CHAPTER 1
Organic Radicals in Solution
BY B. J. TABNER
1. Introduction
On this occasion my report covers the literature published between June 1984 and May 1985. This is a shorter period than in the corresponding Chapter in Volume 9. However, bearing this in mind, the number of papers published per annum in which e.s.r. spectroscopy has been applied to the study of organic radicals in solution remains fairly constant. I have retained the same general review areas as in Volume 9 but have decided to include an additional section dealing with CIDEP. This latter technique has attracted the interest of a number of research groups and a useful review dealing with the origins of spin polarization and some of the characteristics of CIDEP effects has recently appeared.
The application of e.s.r. spectroscopy to the study of organic radicals and radical ions continues to show a remarkable diversity. Objectives vary from obtaining information on the structure of these species, to establishing their presence as reaction intermediates, and to the study of the rates of their reactions. Radical cations generated at low temperature in Freon and related matrices have continued to attract considerable attention and readers interested in this area will find a recent review by Symons extremely valuable. The ENDOR technique, which can overcome some of the resolution limitations of conventional e.s.r. spectroscopy, has also been reviewed.
Applications of microcomputers to e.s.r. spectroscopy continue to appear. These vary from systems designed to improve spectral quality and display, to applications of the fast Fourier transform method, and to the analysis of e.s.r. spectra employing autocorrelation techniques.
2 Carbon-centred Radicals
2.1 Alkyl Radicals.-As in previous reports I have divided the first part of my report, dealing with carbon-centred radicals, into two sections. The first section covers 'simple' alkyl radicals and the second section delocalized radicals. Recent investigations of alkyl radicals again include a wide range of interests such as the examination of their structure, the selectivity of their attack, and some of their addition and rearrangement reactions. There are also several publications to report covering attempts to clarify various reaction mechanisms and the determination of absolute rate constants for some fundamental radical reactions.
The structure and conformation of radicals remains a topic of fundamental importance and several recent publications illustrate this interest. Radicals of the type R3MCH2CH2 (M = Si, Ge, or Sn) exist in a preferred conformation with the metal substituent in an eclipsed position relative to the singly-occupied p-orbital. The e.s.r. spectra of (RCH2)2CX radicals (X = H, Me, or OSiMe3) show two different methylene proton couplings indicating that the two methylene groups have different orientations relative to the singly-occupied molecular orbital. It appears that the ~n-conformation is preferred over the anti-conformation. The activation energy for the 'flip-flop' motion of these radicals lies in the range 12-40 kJ mol-1. The main feature of the spectrum of the F-2,4-dimethyl-3-ethyl-3-pentyl radical consists of a doublet of 4.50 mT. This is consistent with a conformationally locked structure in which one S-F of the C2F5 group is nearly eclipsed with the half filled Q-orbital, while the other lies near its nodal plane. These radicals are remarkably stable with a half life of 1 hour at 373 K. A further illustration of the information obtainable from a study of long range couplings has been reported in which the conformations of radicals with the structures ·C(R1)(R2)OC(O)CH(R3)(R4) and CH2(R1)OC(O)C(R2)(R3) have been determined from the temperature variation of their [??](δ-H) values. This method provides a useful alternative approach to establishing preferred conformations. The spectra of some sterically crowded radicals [e.g, (Pri)2(But)C·] are independent of temperature and these radicals appear to exist in a frozen cogwheel conformation. In a study of a somewhat different nature the isooctyl radical [a(6H) 2.30 and a(2H) 1.18 mT] has been trapped in a thiourea canal complex. The resulting e.s.r. spectra are well resolved and indicate that the radical can undergo random motion if the diameter of the canal is large. The methyl groups appear to rotate freely at 77 K.
The selectivity of radical attack has been studied for many years and continues to attract interest. Gilbert et al. have undertaken a comparative study of the reactions of 'OH (an electrophilic radical), SO4- (a powerful one-electron oxidant), Ph (a relatively nucleophilic radical), and CO2- (a one-electron reductant) with a series of halogen-containing substrates. It is interesting that despite their different polar and steric characteristics OH, Ph and CO2- all show similar trends in their attack, with iodine-atom abstraction preferred to hydrogenatom abstraction in iodine containing carboxylic acid derivatives. However, OH abstracts hydrogen in preference to chlorine or bromine, a result which can be rationalized on thermodynamic considerations. The t-butoxy radical has been employed to study the regioselectivity of hydrogen-atom abstraction from fatty acids. For example, the spectra reveal that reaction of this radical with pentanoic acid results in the formation of three radicals CH3CH2CH2CHCO2H, CH3CHCH2CH2CO2H, and CH3CH2CHCH2CO2H (relative proportions 0.41:0.47:0.12). Thus hydrogen-atom abstraction from carboxylic acids by t-butoxy radicals appears to be a fairly random process.
With many substrates one-electron oxidation, radical addition, and atom abstraction reactions are all possible and some of these possibilities are observed in the reactions of 'OH, Cl2-, and SO4- with alkenes and dienes. Thus it appears that both SO4- and Cl2- are capable of effecting one-electron oxidation with some substrates, but there is clear evidence for the formation of so4: and Cl2- adducts with simple substrates. A reaction of fundamental importance in the study of radical polymerization is the addition of an initiating radical to a suitable vinyl monomer. It is accepted that such addition will normally occur at the least substituted carbon atom of the C:C bond. However evidence has now been presented that the benzoyloxyl radical can add to both carbon atoms of the C:C bond of vinyl acetate. The experiments, employing 2,4,6-tri-t-butylnitroso-benzene as a spin-trap, reveal overlapping spectra of two trapped radicals following addition of the initiating radical to both carbon atoms. In another study of addition reactions the photolysis of disulphides, RSSR, in the presence of acylsilanes [R1C(O)SiMe3] in cyclopropane leads to the formation of R1(RS)COSiMe3 radicals rather than R1CMe3Si)COSR radicals. Although the mechanism of formation of these radicals is unclear this conclusion was confirmed when the same spectra were obtained following the reaction of Me3Si radicals with R1C(O)SR.
Hatano et al. report the use of the combination of the spin-trapping technique (employing 2-methyl-2-nitrosopropane) with high performance liquid chromatography, in which the individual spin adducts are isolated from their mixtures. In all cases the primary radicals were produced by γ-irradiation. It appears that amino acids can give spectra resulting from reaction with the hydrated electron or by attack of the 'OH radical and in some cases diastereomeric pairs of radicals can be individually separated. Diastereomeric pairs of radicals were also individually separated following γ-irradiation of prolines (such as cis-4-chloro-L-proline). The relatively small g(β-H) values for these diastereomers suggest that the β-hydrogen is approximately in the nodal plane of the 2pz orbital on the aminoxyl nitrogen. In the analogous study of some tripeptides four types of spin adducts were separated and identified corresponding to both backbone and side-chain adducts. The spin-trapping technique has also been used to study radicals produced by the autoxidation of monosac-charides in which the main species trapped were hydroxyalkyl radicals.
E.s.r. spectroscopy provides an informative method for studying ring-opening and rearrangement reactions involving radicals. The ring-opening reaction of cyclopropylalkyl radicals has been studied extensively but it is interesting to compare the ring-opening reaction of the 2-methylaziridine-boryl radical (1) with its isoelectronic carbocyclic analogue. 23 Two radicals are observed from both cis-and trans-(1). From the cis form radical (2) [a(H) 2.16, a(2H) 2.80, a(3H) 2.49, and a(N) 0.32 mT] dominates, whilst radical (3) [a(2H) 2.21, a(H) 3.40, and a(N) 0.34 mT] dominates from the trans form. An illustration of the application of e.s.r. to the study of rearrangement reactions is found in β-(alkylthio) ethyl radicals. Although 1 ,4-hydrogen shifts are seldom observed there is clear evidence for such a shift in the case of this particular radical. Photolysis of dimethylsulphide with ethene in cyclopropane gives the spectrum of the CH3SCH2CH2 radical [a(2,α-H) 2.03 and a(2,β-H) 1.29 mT] together with a secondary spectrum [a(α-H) 1.625 and 1.725, and a(2,γy-H) 0.14 mT] assigned to CH3CH2SCH2. Presumably the driving force for this rearrangement is the greater stability of the latter radical. There is clear evidence that the bromine atom migrates between two equivalent sites in the Me2CC(Br)Me2 radica1. In the asymmetric structure six strongly coupled and six weakly coupled protons would be expected. In fact coupling to twelve equivalent protons is observed with a splitting constant half that of the slow exchange value indicating that the migration of the bromine atom is very rapid at 100 K.
One of the attractions of e.s.r. spectroscopy is its potential as an aid to establishing intermediates in a variety of reactions. Several publications have appeared illustrating this application. For example, when a dithiocarbonate, R1OC(O)X (X = SMe, R1 = Me), is subjected to u.v. irradiation in methylcyclo-hexane in the presence of hexamethylditin and di-t-butyl peroxide an e.s.r. spectrum with a(3H) 0.130 mT, assigned to MeOCS, is observed (Reaction 1). Thus the identification of alkoxythio-carbonyl radicals as primary products suggests that the Barton deoxygenation of alcohols might involve the same intermediates. Two other publications give further illustrations. One of these describes addition/elimination in the reduction of nitro-benzene by α-hydroxyalkyl radicals the other describes the capto-dative radical Me2C(CN)CH2C(CN)SEt. The hyperfine splitting constants of this latter radical are markedly effected by coordination with SnCl4 and further investigations in this area may prove helpful in understanding the role of Lewis acids in radical polymerization. A particularly interesting study has been made of the generation of radicals during the ultrasonic decomposition of organotin compounds. Although it has been appreciated for many years that intense ultrasonic waves induce chemical reactions the mechanisms of such reactions are largely unknown. Hexabutyltin (in benzene) gives two e.s.r. spectra (employing nitrosodurene as a spin trap) one with a(N) 1.349 and a(2,β-H) 1.05 mT corresponding to the trapping of Me3C· radicals and one with a(N) 1.053, a(2H) 0.270 and 0.075, and a(119Sn) 0.76 mT corresponding to the trapping of R3SnC6H4 (formed by reaction with the solvent). These results strongly suggest sonochemical cleavage of the Sn-C bond.
MeO-CS2Me + Me3Sn' [right arrow] + MeOC=S + Me3SnSMe (1)
One of the most fundamental applications of e.s.r. spectroscopy is to the determination of absolute rate constants for reactions involving radicals. Fischer et al. have explored a number of reactions of transient radicals. For example, the photolysis of diisopropyl ketone gives Me2CH radicals which can self-terminate by disproportionation or combination. (Disproportionation is favoured at 293 K). The isopropylol radical is particularly interesting since it can exist in both neutral (Me2COH) and charged (Me2CO-) forms. Consequently if termination is studied in the region pH ca. pK information on the cross-termination of the neutral radicals with their ionic counterparts can be obtained. Self-termination of the neutral radicals is close to the diffusion-controlled limit but the smaller rate constants for cross-and self-terminations involving the anion reflect the effect of charge repulsion. Fischer et al. have also employed the intermittent photolysis technique to study a rather more complex system in which both isopropylol and t-butyl radicals are present. Again cross-and self-termination rate constants were determined (ca. 109 M-1 S-1 at 300 K) and their relatively low values ascribed to steric constraints. The stopped-flow technique has been used to study the second order self-decay reaction of malonyl radicals [·CH(CO2H)2, a(α-H) 2.02 mT, k ca. 109 M-1 S-1]. One other kinetic study to report is chlorine atom abstraction from CCl4 by Cl3C(CH2CHCl)n radicals (k ca. 10-3 M-1 s-1 at 313 K).
Only a comparatively small number of publications have involved the study of cyclic alkyl radicals. Several of these are concerned with the conformation of these radicals. For example, the magnitude of the β-proton hyperfine coupling in a series of 2-cyclohexanonyl radicals suggests that they have a half-chair structure in which four of the ring carbon atoms lie in a plane. The selective line broadening observed in the spectra of these radicals has been interpreted in terms of a two site model for ring inversion (for the 2-cyclohexanonyl radical itself ΔH+ = 15.5 kJ mol-1). The hyperfine splitting constants in tetraalkyl-glucosyl radicals, formed by photolysis of the corresponding ketone with hexamethyltin, are reconcilable with a boat conforMation. Evidence for this structure is found in the relatively large a(γ-H) values of 0.34-0.38 mT. The inversion barrier in a seven membered ring system, the cycloheptyl radical, has been determined. At low temperature the 'frozen' conformation has a(α-H) 2.10 and a(2,β-H) 3.14 and 2.10 mT. As the temperature is raised the classic alternating linewidth effect is observed until at 275 K all four S-protons become equivalent (Ea [??] 14.3 kJ mol-1). Two papers by Gilbert et_al. deal with cyclic alkyl radicals produced by reaction with 'OH. Attack of 'OH on Ph(CH2)3CO2H depends upon pH. Above pH 3 the spectrum of radical (4) is observed but as the pH is lowered this is succeeded by the spectrum of (5) and at pH < 1.5 by that of (6). It is proposed that the precursor to (6) is a radical cation. The fact that the 'OH radical is often very unselective in its site of attack is further illustrated in the study of its reaction with polysaccharides. However, some degree of selectivity is observed with D-galacturonic acid where abstraction predominantly occurs at the carbon atom adjacent to the carbonyl group.
Finally in this section there are two papers to report which describe further work on acyl radicals. The photolysis of di-t-butyl peroxide with aromatic aldehydes in cyclopropane produces two conformers in the case of furaldehyde. Conformer (7) has a(2H) 0.02 and a(H) 0.265 mT and conformer (8) has a(H) 0.23 and 0.04 mT. The spectrum of the 3-thenoyl radical (9) shows an alternating linewidth effect as a consequence of restricted rotation about the C-CO bond. In the second paper Pedulli et al. have studied the photoreaction of tetraethyl pyrophosphite with aroylsilanes. Two radicals are observed in the spectrum one of which can be assigned to Ar[(EtO)2(O)P]C-OSiR3. The spectrum of the other radical (which is stable up to 420 K) has A(31p) ca. 2.6 mT and is assigned to Ar[(EtO)2(O)P]C-OSiR3. The mechanism proposed to explain the formation of these radicals involves the homolytic cleavage of the C-Si bond to form the ArC=O and R3Si radicals.
