4 Reaction Mechanisms Part (iii) Electron Spin Resonance Spectroscopy and Free-radical Reactions By A. T. BULLOCK Department of Chemistry University of Aberdeen Old Aberdeen Scotland A69 2UE This year three specific areas have been selected for review. These are kinetic studies chemically induced electron polarization and equilibria involving free radicals and radical ions. 1 Kinetics and Mechanism In their continued studies of kinetic applications of e.s.r. Ingold and his co-workers have introduced the useful term 'persistent' to distinguish radicals which whilst not being stable in the thermodynamic sense are usually for steric reasons long-lived in solution. For example each radical in the formation of the alkyl radical Me,Si(Me,C)CCH(SiMe,) (1) has been observed2 (Scheme 1).At Me,Si. + Me,CC-CH + Me,CC=CHSiMe li Me,Si( Me C)CHCHSi Me +-Me,CCH=CHSiMe + Me,Si-4-H. Me,Si(Me,C)C=CHSiMe % Me,Si(Me,C)CCH(SiMe,) (1) Reagents i Me,SiH ;ii Me,Si. radical. Scheme 1 50 "C ti the half-life of (1)at an initial concentration of 3 X moll-' was 23 h. Further cases of radicals which are 'destabilized' yet persistent are certain vinyl and ally1 radicak2 One example must suffice. Radical (2) has been produced by the D. Griller and K. U. Ingold J. Amer. Chem. SOC.,1975,97,1813. * D. Griller J.'W. Cooper and K. U. Ingold J. Amer. Chem. SOC.,1975,97,4269. D. Griller K. Dimroth T. M. Fyles and K. U. Ingold J. Amer. Chem. Soc. 1975,97,5526. D. Griller L. R. C. Barclay and K. U. Ingold J.Amer. Chem. Suc. 1975,97,6151. 5 R. A. Kaba L. Lunazzi D. Lindsay and K. U. Ingold J. Amer. Chem. Soc.,1975,97,6762. 88 Part (iii) Electron Spin Resonance Spectroscopy and Free -radical Reactions addition of the trimethylsilyl radical to the appropriate di-t-butylvinylidenecyclopropaneaccording to Scheme 2. a(l3C,)=4.28 mT," which implies that the spin density on C is close to unity and that therefore the symmetry axis of the 2pz orbital on C has been twisted ca. 90" out of conjugation with the double bond. Clearly this represents maximum destabilization of (2) and yet the radical is very persistent. A comparative study4 has been made of the kinetic behaviour of the structurally related 2,4,6-tri-t-butyl-benzyl,-aniline -phenoxy and -phenylthiyl radicals (3a-d).Radical (3a) was readily prepared by photolysis of ArCH in di-t-butyl Bu' But (3) a X = -CH2; b X = -NH; c x = -0; d:X= -S peroxide. Cutting off the light resulted in a second-order decay with a rate coefficient of (5 f2) x 10' 1mol-' s-' at room temperature. Product analyses indicated the formation of the head-to-head dimer thus kl 2ArCH2 + ArCH2CH2Ar (1) (34 In contrast the anilino radical ArNH (3b) was found to be rather stable. Indeed the concentration of (3b) increased and decreased reversibly on raising and lowering the temperature ArNHNHAr 2ArNH k-z In the temperature range -40 to -65 "C AH = -54.8 f2.1 kJ mol-' and AS = -113f8 J moF1 K-'. Temperature-jump measurements enabled a determination of k-2 to be made and hence from the equilibrium constant K( = k2/L2),k was found.Expressed in Arrhenius form log(k-2/lmol-' ~-~)=6.3(=kl.0)-10.5(f3.3)/i9 (34 and log (k,/s-') = 12.2-65.3/e (3b) where 8 =2.303RT kJ mol-l. At room temperature k- =3 X lo41mol-' s-' which is some four orders of magnitude slower than the dimerization rate of the benzyl radical (3a). The arylthiyl radical (3d) showed complex kinetic behaviour but equilibrium constants for the dissociation of the dimer were 1.7x lo-'* moll-' and 1.3 X mol 1-1 at 18"C and 28 "C respectively (AH= -97.5 *8.4 kJ mol-'). In * All hyperfine coupling constants (hfcc's) are given in mT (1 mT= 10G) A. T.Bullock contrast the phenoxy radicals showed no tendency to dimerize or decay between room temperature and -100"C.There have been other reports of kinetic studies using similar techniques to those above namely monitoring the decay rate of photolytically generated radicals after the light was cut As an example we take the virtually thermoneutral hydrogen-atom transfer6 shown in Scheme 3. Under steady-state conditions a broad OOH k ROO* + d 3,ROOH + \ (4) a R = But \ b R = C2H4C(CH,)2 (5) (4) Scheme 3 singlet was observed with g = 2.015 (ie. ROO-). The concentration of this species was found to be proportional to the intensity of the light and hence to the rate of radical production but inversely proportional to [(5)]. This indicated that the alkylperoxy radical detected was (4a b) and that (6) did not contribute.The following rate law was found by following the rate of decay of the signal after shuttering the light -d[Bu'OO.]/dt = 2k3[ButO0.][~-CloH1 100HJ (4) (4b) Also followed equation (4) with values of k identical to those found for (4a). The results fitted the equation log (k3/l rno1-l s-l) = (6.0*0.5)-(18.8*2.0)/e (5) where again 8=2.303RTkJmol-'. A very large isotope effect was noted with a-C,,H,,OOD; at 21 "C k3H/k,D = 9. Further studies using 9,10-dihydro-9- hydroperoxyanthracene s-butyl hydroperoxide and triphenylmethyl hydroperox- ide showed identical kinetic behaviour. Furthermore from the absolute values of k it seems that the nature of the alkyl moiety associated with the hydroperoxide has little influence on the rate of the transfer reaction.An example of a rapid reaction which required the use of rotating-sector and signal-averaging techniques is the second-order decay of SF generated photolytic- ally by Scheme 4(a) and (b).' A plot of l/[SF,] versus time was linear; hence the rate equation is -d[SFS]/dt = 2k6[SFJ2 (6) (4 SFSCl SF + C1. (b) SF,OOSF % 2SF,O. SFSO-+ PF --* [SF,OPF,] -+ SF + OPF Scheme 4 J. H. B. Chenier and J. A. Howard Canad. J. Chem. 1975,53,623. ' J. C. Tait and J. A. Howard Chad. J. Gem. 1975,53,2361. Part (iii) Electron Spin Resonance Spectroscopy and Free-radical Reactions 91 In cyclopropane and CF2C12 values of k were identical and fitted the equation log (2k6/l mol-' s-')= (10.3*0.6)-(7.1 f2.1)/8 (7) with 8 defined as before. This is close to the diffusion-controlled limit.It would have been interesting if the authors had quoted values of the activation energy for viscous flow of the solvents. In the same study the adduct But2CCH2SF5 (7) was observed when SF5C1 was photolysed in the presence of 1,l-di-t-butylethylene [a(2H) = 0.975 a(4F) = 1.412 a(,,S) =9.171. The steady-state concentration of (7) had a light-intensity exponent of unity and decayed with first-order kinetics -4(7)i/dt = k8[(7)1 (8) with log (k*/~-')=(13k0.4)-(41.8*0.8/8) (9) This is strong evidence that the decay occurs by the unimolecular p-scission process (Scheme 5a) rather than the disproportionation shown in Scheme 5b. The authors BuiC=CH + SF Bu\CHCH2SF + Bu\C=CHSF Scheme 5 point out that the value of the pre-exponential term is consistent with a unimolecular scission process and that the low value of the activation energy is typical of the reversible addition of sulphur-centred radicals to olefins.There have been other examples of e.s.r. studies of p-scission in simple radicals. A series of ten a-alkoxyalkyl radicals have been produced by hydrogen-atom abstrac- tion from the ether precursors using Bu'O. radical.8 The p-scission process may be represented as R10CR2R3 2 R1 + 0=CR2R3 Certain trends were noted. In the series Bu'OCH, Bu'OCHCH, and Bu'OC(CH,), the first did not fragment at 0 or 30"C the second did not fragment at 0°C but &scission occurred at 30°C and a mixture of Bu'OCHCH and (CHJ3C was observed. Finally the radical Bu'0C(CH3) was found at neither 0 nor 30 "C the sole observable species being the scission fragment (CH3),C.Cumyl and benzyl ethers were also studied. 8-Scission has also been found in the fragmentation of several cyclic and acyclic dialkoxyalkyl radicals.' Measurements of the steady-state concentrations of the dialkoxyalkyl radicals [(R'O),CR2k showed a square-root dependence on the light intensity (I)at low temperatures (-60 to -136 "C) indicating that de!ay occurs by a second-order process. At higher temperatures however [(R'O),CR2] aI showing the dominance of the first-order p -scission process in removing the dialkoxyalkyl radicals. Simultaneous measurements of S. Steenken H.-P. Schuchrnann and C. von Sonntag J. Phys. Chem. 1975,79,763. M. J. Perkins and B.P. Roberts J.C.S. Perkin 11 1975 77. A. T.Bullock [(R'O),CR*] and [R']= lead to a value for 2kll/k10 where k, is the rate coefficient for the recombination process kll R' +R' +non-radical products (1 1) Estimates of .kll were available and hence klo the rate coefficients for the p-scission were obtained. An important finding was the large difference in rates for B-scission for the di-t-butoxymethyl radical and its cyclic analogue the 4,435 t e t rame t h yldioxolan y1radical (Scheme 6). rkH-1 klO(72"C)= 5.8 x 103s-' + O=CH-OCMe,CMe, 0-(CMe,),-0 Scheme 6 An extensive kinetic study has been carried out on a long series of aminophos- phoranyl radicals in solution." Both a-and p-scissions were found to occur. Kinetics of scission were measured using rotating-sector techniques combined with signal averaging.Before leaving the topic of @-scission and its kinetics an experiment designed to test for the presence or absence of a 'memory' in the @-scission of the tetra-t- butoxyphosphoranyl radical should be mentioned.' The problem may be briefly stated. Alkoxy radicals oxidize trialkyl phosphites (8) to trialkyl phosphates (10) according to Scheme 7. The tetra-alkoxyphosphoranylradical (9) is known to be the 4R. + OP(OR*)(OR) (9) (b) (10) Scheme 7 intermediate in this reaction. If the attacking alkoxy radical R*O. is identical to the alkoxy-groups RO of the phosphite then the question arises as to whether the &scission process has any 'memory' with respect to the attacking radical.In other words are either of routes (a)or (b) favoured in Scheme 7? The question has been resolved by the in situ photolysis of (i) (CD3),C02C(CD3) with [(CH,),CO],P and (ii)(CH,),CO,C(CH,) with [(CD,),CO],P in the cavity of the e.s.r. spectrometer. If the reaction has no memory then the ratio [(CH3),C*J/3[(CD,),C*] in experiment (i) should equal 3[(CH,),C]/[(CH,),C] in experiment (ii) (there is an obvious statisti- cal reason for the factors of 3). This was found to be the case over the temperature range -70 to +10"C although the individual ratios were all substantially greater than unity indicating a secondary isotope effect. That phosphoranyl radicals have a trigonal-bipyramidal structure with two apical and two equatorial ligands has been lo R.W. Dennis and B. P.Roberts J.C.S. Perkin ZZ 1975 140. Part (iii) Electron Spin Resonance Spectroscopy and Free-radical Reactions well established both experimentally'' and theoretically. l2 This leads to three possible interpretations of the absence of memory in the &scission process the most probable of which seems to be ligand exchange by for example pseudorotation (Scheme 8). For this interpretation to be correct kexch>> k, where k is the rate OR* OR I ,,OR* OR & .p' ,I' .P I'OR 1.0, OR OR Scheme 8 coefficient for the scission process. At -60°C k for P(OBu') lies in the range 1-3 x lo2s-*,depending on ~olvent.'~ Whilst there are no measurements of kexchfor this radical there have been some measurements of temperature-dependent linewidth effects in a series of alkoxyfluorophosphoranyl radicals; such effects clearly indicate rapid exchange of apical and equatorial fluorine ligand~.'~ Analysis using modified Bloch equations gave kexch(-70 "C)= 5 X lo6s-' for Bu'OPF,.This is 4-5 orders of magnitude faster than k for P(OBu'), and it seems improbable that the ligand-exchange frequency for this latter species will be significantly less than in the fluorophosphoranyl case. The pseudorotation model thus seems to be the correct interpretation of the 'memory loss'. The spin-trapping technique has been used to determine the rate of addition of phenyl radicals to benzene15 and the rate coefficients for hydrogen abstraction by the phenyl radical from methanol ethanol and propan-2-01.'~ In the first of these the kinetic Scheme 9 was found to hold.Analysis yields an expression for the initial rate PhN=NCPh 2 Ph. + N + Ph,C. (12) (not PAT trapped) Ph-+ PhCH=N(O)CMe Ph,CHN(O-)CMe PBN Ph-SA Ph-+ PhH k (not trapped) Scheme 9 l1 A.J. Gdussi J. R. Morton and K. F. Preston J. Phys. Chem. 1975,79,651. l2 J. M. F.van Dijk J. F. M. Penning and H.M. Buck J. Amer. Chem.Soc.,1975,97,4836. l3 G. B. Watts D. Griller and K. U. Ingold J. Amer. Chem.Soc. 1972,94,8784. l4 I. H.Elson M. J. Parrott and B. P. Roberts,J.C.S. Chem.Comm. 1975,586. l5 E.G. Janzen and C. A. Evans J. Amer. Chem.Soc. 1975,97,205. l6 E.G. Janzen D. E. Nutter jun. and C. A. Evans J. Phys. Chem. 1975,79,1983. A. T. Bullock of formation of phenyl adduct (Ph-SA) as a function of the nitrone concentration [PBN] uiz.It was found from the intercept that k12 (30 "C)is (1.5f0.2)x 10-5s-' in good accord with an earlier report on the kinetics of thermolysis of PAT (k12= 1 x 10" s-').I7 Further klJk13 = 0.0065. From an independent determination of k13(1.8f0.1 X lo71rno1-l s-')* the rate coefficient (30 "C)for the phenylation of benzene k14 was found to be 1.2 x lo51 mol-' s-'. Similar techniques were used to estimate the rate coefficients for hydrogen-atom abstraction by the phenyl radical from methanol ethanol and propan-2-ol16 accord- ing to the general reaction Ph. + R1R2CHOH 2R'R'COH + PhH (16) For methanol ethanol and propan-2-01 kI6was found to have the values 1.4X lo5 (2.3*0.1) X lo5 and (4.1kO.1) X lo51mol-' s-' respectively.The relative reac- tivities per a-hydrogen atom are thus in the ratio 1.0 2.4 8.8. These are perhaps fortuitously very close to the ratios of reactivities shown by these alcohols to the methyl radical in the gas phase namely 1.0 2.6 9.2 per a-hydrogen atom.17 There has been a long-felt need for the application of e.s.r. to the study of the kinetics of free-radical polymerizations. Apart from early flow-system work using redox initiators," it is only in the past three year^'^*^* that systems have been described which more closely resemble those used in conventional kinetic studies. The most recent study was of the photoinitiated radical polymerization of meth-acrylonitrile in toluene (Scheme Several assumptions of varying plausibility (CH,),C(CN)N=NC(CN)(CH,) !% 2(CH3),CCN + N (R-1 R.+ CH2=C(CN)CH3 RCH,-C(CN)CH (M) (P.1 P. + CH,=C(CN)CH PCH,-C(CN)CH (P.) 2P. 5 non-radical polymer Scheme 10 S. W. Benson 'The Foundations of Chemical Kinetics' McGraw-Hill New York,1960 p. 297. 18 H. Fischer and G. Giacometti J. Polymer Sci.,Part C,Polymer Symposia 1967,16 2763. l9 P. Smith and R.D. Stevens J. Phys. Chem. 1972,76,3141. 2o P. Smith L. B. Gilman and R. A. DeLorenzo J. Magn. Resonance 1973,10 179. *l P. Smith R. D. Stevens and L. B. Gilman J. Phys. Chem. 1975,79,2688. * This is a revised value from ref. 15. Part (iii) Electron Spin Resonance Spectroscopy and Free -radical Reactions 95 were made in the kinetic analysis.The first two are that ki= k and that k is independent of the chain length. Less reliable is the assumption that all radical recombinations (i.e. R-+P. R.+R.,and P. +Pa) have the same rate coefficient k,. A stationary-state treatment of the kinetics leads to ([P.I/rR*l)(rP.l+ [R.I)/[Ml= k,/2k (17) Measurements of [P-J and [ReJ for a range of values of monomer concentration [MI yielded a value of ca. 1x for the ratio k,/2 k,. The agreement with the literature value of 1.2 x is not impressive. The probable reason for the discrepancy is that k will almost certainly be strongly dependent on chain length for short chains. Unfortunately to detect the propagating radicals by e.s.r. it is necessary to have a rapid rate of initiation. Since this inevitably leads to short chains there is serious doubt whether such studies can appreciably contribute to our knowledge of the kinetics of free-radical polymerizations which lead to high polymers.An interesting difference between the mechanisms of the photochemical reactions of organoboranes with ketones and imines respectively has been found.22 It seems well established that the ketone reacts in its triplet state (Scheme 1la). Since imines (a) [R:CO13 + R3B --* RiCOBRi + R? c -10°C (b) Bu:C=NH + Et,B n-pentaner BuiCNHBEt + Et. Scheme 11 and ketones are isoelectronic it was supposed and found that photolysis of 2,2,4,4-tetramethylpentan-3-imine(11) with triethylborane (12) should give rise to the substituted alkyl radical (13) and the ethyl radical (Scheme llb).The hfcc's of (13) were a(14N)=0.281 a("B) = +1.022 and aEH= 3.345 mT. If the imine reacted via its triplet state (excitation energy 314-356 kJ mol-') then the reaction should be quenched by the addition of a conjugated diene. However the addition of butadiene or trans-penta-l,3-diene caused a considerable enhancement of the signal from (13). The authors" ruled out the possibility of triplet sensitization by the dienes since their triplet states are lower than those of the imines. However "B n.m.r. spectroscopy showed the formation of a strong 1 1complex of the imine with the triethylborane. It was suggested that the sensitizing effect of the dienes was attributable to their ability to transfer singlet energy to the imine-borane complex. The decay of (13) followed second-order kinetics with a rate coefficient (-80 "C)of 3 X lo31mol-' s-'.This is much more rapid than most other 1,l-di-t-butylalkyl radicals.23 Attention should be drawn to the first of the oxaziridinyl radicals (14a b) the preparation of which is summarized in Scheme 12 (14b) showed no significant decay at room temperature over a period of 24 h whereas (14a) decayed over a 22 J. C. Scaiano and K. U. Ingold J.C.S. Chem. Comm. 1975,878. 23 G. D. Mendenhall,D. Griller,D. Lindsay T. T. Tidwell and K. U. Ingold J. Amer. Chem. Soc. 1974,% 244 1. 24 R. F. Hudson A. J. Lawson and K. A. F. Record J.C.S. Chem. Comm. 1975,322. A.T.Bullock (14) a R = Ph; b R = Bu' Reagent i PbO, CCI Scheme 12 period of 1h at 50°C to give a strong e.s.r.signal of the corresponding oximino radical. The results of the investigation of the mechanism of this rearrangement are awaited with interest. Work has continued on alkylhydrazyls with a report on the kinetics mechanism and products of decay of some mono- 1,2-di- and tri-alkylhydrazyls.' The principal results were (i) that l-alkylhydrazyls undergo rapid diff usion-controlled second- order decays (ii) 1,2-di-isopropylhydrazylalso undergoes a rapid second-order decay which however is via a P-disproportionation and (iii) that trialkylhydrazyls decay by one of two routes according to structure. These are a fast second-order P-disproportionation when there is a small alkyl group bonded to the bivalent nitrogen and a slow p-scission for more hindered cases.These points are sum-marized by the examples in Scheme 13. 2CH3NNH * products k(25"C)= (1.0 & 0.5) x lO9Imol-'s-' 2Pr'NHNPr' -+ Pr'NHNHPr' + Pr'N=NPr' k(20"C)= (9.8 f 1.3) x lo7 Imol-'s-' k = (1.2 & 0.2) x lo71 mol-s-(E < 8 kJ mol-') log(k/s-') = (6.7 _+ 1.2) -(38 & 7)/d Scheme 13 There are well-established intramolecular 1,5-hydrogen shifts from carbon atoms to oxygen radical centres2' A reinvestigation of the hydrogen abstraction by Bu'O. 25 J. W. Wilt in 'Free Radicals' ed. J. K. Kochi Vol. I Wiley New York,1973 p. 17. Part (iii) Electron Spin Resonance Spectroscopy and Free -radical Reactions from substituted 1-cyclopropylcarbinols has revealed the occurrence mechanism and kinetics of a 1,5-hydrogen shift in the reverse direction i.e.from oxygen to a carbon radical centre.26 The mechanism of the rearrangement and subsequent hydrogen-atom shift is summarized in Scheme 14. For R=Me a steady-state trans-(16) (R = H Me 4,or Ph) Scheme 14 analysis of the relative concentrations of cis-(16) and (17) leads to an activation energy for the rearrangement of (20*4) kJ mol-' and a frequency factor of lo8s-l. The conclusion that trans-(l6)does not rearrange to (17) was reached by analysing the temperature dependence of the concentrations of cis-(16) trans-( 16) and (17; R =H). The authors26 account for the occurrence of this unusual shift in terms of the facile formation of a six-membered-ring transition state together with resonance stabilization.The only case where (15) could be observed was for R =Ph. There have been several e.s.r. studies of atom-molecule reactions in the gas phase using the fast-flow-microwave discharge method. The reactions of oxygen atoms ('P) with CH,Br and CH3ClZ7 over the temperature range 500-1000 K have the same rate coefficients (*lo%) given by k(O+CH,X) =3.5 x lo1 exp(-4560/T) cm3mo1-l s-l. Perhaps surprisingly the halogen has little if any effect on the rate of the primary reaction O('P) +CH,X -+ CH2X+ OH. This is further emphasized by the fact that k(O+CH,X) is very similar (especially in activation energy) to k(O +CH,) =2.1 x 1013 exp(-4550/ T)cm3 mol-' s-l. Other gas-phase kinetic studies include the systems O(3P)+CS -P CS+SO; O(,P)+ OCS -+ CO +SO (both ref.28); 0+SO +M -+SO +M" (M usually He);29 and H +ClF -B HF+ClF,.,' Finally it has been shown that two consecutive addition 26 H. Itzel and H. Fischer Tetrahedron Letters 1975,563. 27 A. A. Westenberg and N. deHaas J. Chem. Phys. 1975,62,4477. 28 C.-N. Wei and R. B. Timmons J. Chem. Phys. 1975,62 3240. 29 A. A. Westenberg and N. deHaas J. Chem. Phys. 1975,63,5411. 30 S. J. Pak R. H. Krech D. L. McFadden and D. 1. MacLean J. Chem. Phys. 1975,62,3419. 98 A. T.Bullock steps occur in the reaction of atomic fluorine with phosphorus trifl~oride;~' F. +PF3 + PF4-k(300 K) =(8.6* 0.6)X 10" cm3mol-' s-' F.+PF,. * PF5 k(300K)=(1.2*00.2)x10'3~3mol-'s-1 2 Chemically Induced Dynamic Electron Polarization Both time-resolved and steady-state measurements of chemically induced dynamic electron polarization (CIDEP) have been reported using pulsed laser^,^'-^^ pulsed radioly~is,~~ and conventional U.V.source^^^,^^ as means of generating radicals. The reader is referred to a recent review for details of the various theories proposed to account for the polarization phenomenon38 but the two principal mechanisms may be briefly described as follows. In a homolytic scission a level-crossing process is envisaged during the relative diffusion of the two radicals and gives rise to spin polarization. This is the radical-pair mechanism (RPM). The other mechanism is the triplet mechanism (TM),in which an intersystem-crossing step gives rise to polariza- tion provided that radical formation takes place in a time comparable to or less than the triplet spin-lattice relaxation time.Photolysis of solutions of pi~alophenone~~ in solvents of differing viscosities using 20 ns pulses from a nitrogen laser (A =337 nm) generates the benzoyl and t-butyl radicals both showing spin polarization. The evolutions of the polarizations with time were generally complex. Naphthalene quenching studies showed that the photolysis occurs through a triplet state which could be responsible for the weak emission in the spectrum of benzoyl. However the polarization of t-butyl showed a marked hyperfine dependence indicative of the RPM. Briefly the results were qualitatively interpreted in terms of a process involving two rates of separation of the initial radical pair. The first of these rates is separation from a triplet pair and the second is the slower departure from a singlet pair which arises from the initial triplet pair by the perturbations from the large t-butyl hyperfine interactions.It was pointed out that this interpretation was in accord with the theoretical model devised by Pedersen and Freed3' in which the rates of separation of the radicals are dependent on the exchange interaction. It is expected that the singlet pair separation would be retarded by this interaction. Polarized transient radicals from aqueous methanol and aqueous acetate solutions have been produced by pulse radiolysis (100ns electron pulses) and observed in the submicrosecond domain typically 300 ns after the The CH,CO radical has been studied in some detail.When produced by the reaction of -OH with CH3C0; the low-field emission high-field-enhanced absorption spectrum charac- teristic of the radical-pair model was observed. However the polarizations though 31 I. B. Goldberg H. R. Crowe and D. Pilipovich Chem.Phys. Letters 1975 33 347. 32 P. W. Atkins A. J. Dobbs and K. A. McLauchlan J.C.S. Faraday II 1975,71 1269. 33 B. B. Adeleke K. Y. Choo and J. K. S. Wan J. Chem. Phys. 1975,62,3822. 34 A. J. Dobbs and K. A. McLauchlan Chem. Phys. Letters 1975,30,257. 35 A. D. Trifunac K. W. Johnson B. E. Clifft and R. H. Lowers Chem.Phys. Letters 1975,35,566. 36 P. B. Ayscough T. H. English G. Lambert and A. J. Elliott Chem. Phys. Letters 1975,34,557. 37 J. B. Pedersen C. E. M. Hansen,H. Parbo and L. T. MUUS J. Chem.Phys. 1975,63,2398. 38 J. K. S. Wan S. Wong and D. A. Hutchinson Accounts Chem.Res. 1974,7 58. 39 J. B. Pedersen and J. H. Freed J. Chem. Phys. 1973 59 2869. Part (iii) Electron Spin Resonance Spectroscopy and Free -radical Reactions opposite were not equal (as predicted by RPM). It was suggested that since the spin-relaxation time TI, of .OH is expected to be very short then it would react with CH,CO with its spin states in equilibrium. Some transfer of this equilibrium population to *CH,CO would result in the observed decreased emission of the low-field line and enhanced absorption of the high-field line. This interpretation was supported by the nearly equal intensities of the polarized low- and high-field lines in the same radical produced by the reaction of the hydrated electron (e;J with CH,ICO; and CH,ClCO,.T, for e& is of the order of microseconds and thus cannot relax before the production of -CH,CO; radical. In discussing e.s.r. spectra showing CIDEP it is useful to define an enhancement factor Vqsuch that Vq=[S(4)-S,(q)]/S,(q) where 4 is an index of the nuclear spin state associated with the line under consideration S(4) is the amplitude of the enhanced signal and &(q) is the amplitude of the signal when there is no enhancement. An investigation of the dependence of Vq'son radical concentration and radical structure has been made.36 The radicals were generated under steady- state concentration conditions by 1ms pulses from a mercury lamp according to Scheme 15. Radicals (18) (19) and (21) all showed a linear dependence of Vqon n n hv Bu:02 -hv CH3CHO -CH3CHOH hv H2°2 -H°CH2CHoH (21) Scheme 15 the radical concentration as predicted by the theory due to Fe~senden.~' Thistheory predicts that the enhancement factor Vq,for a given pair of lines (high- and low-field lines with the same 9) is proportional to TT,k,n where T is the temperature TIthe spin-lattice relaxation time and k the radical-encounter rate constant.Values of VJn for (20) (19) and (18)were found to be 0.68,2.2,and 7.2 1mol-' respectively. It was concluded that these differences could not be explained in terms of the values of T,k for the three systems and that Vqmust depend markedly on radical structure. In particular the authors pointed out that the largest emission- absorption polarizations are found in cyclic radicals.It is possible that an anisotropic exchange interaction its dependence on radical structure and its effect on the separation rate of the radical pair need to be considered in the theory of Pedersen and Freed39 before the effects of radical structure on CIDEP can be explained. R.W.Fessenden J. Chem. Phys. 1973,58,2489. 100 A. T.Bullock Time-resolved techniques have been used to study the polarization mechanisms in the neutral p-benzoquinone radical (PBQH.) produced by photolysis of p-benzoquinone in solution by the steps indicated in Scheme ,16.37 The curves of the PBQ+hv + 'PBQ 'PBQ -+ 'PBQ 'PBQ + RH + PBQH. + R-solvent (PBQH. PBQ; + H') Scheme 16 time evolution of the intensity were fitted to theoretical curves.41 Two contributions to the polarization were observed when ethylene glycol was the solvent.These were (a)an initial hyperfine-independent polarization due to a triplet mechanism and (b) a hyperfine-dependent polarization caused by the RPM due to the radical-termination reaction. Observed values of the TM polarization indicated a triplet lifetime in the range 0.1-1011s and a zero-field parameter IDI>70mT. The magnitude and hyperfine dependence of the RPM polarization was in accord with the theory for an exponentially decaying exchange interaction J(r) with a maximum value Jo > lo9rad s-'. The photoexcited TM has received support from two groups who used plane- polarized Both studies have as their origin the calculations of Adrian:' who showed that if the TM is operative then the spin polarization of the radicals will depend on the orientation of the electric vector E of the polarized light with respect to B the external magnetic field.We conclude this section by reporting on the more extensive of the two experimen-tal The T-T* singlet-singlet transition in 1,4-benzoquinone is known to be polarized along the 0-0axis. Adrian's calculations predict that for a CIDEP process involving this species the TM using polarized light would result in a ~W/O increase in polarization going from EllB to E IB. In complete agreement with this both the benzoquinone and the 2,6-di-t-butylphenoxy radicals showed this polariza- tion when a solution of benzoquinone and 2,6-di-t-butylphenol (DTBP) was photo- lysed with plane-polarized light from a pulsed 20 kW nitrogen laser.In contrast to benzoquinone the T-T* singlet-singlet transition in 2-methylanthraquinone is perpendicular to the 0-0 axis. For a solution of this quinone with DTBP it was found that the emission magnitude of the phenoxy radical decreased by ca. 20%in going from EllB to ElB (the signal from the quinone radical was too weak to be useful for quantitative measurements). Again this result is substantially in agree- ment with Adrian's calculations although the predicted decrease is 10%. 41 J. B. Pedersen J. Chem. Phys. 1973,59,2656. 42 F.J. Adrian J. Chem.Phys. 1974,61,4875. Part (iii) Electron Spin Resonance Spectroscopy and Free -radical Reactions 101 3 Equilibria The use of precise measurements of g for the determination of equilibrium constants in a rapid dynamic equilibrium situation has been described43 for the ion-pair equilibrium between 2,6-di-t-butyl- 1,4-benzosemiquinone and IS+in hexamethyl- phosphoramide (HMPA).The relevant equation is l/AE =K,,/Ag'[K+]+ l/Ag' CW where Ag =gokneci-gfree ion Ag' = gfre ion -gion pair and Keg =[Ryl[K+I/CRT K+l i.e. the ion-pair dissociation constant. At 28 "C Keg=0.076 f0.009 which agrees well with the value of 0.091 f0.01 obtained by measuring time-averaged hfcc'~.~~ In this latter work ion-$air dissociation constants were measured over a range of temperatures for K' with the radical anions of 2,6-di-t-butyl- 1,4- benzosemiquinone p-dinitrobenzene p-nitrobenzaldehyde and p-nitrobenzophenone.All the relevant thermodynamic parameters were obtained for these systems. The equation is precisely analogous to (18) namely 1/(A -A")=Keq/[K+](A'-A")+ l/(A' -A") (19) where A is the observed hfcc for a particular magnetic nucleus A" is for the free ion and A' for the ion pair. Equations (18) and (19) are only valid when the frequencies of the forward and backward reactions are large compared to the relevant changes in the measured spectral parameters expressed in frequency units i.e. Ag'PB,/h [equation (l8)J and (A'-A") [equation (19)J. When this situation does not pertain two superimposed spectra are obtained and Keqis found by straightforward con- centration measurements on the two components.For ion-pair equilibria of K+with the radical ions of 2,6-di-t-butyl-l,4-benzoquinone, 9,10-anthraquinone and 1,4- naphthoquinone single spectra and the various thermodynamic parameters (K,AW AS") were obtained using g measurements. However for the unsubstituted benzosemiquinone slow exchange was observed and it was suggested on the basis of INDO calculations that fast exchange occurs if the sum of the electron densities on the oxygens <6.46; otherwise the exchange is slow. Another example of a fast exchange equilibrium studied using equation (19) includes the hydrogen-bond exchange reaction shown in Scheme 17.46 The results showed a linear correlation between the Taft parameter for the para-substituent u+and In Kes. XGN02 + HMPA--HOMe NO --HOMe+ HMPA .I .-I (X = H Bu' CN NO, CO,Me Cl or COMe) Scheme 17 Slow and fast exchange has been found in certain acid-base equilibria.The P-H hfcc's of seven P-hydroxyalkyl radicals have been found to be dependent upon 43 G. R. Stevenson and A. E. Alegria J. Phys. Chem. 1975,79 1042. 44 A. E. Alep'a R. Concepci6n and G. R. Stevenson J. Phys. Chem. 1975,79 361. 4s G. R. Stevenson A. E. Alegria and A. McB. Block J. Amer. Chem. Soc. 1975,91,4859. 46 G. R. Stevenson L. Echegoyen and H. Hidalgo J. Phys. Chem. 1975,79 152. 102 A. T.Bullock PH.~’The equilibrium is >H20H +OH->H20-+H20 A B and it may be shown that PKa =pH +loglo [(a-a’)/( uA-U)J (21) where u is the observed P-H hfcc and aB,(I~ are those for pure B and pure A respectively.Where comparisons are available pKa (radical)=pKa (alcahol)-(1.3*0.2). A similar result was found for radicals of the type HSCH2CXY,where again the radical pKa’s were lower than for the parent thiols (ca. 1.5-2 In the sulphur radicals however exchange was slow and the Ka’s were determined by relative concentration measurements of the ionized and unionized radicals as a function of pH. Acid dissociation constants have also been reported in an extensive study of the eiq adduct to fumaric acid.49 These may be represented thus pK,=8.1 pKa=10.8+0.l monoanion (two forms) dianion (one form) . trianion Resuits on three disproportionation equilibria have been reported in which radical monoanions are in equilibrium with their diamagnetic dianions and hydrocarbon precursors.The equilibria may be expressed as .rr2-+ n S 2nT. The enthalpy changes involved in this type of equilibrium for bemcyclo-octatetraene and naphthocyclo-octatetraene when compared with the parent cyclo-octatetraene did not reflect the expected strong decrease in electron-electron repulsions for the dianions.” The thermodynamic parameters controlling the disproportionation of the [16]annulene anion radical depended on the counterion (Li+ Na+ K+).51 The variations were ascribed to metal association with the dianion the monoanion existing essentially as a free ion in the solvent used (HMPA). The morphamquat radical cation (22) exists in equilibrium with its diamagnetic dimer in methanol (22) ~~lution.~~ For this system AH” = -45.05 *0.3 kJ mol-’ AGO298 = -10.6 kJ mol-’ and = -115.6J mol-’ K-’.Finally we conclude with two cautionary notes. The first concerns the use of crown ethers to complex with alkali-metal counterions and hence obtain spectra of 47 Y. Kirino J. Phys. Chem. 1975,79 1296. 48 Y. Kirino and R. W. Fessenden J. Phys. Chem.,1975,79 834. 49 0.P. Chawla and R. W. Fessenden J. Phys. Chem.,1975,79 76. 50 G. R. Stevenson M. Colbn I. Ocasio J. G.Concepcih and A. McB. Block J. Phys. Chem.,1975,79 1685. 5l J. G. Concepci6n and G. Vincow J. Phys. Chem. 1975,79,2037. 52 A. G. Evans J. C. Evans and M. W. Baker J.C.S. Perkin ZZ 1975 1310. Part (iii) Electron Spin Resonance Spectroscopy and Free -radical Reactions 103 'free' radical anions.The first report of the anion of mesitylene has appeared,53 prepared by the reaction of a solution of 18-crown-6 in mesitylene with a potassium mirror. In addition to the couplings to protons in the radical ion (a,", =0.491 mT a&=0.257mT) a further coupling to six equivalent protons was found (0.018mT). The authors suggest a model in which there is an ion pair between [mesitylene]' and [K -* -crown]+ ions. This was confirmed using the anion of toluene when two sets of hfcc's to the crown occur 0.012 mT (4H) and 0.018 mT (ZH).Secondly in a study of the association of di-t-butyl nitroxide with the Schardinger dextrin cyclohepta-amylose it was found that the concentration quotient for a simple association equilibrium was not constant.54 However a simple linear relationship between In yR (where yR is the activity coefficient of the radical) and d the total dextrin concentration adequately dealt with deviations from ideality.53 G. V. Nelson and A. von Zelewsky J. Amer. Gem. Soc.,197597,4279. 54 N.M.Atherton and S. J. Strach J.C.S. Faraday I 1975,71,357.