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 A89 2UE This yekr’s Report is restricted to two topics namely kinetic studies and chemically induced electron polarization. The year saw the publication of Volume 3 of ‘Electron Spin Resonance’.’ Whilst the standard is excellent throughout of particu- lar relevance to this Report are Chapter 9 by Sealy and that part of Chapter 3 by Atkins which deals with polarization phenomena. 1 Kinetics and Mechanism There have been two independent reports on the observation and kinetic behaviour of several thionitroxide radical^,"^ (l),(2) and (3). Compared with the parent a:/rnT2 * 1.16 0.80 1.09 2.0171 2.017 2.015 g nitroxides it is seen that the thiyl radicals have lower uNvalues and larger g factors indicating more localization of spin on the sulphur atoms.’ The radicals have been prepared by photolysis or thermolysis of the parent disulphides.Using the former technique and following the radical concentration after shuttering the light it was found that (1) and (3) both decayed with second-order kinetics and closely similar rates such that log (k-’/l mol-’ s-’) = 11.8 -16.7/8 (8 = 2.303 RT kJ mol-’). The equilibrium constant k l/k-l [equation (l)]was determined by measuring radical ‘Electron Spin Resonance’ ed. R. 0.C. Norman (Specialist Periodical Reports) The Chemical Society London 1976 Vol. 3. W. C. Danen and D. D.Newkirk J. Amer. Chem. SOC.,1976,98 516. 3 B. Maillard and K.U. Ingold J. Amer. Chem. SOC.,1976 98 520. *lmT=lOG. 71 A. T.Bullock concentrations in solution at different temperatures. From this together with the kinetic data it was found that log (k1/s-') = 16.8-129.7/8 8 being defined as bef~re.~ The activation energy was in good agreement with the result obtained by Danen and Newkirk,' who measured the rate of thermal decomposition of bis(pyrrolidy1-1) disulphide (4) using Banfield's radical as a scavenger. It has been noted3 that the thionitroxides are unique in that they recombine at or near the diffusion-controlled limit but are extremely unreactive to several substrates (e.g. 02 alkenes triethylphosphite and triethylphosphine) which react readily with most free radicals.(4) Following earlier inve~tigations,~*~ an extensive study of the kinetics and products of decay of a series of N-alkoxy-N-alkylamino N-alkoxyamino and N-alkoxy-N- anilino radicals has been described.6 The N-alkoxyamino radicals decayed rapidly with second-order kinetics (k/l mol-' s-l -1-4 x 10' at -60 "C). The rate of the decay was ascribed to lack of steric protection. On the other hand the N-alkoxy-N- anilino radicals decayed very slowly and were found to be in equilibrium with their dimers over the temperature range -63 to +74 "C [equation (2)]. For R = But AH" = -52 f2 kJ mol-' and AS"= -88 * 8 J mol-' K-'. ~RONC~HS $ (RONC~HS)~ (2) The kinetics of addition of a series of radicals R,M- to di-t-butylthioketone have been studied (Scheme l).' From a comparison of aMin radicals (5) and analogous R,M.+ S=CB& -+ R,MS~BU'~ [R,M = CH3 C(CH3)3 CF3 SiH3 Si(CH3)3 S~(BU")~ Sn(CH3)3 S~(BU")~, or Et02P=O] Scheme 1 radicals formed by the addition of R,M* to di-t-butylethylene it was proposed that the adduct group occupied the eclipsed position with respect to the 2p orbital on the a-carbon (6).Support for this structure came from a comparison of the g values W. C. Danen and C. T. West J. Amer. Chem. SOC.,1971,93,5582. W. C. Danen C.T. West andT. T. Kensler J. Amer. Chem. SOC.,1973,95 5716. 6 R. A. Kaba and K. U. Ingold J. Amer. Chem. SOC.,1976,98,7375. J. C. Scaiano and K. U. Ingold J. Amer. Chem. SOC.,1976 98 4727. Reaction Mechanisms-Part (iii) Electron Spin Resonance Spectroscopy 73 (2.0024-2.0033) with those found in unhindered alkylthiyl radicals where delocali- zation onto the sulphur atom c2n take place (g -2.0044-2.0049).The kinetics of the addition of R,M* to the thione were obtained from two competition experiments. Bu'N=NBu' 42Bu'. + N2 But.+ S=CBu'2 -%Bu'SeBu'2 Bu'*+O2 -% Bu'OO. Scheme 2 The first of these involved Scheme 2. At -80°C the adduct was found to be persistent and its concentration increased with time t. Analysis yielded the equation -d[Bu'OO*] kl = k2[Bu'SeBut2],[02]/ dt X t X [Buf2C=S] (3) At -80 "C k = (1.3k0.6)X lo61 mol-' s-l. The second experiment involved competition between the addition of the methyl radical to the thioketone and to Bu'N=O.Analysis of these results gave a value for kl the rate coefficient for addition to the thioketone of 21.1X lo61mol-' s-' at -40°C. Decay kinetics of the adduct radicals (5) were usually complex and of non-integral order. However the methyl adduct was found to be in equilibrium with its dimer. 2CH3SCBut2 dimer (4) Application of the van't Hoff isochore yielded AW=40*6 kJ mol-' and AS"= 134* 13 J mol-' K-' for the dissociation of the dimer. Radical addition to di-t-butylsulphurdi-imide (7) produces the nitrogen-centred radicals (8) according to equation (5).' The adduct radicals were F3C. Me3Si- Bu"Si*,(EtO) P=0,and F3CS*.All the adduct radicals (8)decayed with first-order kinetics and at 20 "C k/s-' = 4 x <lo-' 0.25 and 0.30 respectively.R,M* + Bu'N=S=NBd + Bu'NSN(Bu')MR (5) (7) (8) The persistence of these radicals was attributed to two factors stericprotection of the a-nitrogen and the absence of a-hydrogens. The same paper8 reported addition of CF,O* to di-t-butylcarbodi-imide (9) to give the diaza-ally1 radical (10) [equation (6)]. Radical (10) was found to be appreciably more persistent than previously CF30.+BU'N=C=NBU' -+B~'N=-C:-NB~~ (6) I OCF3 (9) (10) reported diaza-allyls' and decayed with second-order kinetics (presumably to dimer). The rate coefficient was found to fit the expression log (2k,/l mol-' s-I) = (7k l)-(lOk5)/6. The low pre-exponential factor indicates a high entropy of * G. Brunton J. F. Taylor and K. U. Ingold J. Amer. Chem. SOC.,1976,98,4879. 9 W.Ahrens and A. Berndt Tetrahedron Letters 1974 3741. 74 A. T.Bullock activation. Thus although the activation energy is very low the duration of a radical-radical encounter in solution would not normally be long enough to allow the establishment of a configuration favourable for reaction. The dicyanomethyl radical Ht(CN), has been found to decay with second-order kinetics at the diffusion-controlled limit" despite extensive delocalization onto the cyano-groups.'o However in the same work it was found that the rate of decay of the tricyanomethyl radical C(CN) was some two or three orders of magnitude sbwer. No convincing explanation of this was given. Several reports of radical scissions have appeared. 12-14 It is always gratifying when e.s.r.results provide unequivocal evidence in support of a previously proposed reaction mechanism. This was found to be the case in a study of imidoyl radicals and their subsequent fragmentation.' A series of these radicals was produced by radical addition to alkyl isocyanides [equation (7)] and all had g factors in the range 'X 2.0011-2.0016. In particular the truth of an earlier pr~posal'~ was established namely that oxidation of t-butyl isocyanide to the isocyanate by t-butoxyl radicals involves p-scission of an intermediate imidoyl (Scheme 3). Spectra of the imidoyl BU'O.+BU'N=C -+ BU~N=~ /@ 'OBu' Bu'N=C=O +But. Scheme 3 and t-butyl radicals were observed simultaneously. Further when (CD3),COOC(CD3),was used as the source of radicals the imidoyl and perdeuterio- t-butyl spectra were observed thus confirming the p -scission mechanism.The rate coefficient ks was determined relative to the known value for self reaction of the t-butyl radicals and was given by loglo (kp/s-*)= 12.8 -42.2/8. Interest has been maintained in scission reactions of phosphoranyl radicals and configurational effects in the a-scission of such radicals have been reported.I3 The phosphoranyl radicals were produced by the addition of photochemically generated alkoxyl radicals to alkoxyalkylphosphines and their decay with time was measured. It had previously been proposed'6 that in the step-wise reaction given in equation (8) RO.+ROPR -+ (RO)zPRz + (RO),PR+R* (8) the a-scission was a configurationally selective process in which an apical P-C bond 10 R.A. Kaba and K. U. Ingold J. Amer. Chem. SOC.,1976,98,523. 11 A. T. Bullock G. M. Burnett and C. M. L. Kerr European Polymer J. 1971,71 1011. 12 P. M. Blum and B. P. Roberts J.C.S. Chem. Comm. 1976 535. l3 J. W. Cooper and B. P. Roberts J.C.S. Perkin II 1976 808. 14 P. G. Cookson A. G. Davies N. A. Fazal and B. P. Roberts J. Amer. Chem. SOC.,1976,98,616. 15 L. A. Singer and S. S.Kim Tetrahedron Letters 1974 861. 16 A. G. Davies R. W. Dennis and B. P. Roberts J.C.S. Perkin IZ 1974 1101. * A similar conclusion was reached some years ago" for the recombination of 2-cyano-2-propyl radicals (CH&CCN. Reaction Mechanisms-Part (iii Electron Spin Resonance Spectroscopy was cleaved more rapidly than an equatorial P-C bond.The kinetic res~lts'~ for a series of phosphoranyls having different alkyl substituents suppcrted this. It was further suggested that the increase in the rate coefficient for a-scission in the order L2P(OBu') <L P(0Bu')OEt <L2P(OEt) (L = alkyl or R,N) was due to the increase in the proportion of the isomer with an apical P-L bond as the number of ethoxy ligands increased. However if configurational effects were absent then changing the nature of the alkoxy substituents had little effect on the rate of a-scission. This paperI3 also presented the first reported measurement of a bond dissociation enthalpy for a phosphoranyl radical. The temperature variation of the equilibrium constant for reaction (9) gave AW(dissociation) = 29 kJ mol-'.The authors suggested that the low selectivity observed for the overall displacement of alkyl radicals from R'RZP by t-butoxyl radicals was due to the low selectivity of the product-controlling radical formation step. There has been a preliminary report of an e.s.r. study of the reactions of a series of trialkylsiloxyl radical^.'^ These were generated by the photolysis of some t-butyl trialkylsilyl peroxides (Bu'OOSiR3) and bis(trialkylsily1) peroxides (R3SiOOSiR3) in the temperature range -120 to -20 "C. Many of the reactions are summarized in Scheme 4. In addition to these it was noted that for the higher bis(trialkylsily1) XOOSiR3 R Me Et Pr (la) XOOSiR2R(-H) Me Et (lb) 1.5-H CH~ Pr (1c) transfer b R~S~(OH)CH~CH~ 7C2H4 R3SiOCH2CH2 Me Et (Id) (X = Me$ or R3Si) Scheme 4 peroxides (R = Et or Pr) @-scission occurred giving rise to the appropriate alkyl radicals according to reaction (10).This reaction was not found in the t-butyl R3Si0. + R*+ R2Si0 (10) trialkylsilyl peroxides. It is expected to be strongly endothermic which leads the authors to suggest the concerted process shown in Scheme 5. The 1,5-hydrogen atom transfer from carbon to oxygen [Scheme 4 (lc)] is well known in the chemistry A. T.Bullock R3Si-,O R R2Si-0 II 0'-SiR3 O-SiR2.R Scheme 5 of alkoxyl radicals1' but does not seem to have been previously demonstrated by e.s.r. spectroscopy. The trimethyl- and triphenyl-germyl peroxyl radicals (1 1) have been studied." At low temperatures (<-50 "C) they behave in a similar fashion to their carbon analogue^'^ in forming tetroxide dimers [equation (ll)].The variation of K the 2R3Ge02. $ R3Ge04GeR3 (11) (11) R=MeorPh equilibrium constant with temperature gave AW =-48 kJ mol-' and AS"= -1 11J mol-' K-' for both radicals. AH" for dimerization of the germyl peroxides was thus found to be at least 11 kJ mol-' more negative than for alkyl peroxides" whereas the entropy changes were about the same. Above -50°C the radicals decayed according to equations (12) which fits the observed rate equation (13) provided that kl >>kz. The nature of reaction (12b) was not defined. 2R. & R2 (124 k-1 k R. 2,products (12b) -d[R*] k2[R*] -= dt 1+4K[R*] In view of the postulated mediation of electron donor-acceptor complexes in a variety of chemical reactions the first kinetic study of the thermal ionic dissociation of such a complex2o should be noted.It has been shown that tetracyanoethylene in solution in dimethyl sulphoxide and in the absence of light changes quantitatively to its radical anion with first-order kinetics. Over the temperature range 25-45 "C the first-order rate coefficient was given by log (kls-') =6.3 -58.6/8. The authors claim that the slowness of the reaction supports Mulliken's suggestion*' that in polar solvents dissociation of a complex into ions is govered by a slow stabilizing solvation process. The decay kinetics of a series of monosubstituted p-benzosemiquinones (12) have been measured.22 The radical anions were generated by electrolysis in methanol using 0.1M tetraethylammonium perchlorate as the supporting electrolyte.The decay was second-order and the following mechanism was proposed K Et4NQ* +Q'+Et4N+ (144 2QS k2 Q2-+Q (14b) J. K. Kochi in 'Free Radicals' Vol. 11 ed. J. K. Kochi Wiley New York 1973 pp. 686-688. J. A. Howard and J. C. Tait Canad. J. Chem. 1976 54 2669. 19 J. E. Bennett D. M. Brown,and B. Mile Trans.Faraduy Soc. 1970 66 397. 2o N. Kushibiki and H. Yoshida J. Amer. Chem. SOC.,1976,98 268. 21 R. S. Mulliken and W. B. Person in 'Molecular Complexes A Lecture and Reprint Volume' Wiley New York 1969. 22 A. B. Sullivan and G. F. Reynolds J. Phys. Chem. 1976,80 2671. Reaction Mechanisms-Part (iii) Electron Spin Resonance Spectroscopy 0- Q" 0-(12) X =OMe Me Ph COOMe H COMe Br C1 or CN whence -d[Et,NQ.]/dt = (k2K/[Et4N'])[Et4NQ.l2 (15) A linear correlation was found between log kobsd where kobsd= (k2K/[Et4N']) and the normal Hammett substituent coefficients.The observed second-order rate coefficients varied by three orders of magnitude from kobsd/l mol-' s-' = 3.3 (-OMe) to 3300 (-CN). It was suggested that the substituent effect operated on the ion-pair dissociation constant K. Unfortunately the authors did not test the predicted [Et4N+]-l dependence of kobsd. The theories of Marcusz3 have been tested in a study of the rates of electron exchange and electron transfer involving the quinones (13)-(16).24 The rate 6 I It 6 I MI e MeQMe t 1;I ~ I Me Me Me 0 0 0 0 (13) (14) (15) (16) coefficient for the exchange process [Equation (16)] was determined by measuring the e.s.r.linewidth variation with [a]. Values of k, for all four quinones together Q+QS &% Q7+Q (16) with Kii the equilibrium constants for the transfer reactions [equation (17)] allowed k and kb to be calculated from The constants kf and kb were obtained spectrophotometrically by pulse radiolysis of mixtures of the quinones (in pairs) and in general the agreement between theory and experiment was satisfactory. Q~'+Q~& Qi+QjT (17) kb Spin-trapping experiments continue to be used and extended. Phenyl-t-butylnitrone (17)was the trapping agent in a study of the y-radiolysis of nitriles in the liquid phase." The radicals trapped and identified were *CH2CN (acetonitrile) H-and CH3CHCN (propionitrile) *CH(CN) (malononitrile) and Ha -CN and *CH2CH2CN (succinonitrile).The same trap has been used in an investigation of the photoreduction of five quinones by a series of alcohols.26 Alkoxyl radicals were 23 R. A. Marcus Ann. Rev. Phys. Chem. 1964,15,155. 24 D.Meisel and R. W. Fessenden J. Amer. Chem. SOC.,1976,98 7505. 25 S.W.Mao and L. Kevan J. Phys. Chem. 1976,80,2330. 26 K. A. McLauchlan and R. C. Sealy J.C.S. Chem. Comm.,1976,115. A. T.Bullock observed in every case whereas hydroxyl radicals were never trapped. The authors point out the presence of RO. radicals must be taken into account when interpreting electron polarization experiments. Many spin-trapping experiments in the radiolysis of aqueous solutions have involved the use of CH2=N02- the aci-anion of nitromethane.This restricts such studies to alkaline solutions. The restriction has now been removed in a comparative study of the merits of phenyl-t-butylnitrone (17) 5,S-dimethylpyrroline- 1-oxide (18) and 2-methyl-2-nitrosopropane (19).27 The last-mentioned was found to be unsatisfactory largely because of its low solubility. It seems that the best spectra were obtained from (18) and the solvated electron e& H- and *OHwere all successfully trapped during the in situ radiolysis of water with 3 MeV electrons. oCH=y-Bul 1 0 0 BU'NO (17) (18) (19) The investigation of triplet species in solution by e.s.r. is seriously hampered by the fact that for low viscosity solvents at moderate temperatures the short rotational correlation time T, together with the strong dipolar coupling produces very efficient spin relaxation.The lines are thus broadened beyond detection. Two reports of techniques which circumvent this problem have appeared. It has been shown2* that nitric oxide acts as an efficient 'triplet-trap'. Ten substrates known to produce carbenes were photolysed in the presence of NO [equation (18)]. i'.+NO -P T=NO* (18) (20) Characteristic spectra of the resultant iminoxyls (20) were observed. Direct obser- vation of the Ams=*l transition of the triplet species (21) in solution has been reported.29 This was achieved by using a solvent with a viscosity coefficient such that rC 3 8 x s.A mixture of Pr'OH and PrOH was used at temperatures no higher than 143.5 K. The triplet was produced by photolysis of the appropriate diazene as shown in Scheme 6. Cutting off the light resulted in a 'clean' second-order decay with Scheme 6 a rate coefficient (143.5 K) of (2*O0.8)x lo31 mol-'s-' i.e. about 0.13 times the diff usion-controlled encounter frequency. The decay could occur either through a 27 F. P. Sargent and E. M. Gardy Canad. J. Chem. 1976,54 275. 28 A. R. Forrester and J. S. Sadd J.C.S. Chem. Comm. 1976 631. 29 M. S. Platz and J. A. Berson J. Amer. Chem. Soc. 1976 98 6743. Reaction Mechanisms-Part (iii)Electron Spin Resonance Spectroscopy 79 singlet +triplet reaction or uia a triplet +triplet dimerization. Earlier work3' had shown that the singlet state (S) of (21) lay at least 2.5 kJ above the triplet (T).From this it was estimated that [S]/[T] -0.04 and hence if the singlet +triplet mechanism were correct the collisional frequency would have to be at least three times greater than the diffusion-controlled limit to account for the observed value of the second- order rate coefficient. On this basis it was decided that the dimerization took place uia the triplet +triplet reaction. The kinetics of cycloaddition of (2 1) to a number of olefins were followed and shown to be first-order in each reactant. The second-order rate coefficients were claimed to be very similar to those observed for the addition of radicals to 01efins.~~ The field of radical rearrangements remains a fertile one for e.s.r.spectroscopists. The kinetics of isomerization of cyclopropylcarbinyl to allylcarbinyl have been as have those of several neophyl rearrangement^.^^*^^ A valence bond isomerization of two hemi-Dewar naphthalenes to the corresponding naphthalenes is said to take place via the radical anions and the diamagnetic dianions of the naphthalene^.^^ The formation and decay kinetics of the e.s.r. signal of the radical anion of 1,3,7,9-tetra-t-butyl hemi-Dewar naphthalene were measured. Activation energies for formation and decay were 4.2 f2.1 and 17 f5 kJ mol-' respectively. It has been pointed that e.s.r. observations of radical rearrangements are often frustrated by the fact that the rearrangement unless occurring with a very small activation energy cannot compete kinetically with recombination and disproportio- nation reactions.Isolation in an adamantane matrix was therefore used to study the sigmatropic and electrocyclic reactions of the bicyclo[ 3,l ,O]hexenyl radical (22); X2= Y3= H). The radical was produced by X-irradiation of bicyclo[3,1,0]hex-2- ene in adamantane at -196 "C. Below -60 "C (22) was stable but above -60 "C it isomerized to the cyclohexadienyl radical (23) as shown in Scheme 7. Kinetic measurements on the growth of (23) gave a free energy of activation (-50 "C) of AG' = 60.6 kJ mol-'. When X =D and Y = H the cyclohexadienyl spectrum was found to be a composite of monodeuteriated radicals with deuterium in positions 1 2 and 3 in the statistical ratio 2 :2 1. The same result was found for X = H Y = D.The rearrangements shown in Scheme 7 are clearly more rapid than the ring-opening process. The kinetics of isomerization of 2,4,6-tri-t-butylphenyl (24) to 3,5-di-t-butylneophyl (25) have been measured over the temperature range -26 to -160 0C.37 In addition the same radicals deuteriated in the Butgroups were studied between 20 and -150 "C. The authors listed four criteria necessary to establish the presence of quantum mechanical tunneling and showed that their data satisfied all of them. The criteria are (i) a large kinetic isotope effect -k,/k = 80 at -30 "C 1400 at -100 "C and 13000 at -150"C; (ii) non-linear Arrhenius plots -tunneling should become relatively more important with decrease in temperature; (iii) a large 30 M.S. Platz J. M. McBride R. D. Little J. J. Harrison A. Shaw S. E. Potter and J. A. Berson J. Amer. Chem. SOC.,1976,98 5725. 31 J. M. Tedder and J. C. Walton Accounts Chem. Res. 1976,9 183 and references therein. 32 B. Maillard D. Forrest and K. U. Ingold J. Amer. Chem. Soc. 1976,98 7024. 33 B. Maillard and K. U. Ingold J. Amer. Chem. SOC. 1976 98 1224. 34 B. Maillard and K. U. Ingold J. Amer. Chem. Soc. 1976,98 4692. 35 I. B. Goldberg H. R. Crowe and R. W. Franck J. Amer. Chem. SOC. 1976,98 7641. 36 R. Sustmann and F. Liibbe J. Amer. Chem. SOC.,1976,98,6037. 37 G. Brunton D. Griller L. R. C. Barclay and K. U. Ingold J. Amer. Chem. SOC. 1976,98 6803. A. T.Bullock 1 1 1 D Q HH Scheme 7 But Bu' (24) (25) difference in activation energies for H and D transfer -in the absence of tunneling the maximum value of ED-EHshould equal the difference in zero point energies for H-and D-containing reactants (5.7 kJ mol-I) otherwise ED-EH>5.7 kJ mol-' when tunneling is appreciable (at -30 "C ED-EH= 13.4 kJ mol-'); (iv) a large difference in preexponential factors (experimental values were AH= 106.5s-' and AD= lo7.'s-').Detailed calculations were made37 using a one-dimensional Eckart gaussian and truncated parabolic barriers were tried and good agreement with the experimental data was found for the Eckart barrier i.e. Vcxl= Vo/cosh2(xla). Elementary reactions of atoms with molecules in the gas phase continue to be studied using the fast flow microwave discharge technique.As examples the reactions of atomic hydrogen and deuterium with HBr and DBr have been examined,39 and the kinetics and stoicheiometry of the reaction of atomic oxygen with PF3 have been mea~ured.~' Special mention should be made of two papers which significantly extend the scope of the technique. In the first of these4' the rate coefficients for the reactions F+H2+HF +H H +F2+HF+F and F+ CHF3+HF+CF2 were measured at 298 K. Finite difference techniques were used. 38 R. J. LeRoy E. D. Sprague and F. Williams J. Phys. Chem. 1972 76 546. 39 H. Endo and G. P. Glass J. Phys. Chem. 1976,80 1519. 40 I. B.Goldberg and H. R. Crowe J. Phys. Chem. 1976,80 2407. 41 I. B.Goldberg and G. R. Schneider J. Chem. Phys. 1976,65 147. Reaction Mechanisms-Part (iii) Electron Spin Resonance Spectroscopy 81 These permit the use of e.s.r.in complex systems which may not be amenable to study by simple first-order kinetics and where cavity sensitivity as a function of length along the reaction path may be important. Pressure and velocity gradients may also be taken into account. The second paper was concerned with the kinetics of the reaction between hydrogen atoms and molecular chlorine.42 The authors took the obvious but important step of adding an on-line mass spectrometer to the discharge- flow/e.s.r. system. By this means it was shown that C12 consumption and HCl production were much lower than expected; facts which could be accounted for by postulating an enhanced rate for H + HCl -+ H2+ C1 due to vibrational excitation of HCl produced in the initial step (H + C1 +HCl+ CI).In a subsequent study of the reactions of hydrogen and oxygen atoms with the same workers have taken the further step of adding a gas chromatograph to analyse stable condensables (77 K). 2 Chemically Induced Dynamic Electron Polarization An extended kinetic model for CIDEP has been presented which enables kinetic e.s.r. studies to be made at shorter time intervals than hitherto.44 The model incorporates emissive-absorptive (E-A) polarization due to the radical-pair mechanism (RPM) together with initial polarization (IP) arising from the triplet mechanism (TM) and provides a procedure for measuring true radical concentrations when one or both of these mechanisms are operative. For pairs of peaks with the same [xiaimi),the difference in peak heights represents the E-A polarization whilst the sum incorporates the IP.The model was tested using data obtained from a pulse photolysis study of the behaviour of the semidione radicals of pyruvic acid and biacetyl in alcoholic solvents. A spectrometer with a response time of ca. 0.3 ps has been described4’ which allows the observation of e.s.r. signals at time intervals of less than 1ps after a radiolysis pulse. The e.s.r. time profiles were analysed by means of the Bloch equations modified to take account of changes in radical concentration with time any IP on formation and CIDEP effects produced during decay. It was shown that the time dependence of the e&signal clearly indicated a zero initial magnetization and hence equally populated spin states.The relaxation time varied with dose in a manner which suggested a Heisenberg spin exchange mechanism. Secondary radi- cals produced from e all showed zero initial magnetization whilst secondary radicals produced by the reaction of *OH with several substrates gave time profiles showing initial magnetizations corresponding to the Boltzmann distribution. This was a result of spin relaxation of *OH before reaction and it was shown that T,(*OH)< 1ns. Another example of polarization transmission down a reaction pathway was found in a CIDEP study of the photoreductions of some carbonyl compounds in the presence of amine~.~~ The reaction goes via the triplet state of for example benzophenone thus 3Ph2C0+ NEt3 + Ph2COH+ MeCHNEt (19) 42 P.F. Ambidge J. N. Bradley and D. A. Whytock J.C.S. Faraday I 1976 72 1157. 43 P. F. Ambidge J. N. Bradley and D. A. Wbytock J.C.S. Faraday I 1976 72 1870. 44 P. B. Ayscough G. Lambert and A. J. Elliott J.C.S. Faraday I 1976,72 1770. 45 N. C. Verma and R. W. Fessenden J. Chem. Phys. 1976,65,2139. 46 K. A. McLauchlan and R. C. Sealy Chem. Phys. Letters 1976 39 310. 82 A. T. Bullock Only Ph2COH was observed. The failure to detect the counter radical MeCHNEt, the presence and nature of which had previously been established by flash photo- lysisY4’was ascribed to the fast pseudo-first-order reaction (20). The authors found that the polarization of Ph2COH varies not only with [NEt,] but also with [Ph,CO].MeCHNEt +Ph2C0 k CH,=CHNEt +Ph2COH (20) Thus although the absorption signal changed only slightly as [Ph,CO] varied from 0.02M to OSM the emission was much greater for the 0.5M solution. This was attributed to the rate of reaction (20) being sufficiently fast that an appreciable fraction of the primary polarization in the aminPalky1 radical was transferred to Ph,COH. In other words kZO[Ph2CO] -K’(MeCHNEt,). The RPM for two unlike radicals showing CIDEP predicts emission for the species with the higher g value and absorption for the other. However precisely the opposite has been found in the case of eiq (g = 2.0003),48i.e. for all counter radicals having g >2.0003 the e.s.r. signal for eiq was in the emissive mode. The authors suggested that if the RPM is invoked then it must be assumed that reaction between eiq and the counter radical occurs preferentially into the triplet state of the product.Both time-resolved and steady-state photolyses were used in a study of the pho toreduct ions of substituted benzoquinone nap h thoquinone ,and an thraquinone by 2,6-di-t-b~tylphenol.~~ In all cases it was possible to observe both primary radicals in the radical pair. The RPM and TM each contributed to the polarization but the latter was the major contributor. Various aspects of both theories were tested. An interesting feature of this work lies in some preliminary isotope experi- ments which suggested that CIDEP may be a potentially powerful tool for the investigation of kinetic isotope effects involving excited triplet molecules.Polariza- tion enhancements for the species *CH2COO- and CH(COO-)2 produced by electron pulse radiolysis provided another test of the~ry.~’ The E-A enhancements typical of RPM were studied as a function of ionic strength and radical concentra- tion. With regard to the former parameter the theory of Freed and Pedersen” predicts that for non-spherical radicals and small exchange interactions (J) the enhancements should show a stronger dependence on p the ionic strength at low values of p. This was confirmed. Furthermore it was found that *CH20H -CH,COO- and *CH(COO-) all showed polarization enhancements independent of radical concentrations. It was suggested that the initial non-uniform spatial distribution of radicals in spurs may have been responsible i.e.the observations (2-3 ps after the pulse) were made before a uniform distribution was reached. Finally the power of the correlated use of CIDNP and CIDEP in revealing details of major and minor pathways in radical reactions has been demonstrated in a study of the photolysis of tetrafluro-p-benzoquinone in di~xan.~ The photoreactive species was found to be the excited triplet of the quinone and the TM was mainly responsible for CIDEP. The cage recombination of the primary radicals was revealed by the CIDNP measurements and all observations were accounted for by the reactions 47 S. Arimitsu H. Masuhara N. Mataga and H. Tsubomura J. Phys. Chem. 1975.79,1255. *8 R.W. Fessenden and N. C. Verma J. Amer. Chem. SOC.,1976,98,243.49 B.B.Adeleke and J. K. S. Wan J.C.S. Faraday I 1976,72 1799. 50 A. D.Trifunac J. Amer. Chem. SOC.,1976,98,5202. 51 5. H. Freed and J. B. Pedersen Adu. Magn. Resonance 1975,8,2. 52 H. M. Vyas and J. K. S. Wan Canad. J. Chem. 1976,54979. Reaction Mechanisms -Part (iii) Electron Spin Resonance Spectroscopy shown in Scheme 8. The authors make the important point that conventional kinetic e.s.r. only gave information about disproportionation (kd).53 caged tripletpair 0 6 0 OH g1 ’ g2 FFQ diffusion OH 0 OH OH 0 OH Scheme 8 53 H. M. Vyas and J. K. S. Wan Internat. J. Chem. Kinetics 1974,6 125.