.Paul R.D. Small and J. C. Scaiano J. Amer. Chem. Suc. 1978,100,4520. V.Madhavan R.H. Schuler and R. W. Fessenden J. Amer. Chem. SOC.,1978,100,888.lo R. G. Kryger J. P. Lorand N. R. Stevens and N. R. Herron J. Amer. Chem. Suc. 1977,99,7589. A. G. Davies Full details of Schmid and Ingold's measurements of the rate of spin-trapping of primary alkyl radicals by 2-methyl-2-nitrosopropane,nitrosodurene 2,3,5 -tri- t- butylnitrosobenzene 1,l-di-t-butylethylene and di-t-butyl thioketone," and of Schuh and Fischer's thorough study of the kinetics of the self-reaction of t-butyl radicals in so1ution,12 which were briefly reported last year have now been pub- lished. The isomerization of the 2,4,6-tri-t-butylphenyl (2) radical had been shown to occur in solution by quantum-mechanical tunneling of a hydrogen atom from an ortho t-butyl group. These studies have now been extended down to 28 K in solid matrices.Below 40 K the rates become virtually independent of temperature and show enormous deuterium kinetic isotope effects of greater than 104.13 (2) Work on nanosecond time-resolved e.s.r. spectroscopy is going on in a number of laboratories and many applications of this technique in the study of homolytic kinetics and mechanisms can be expected in the near future. 3 Carbon-centred Radicals A linear relationship has been established between the free activation energy of the thermolysis of a large number of aliphatic hydrocarbons (3) and their strain energies ESl4as estimated by Engler-Schleyer force field" calculations R' IR2-C-C-R2 IR3 R' I k3 -P R' I2R2-C. IR3 (3) (4) AGS(300"C)= -2.51 (*0.13)Es+274.5 (k4.2) kJ mol-' The ground state strain energy is thus crucial in determining the thermal stability about 40% of E prevailing in the transition state; conversely the activation energy for the combination of the radicals (4) corresponds to 40% of the ground state energy of the dimer of the radical.1,1,2,2-Tetra-t-butylethane has a strain energy of 264 kJ mol-' and an activation enthalpy for decomposition of 152 kJ mol-' and as such is the most thermally labile alkane which is known.16 P. Schmid and K. U. Ingold J. Amer. Chem. Soc. 1978,100,2493; S. Y. Maeda P. Schmid D. Griller and K. U. Ingold J.C.S. Chem. Comm. 1978,525. H. H. Schuh and H. Fischer Helu. Chim. Acfu 1978 61 2130. l3 G. Brunton D. Griller L. R. C. Barclay and K. U. Ingold J. Amer. Chem. SOC.,1976 98 6803; G.Brunton J. A. Gray D. Griller L. R. C. Barclay and K. U. Ingold ibid. 1978 100 4197. l4 C. Ruchardt H.-D. Beckhaus G. Hollmann S. Weiner and R. Winiker Angew. Chem. Internat. Edn. 1978 17 593. Is E. M. Engler J. D. Andose and P. von R. Schleyer J. Amer. Chem. Soc. 1973 95,8005. l6 H.-D. Beckhaus G. Hellmann and C. Riichardt Chem. Ber. 1978 111,72. Reaction Mechanisms The heats of formation of the hydrocarbons R-But and R-Me as calculated using the Engler-Schleyer force field have been used to define a frontal strain parameter .Yf for the group R by the expression Yf(R)= AH,"(R-But)-AH (R-Me) + 8.87 ( lo4J mol-') where the constant normalizes .Yf(CH3)to zero. Values for Yf have been listed for many acyclic and cyclic alkyl groups.17 The Yf parameter is useful for analysing the effect of frontal strain on reactivity.For example if the experimental activation enthalpies AH&-AH; for a variety of radicals (5) are plotted against the Yfvalues two straight lines are obtained one for u-radicals and a second one of steeper slope for w-radicals which exert a greater repulsion toward XCC13.'8 t-Alkyl radicals lie on the line correlating n-radicals indicating that they must already have or can readily achieve a near-planar structure Although it is generally accepted that the methyl radical is planar the possibility that t-alkyl radicals may be non-planar continues to attract attention. A sensitive test is the value of a [l3Ca) which should increase monotonically with temperature as the amplitude of vibration increases if the configuration at the minimum of the potential energy curve is planar.A careful study of the e.s.r. spectrum of the Me313C. radical in propane or iso-octane over the temperature range 120-380 K shows that a('3Ca)has a well- defined minimum at 220 K suggesting that there is a small energy barrier for the inversion of the radical. This barrier was calculated to be 1.88 kJ mol-' with the most stable structure distorted by 11.5" from planar." The self-reaction of triphenylmethyl radicals is well established to occur by addition of the central carbon atom of one radical to a carbon atom of a phenyl ring in the second radical. In an elegant application of CIDNP n.m.r. spectroscopy it has now been shown that benzyl radicals give bibenzyl by direct aa coupling and indirectly uia unstable semibenzenes (6) by CYO and arp coupling which can be trapped by acids as u-and p-benzyl-toluenes.*' At 30 "C the total yield of semibenzenes is 19%.The product distribution and temperature dependence point to energetically different bimolar complexes of benzyl radicals acting as product-controlling intermediates.H.-D. Beckhaus Angew. Chem. Internat. Edn. 1978,17,593. ** B. Giese and H.-D. Beckhaus Angew. Chem. Internat Edn. 1978,17,594. D. Griller K. U. Ingold P. J. Krusic and H. Fischer 1.Arner. Chem. SOC.,1978,100,6750. *' H. Langhals and H. Fischer Chem. Ber. 1978 111,543. A. G. Davies 4 Nitrogen-centred Radicals An interesting development in nitrogen radical chemistry is the proposal that succinimidyl (NS*) and related radicals can exist in distinct T (7) and u (8) states which are not readily interconvertible.2' The basis of the suggestion is that suc- cinimidyl radicals derived from different reactions show different selectivities in their reactions with compounds such as neopentane and dichloromethane.If the NS-radical is generated in the presence of bromine by the reaction NBS+Br*-+Br2+NS- it reacts unselectively with neopentane and dichloromethane whereas if an alkene is present to scavenge any bromine so that the NS. radical is generated by the reaction NBS + R*+RBr +NS- the selectivity is about 20 1.22 The results of earlier INDO calculations that the NS. radical from the former route is a ground state wradical(7) but that from the latter is an excited u (probably uN)radical (8) which in its reactions shows a selectivity similar to C1-.ao 043-0 o N N (7) (8) The reactions with hydrogen donors are therefore as follows Br.+NBS -* Br2+.rr-NS* v-NS*+RH -+ NHS+R* R*+Br2 + RBr+Br. and R.+NBS + RBr+c-NS. u-NS*+RH + NHS+R* The u-NS-radical will also add to alkenes and arenesZ4 and is also responsible for the formation of 3-bromopropionyl isocyanate (9) by reversible ring-opening. The 21 P. S. Skell and J. C. Day ref. 6 p. 290;Accounts Chem. Res. 1978 11 381. 22 P.S.Skell and J. C. Day J. Amer. Chem. Soc. 1978,100 1951. 23 T.Koenig and R. A. Wielesek Tetrahedron Letters 1975 2007. 24 J. C.Day M. G. Katsaros W. D. Kocher A. E. Scott and P.S. Skell J. Amer. Chem. SOC.,1978,100 1950. Reaction Mechanisms 75 w-NS- radical does not correlate with the ground state of the ring-opened radical but the u-NS-radical does and the P-scission is rapid and reversible. The reaction of ci~-2,3-[~H~lNBS with neopentane in the presence of bromine (conditions for the formation of the w-NS- radical) gave pure cis -[2H2]succinimide but C~S-[~H~]NCS in the presence of an alkene (now to give the a-NS- radical) gave r2H2]NHS which was 70% cis and 30% trans.25 Attempts to observe the e.s.r. spectra of succinimidyl radicals have as yet been unsuccessful but the hyperfine coupling constants of some other amidyl radicals appear best interpreted in terms of a wconfiguration.26 Other u-amidyl radicals have also been suggested to be implicated in some intramolecular hydrogen abstrac- tions reaction with aromatic rings and additions to alkene~.~~ An interesting comparison can be drawn between the properties of the radicals '+ R2N* R2NH and (Me3Si)2N*.Simple dialkylaminyl radicals R2N-,are rather inert towards alkanes alkenes and arenes whereas the dialkylaminium analogues *+ R2NH are much more reactive; thus the photolysis of tetramethyltetrazene Me2NN=NNMe2 in acetonitrile containing an alkene in the presence of .+ trifluoroacetic acid yields a mixture of products resulting from attack of the Me2NH radical on the double bonds.28 The bis(trimethylsily1)aminylradical (Me3Si)2N* (10) has now been generated by the photolysis of tetrakis(trimethylsilyl)hydrazine (Me3Si)2N-N(SiMe3)2 tris(trimethylsilyl)hydroxylamine (Me3Si),NOSiMe3 or bis(bistrimethylsily1-aminyl)mercury [(Me3Si)2N]2Hg.29 E.s.r.spectroscopy showed that its reactivity is similar to that of an alkoxyl radical. Reaction with isobutane at 170 K gave the isobutyl(l1) and t-butyl(l2) radicals in the ratio of 4.2 :1 whereas t-butoxyl radicals under the same conditions give the ratio 0.02 :1;this probably reflects the greater steric demands of the disilylaminyl radicals. (Me3Si)2NN(SiMe3)2 2(Me3Si)2N* Me2CHCH2+Me3C. (12) " P. S. Skell J. C. Day and J. P. Slanga Angew. Chem. Internat. Edn. 1978 17 515. W. C. Danen and R. W. Gellert J. Amer. Chem. SOC.,1972,94,6853;W.C. Danen and F. A. Neugebauer Angew. Chem. Intemat. Edn. 1975,14 783.27 T. C. Joseph J. N. S. Tan M. Kitadani and Y. L. Chow Canad.J. Chem. 1976,54,3517; S. A.Glover and A. Goosen. J.C.S. Perkin I 1977 1348; P.Mackiewicz R.Furstoss B. Waegell R. Cote and J. Lessard I. Org. Chem. 1978.43 3746 3750. '* L. J. Madgzinski K. S. Pillay H. Richard and Y. L. Chow Canad. J. Chem. 1978,56 1657. 29 B.P.Roberts and J. N. Winter J.C.S. Chem. Comm. 1978 545. 76 A. G. Davies As this would imply bis(trimethylsily1)bromamine will brominate hydrocarbons by a radical chain mechanism AIBN (Me3Si)2N.+RH d(Me3Si)2NH+R. R. +(Me3Si)2NBr __* RBr .t(Me3Si)zN. In the presence of norbornene or t-butylethene to scavenge bromine toluene gives benzyl bromide in 90% yield and ring-substituted toluenes show a Hammett p value of -0.62 (against cr+;cf.Bu'O* -0.35 Me2N-1.08 Br-1.36).30 The interaction of the occupied N 2p orbital with the vacant Si 3d orbital increases the electrophilic power of the radical and accomplishes intramolecularly what protonation does intermolecularly the effect is similar to that which is observed in the series Me3CO* Me3COH and Me3SiO*.31 5 Oxygen-centred Radicals The e.s.r. spectra of alkoxyl radicals cannot be observed in solution because the px and p, orbitals are degenerate allowing the unpaired electron to have orbital angular momentation about the z-axis with consequent broadening of the In the solid state strong hydrogen bonding can remove this degeneracy and quench the orbital angular momentum and under these conditions the e.s.r.spectra of a number of alkoxyl radicals have been observed by the irradiation of biologically important molecules. The first example to be reported was the radical (13),derived from the radiolysis of serine at 4.2K,33to be followed by a number of primary or secondary alkoxyl radicals at temperatures up to 165 K derived from the ribose or deoxyribose moiety of nucleosides [(14) and (15); R =H or OH] or from inositol (16)34and the spectrum of the methoxyl radical itself a(3H) 52 G has recently been observed at 4.2 K in X-irradiated polycrystalline methan01.~' OH OH R 00 OH ' OH (13) (14) (15) (16) Part of the evidence for the identification of the alkoxyl radicals is that the direction of g, in the radical is parallel to that of the C-0 bond in the parent which 30 B.P. Roberts and C. Wilson J.C.S. Chem. Comm. 1978,752. 31 P. G. Cookson A. G. Davies B. P. Roberts and M. W. Tse J.C.S. Chem. Comm. 1976 937; P. G. Cookson A. G. Davies N. A. Fazal and B. P. Roberts,J. Amer. Chem. SOC.,1976,98,616;A.G.Davies ref. 5 p. 399. 32 M. C. R. Symons J. Amer. Chem. SOC.,1969,91,5924;J.C.S. Perkin 11 1974 1618. 33 J. Y.Lee and H. C. Box J. Chem. Phys.. 1973,59,2509. 34 H.C.Box and E. E. Budzinski J. Chem. Phys. 1975,62,197; J.C.S.PerkinII 1976,553;J. Chem. Phys. 1977,67,4726; W.A. Bernhard D. M. Close J. Hiittermann and H. Zehner J. Chern. Phys. 1977,67 1211; E.Sagstuen and C. Alexander J. Chem. Phys. 1978,68,762; H. C.Box E. E. Budzinski and G. Potienko J. Chem. Phys. 1978,69 1966. 35 M. Iwasaki and K.Toriyama J. Amer. Chem. Soc. 1978,100,1964. Reaction Mechanisms forms the strongest hydrogen bond which is the site from which a proton will be most easily transferred. When they decay the alkoxyl radicals R2CHO* are converted principally into the hydroxyalkyl radical R2C0H. The g-values are highly anisotropic with g, ca. 2.028 and the magnitude of the isotropic coupling to the P-hydrogen atoms is given by the expression a(H@)= Bo+B2cos28 where Bo=ca. 0 and B2= ca. 100 G and 8 is the dihedral angle between the PC-H bond and the principal axis of the orbital containing the unpaired electron. This value of p2 is about double that which is observed for carbon-centred ?r-radicals. Against this picture of hydrogen bonding it has recently been claimed that irradiation of deoxycytidine 5'-phosphate showed the spectrum of the 5'-alkoxyl radical by cleavage of the 0-P bond.36 A radical formed at this site could hardly be hydrogen bonded; the g-value was similar to that of the alkoxyl radicals previously reported but the value of a(2HP) at 8.0 G was much lower.6 Metal-centred Radicals The trialkyltin radicals are usually obtained by the photolysis of a hexa-alkylditin or in chain reactions from a trialkyltin hydride. Some new sources have been developed which render these useful radicals more readily available. The simple hexa-alkylditins are thermally stable up to about 200 "C when they decompose irreversibly into hydrocarbons and metallic tin. Very bulky groups however can weaken the tin-tin bond and the first examples of hexa-arylditins have been prepared which undergo reversible thermal dissociation.Hexakis( 2,3,5 -trimethylphenyl)ditin and hexakis( 2,3,5 -triethylphen y1)ditin were prepared by the reaction between the corresponding triaryltin hydrides and azoiso- butyronitrile at 100 "C. The former compound at 180 "C and the latter compound at 100"C show the e.s.r. spectra of the appropriate triaryltin radicals (17; Ar =2,4,6-Me&& or 2,4,6-Et3C6H2) and measurements of the intensities gave the dis- sociation energy of the Sn-Sn bond as 190* 8 and 125 f5 kJ mol-' respectively to be compared with the value of 210-240 kJ mol-' for he~amethylditin.~' Ar3Sn-SnAr3 S 2Ar3Sn. (17) A second new source of R3Sn- radicals depends on the high reactivity of a CH bond @ to tin coupled with the ease of fragmentation of P-stannylalkyl radicals.Bu'O. +HCMe2CH2SnMe3+ Bu'OH +Me2CCH2SnMe3+ Me2C=CH2+.SnMe3 (18) Photolysisof di-t-butyl peroxide in the presence of trimethylisobutyltin (18)shows the e.s.r. spectrum of the trimethyltin radical which demonstrates its usual reactivity towards alkenes and alkyl halides. The reagent is more readily prepared and stored than is hexamethylditin and it is more stable towards oxidizing agents. Trimethyltin isopropoxide HCMe20SnMe3 can be used in the same way.38 36 D. Krilov A. Velenik and J. N. Herak J. Chem. Phys. 1978,69,2420. 37 H.U.Buschhaus and W. P. Neumann Angew. Chem. Internat. Edn. 1978,17,59. 38 A.G. Davies B. P. Roberts and M.-W. Tse J.C.S.Perkin 11 1978 145. A. G. Davies The cyclopentadienyltin compounds CpSnX3 promise to be a useful source of a variety of tin-centred radicals. Among the organotin compounds they are unique in readily undergoing photolysis and for example the Bu3Sn* radical from CpSnBu3 shows its usual reactivity toward reagents such as alkenes and alkyl halides. Tetra- cyclopentadienyltin readily reacts with a variety of protic reagents HX making available a series of new *SnX3 radical^.^' 3HX hw CpNa+SnC14 -B Cp& -CpSnX3 +Cp.+4nX3 39 A. G.Davies and M.-W. Tse J.C.S. Chem. Comm. 1978 353.