CHAPTER 1
Organic Radicals in Solution
BY B. J. TABNER
1 Introduction
My report for Volume 11 A covers the literature published between June 1985 and May 1987. This is double the period covered in my previous report.
It has become apparent during the preparation of this report that the study of organic radicals in solution by e.s.r. spectroscopy continues to retain a considerable interest. This is not only due to the continual development of new areas of study but also reflects recent advances in instrument technology (for e.g., the development of the spin echo technique). Consequently e.s.r. is now being applied to study new problems, particularly those involving transient species. In addition to these new areas, e.s.r. spectroscopy will always play an important role in the investigation of radicals present as reaction intermediates and in the study of the kinetics of their reactions. There is also a wealth of structural information that can be obtained from a study of e.s.r. spectra. With several new techniques available for the analysis of complex hyperfine patterns, particularly ENDOR spectroscopy, it is now possible to examine systems of increasing size and complexity. Thus, with the unpaired electron able to act as a probe, information can be obtained on the structure of larger and more complex systems. The hyperfine data obtained in such investigations not only reveals information on the unpaired electron distribution throughout the π-system but is also available to obtain information on such features as substituent effects and conformational preferences.
Readers interested in ENDOR spectroscopy and its methodology will find a useful review in Volume 10 B of these reports. This latter review covers several recent developments as well as instrumental aspects and advanced techniques. An important nucleus in magnetic resonance is 13C as it provides a means of obtaining information on those positions in a molecule to which no proton is attached. Its hyperfine splitting constant is often quite sensitive to structural changes. Normally 13C-enriched samples are required but deuteriation provides a means of studying 13C ENDOR when the nucleus is present in natural abundance.
It is pleasing to see that the text by Wertz and Bolton, originally published in 1972, has now been reprinted, and also that collections of papers presented at several conferences have now been published. Since my last report the plenary lectures and papers presented at the 22nd AMPERE Congress held at Zurich, and at the Faraday Discussion on 'Radicals in Condensed Phases' have both appeared. In a new innovation the papers presented at the 18th International Conference on ESR in Organic and Bio-organic Systems, held at Leeds, have been published in a special issue of the Faraday Transactions.
Several reviews have appeared which readers interested in organic radicals will find valuable. Symons and Cox have compared the power of µSR and ESR in probing structures of radicals. Due to the greater zero-point energy of muonium (relative to hydrogen) large hyperfine isotope effects are observed in µSR. There appears to be the possibility of preparing some novel species, such as muonated radical cations and radical anions. Another technique that has developed greatly over the last few years is that of ESR imaging. In two reviews Ohno describes the technique and some of its applications. This technique has applications to the study of diffusion of organic radicals in polymers and to the study of the distribution of radicals after the mixing of two flowing reagents.
The application of e.s.r. spectroscopy to the study of small strained radicals has been reviewed by Ingold and Walton. The use of the technique to study the rearrangement reactions of such radicals, and to provide information on their configuration, is described.
The growing interest in identifying radicals very shortly after their formation and following their subsequent reactions is the subject of a review, by McLauchlan, of flash photolysis e.s.r. One of the main continuous-wave techniques for obtaining e. s. r. spectra of transient radicals produced by flash-photolysis involves time-integration. A variant of this method (MISTI), involving switching on the microwave field after radical creation, provides a method for enhancing signals. A further development in the study of radicals very shortly after their creation, and their subsequent recombination reactions, involves the study of the 1: Organic Radicals in Solution effects of magnetic and microwave fields. The problems encountered as a result of non-uniform concentration distributions following photochemical radical generation have also been discussed.
The post-experimental treatment of spectroscopic data continues to be aimed at resolution enhancement (so that information can be obtained on long-range small hyperfine couplings) employing the Fourier transform method, and at the extraction of hyperfine splitting constants in complex or poorly resolved spectra with the aid ofmicrocomputers.
2 Carbon-centred Radicals
As on previous occasions I have divided this, the first part of my report, into two sections. In the first section I shall cover 'simple' alkyl radicals (including cyclic alkyl radicals) and in the second section I shall cover delocalized radicals. Once again the investigation of alkyl radicals by e.s.r. spectroscopy covers a wide range of interests. The study of the structure and conformation of these radicals continues, as does that of their formation by addition and abstraction reactions. There are several reports of rearrangement reactions, and the determination of rate constants for some important reactions has continued to attract attention.
2.1 Alkyl Radicals.-Ingold and Walton have exploited the e.s.r. technique to study the conformation of cyclohexylmethyl radicals. The spectra of the cyclohexylmethyl and (4-alkylcyclohexyl)methyl radicals (at 140 K) have a(β-H) ca. 4.2 - 4.3 mT when the CH2 group adopts the axial conformation, and a(13-H) ca. 3.0 - 3.1 mT when it adopts the equatorial conformation. Thus the two conformations can be readily distinguished from the magnitude of their a(β-H) splitting constants. Both conformers are separately observed up to 400 K indicating that the rate of their interconversion is slow on the e.s.r. timescale. However, it has been possible to determine the rate constant for ring inversion for the cis-4-methy1cyclohexylmethyl radical. Both the quasi-axial (1) and the quasi-equatorial (2) conformers can be observed in the cyclohex-2-enylmethyl radical. The magnitude of a(β-H), and its variation with temperature, shows that these radicals adopt an eclipsed conformation about the Cα-Cβ bond. Ring-opening occurs in some methyl-substituted cyclobutylmethyl radicals. For example, the trans-2-methylcyclobutylmethyl radical has a(2α-H) 2.21, a(β-H) 0.80, a(2γ-H) 0.15, and a(3δ-H) 0.073 mT when prepared at 140 K. However, at temperatures above 240 K ring-opening occurs and the hex-5-en-2-yl radical, with a([apha]-H) 2.20, a(2γ-H) 0.07, and a(5β-H) 2.51 mT, is observed. At 90 K rotation of the methyl group is effectively frozen in the trans-2-methylcyclobutylmethyl radical with one of the δ-hydrogens of the methyl group occupying an all trans-conformation with respect to the radical centre.
Other conformational studies include those of the 2, 2-dimethy lbuty1 radical and some acyclic alkyl radicals. The 2,2-dimethylbutyl radical has a(2α-H) 2.21 and a(8γ-H) 0.09 mT at 250 K. Below 120 K the rotation of both the methyl groups and of the ethyl group are at the slow exchange limit [a(5γ-H) 0.07 and a(γ-H) 0.60, 0.21, and 0.14 mT at 85 K]. The observation of long-range hyperfine couplings are a common feature when radicals have a fixed geometry. Consequently spectral interpretation indicates that many acyclic alkyl radicals, such as 2-methylalkyl and 2,2,3,3-tetramethylbutyl radicals, exist in a preferred conformation about the Cα-Cβ and Cβ-Cγ bonds. Information concerning the conformation about Cγ-Cδ bonds, however, is often limited by the resolution of the appropriate hyperfine splittings.
The Me3CCH2C(OMe)COOMe radical is stable enough to be observed by e.s.r., possibly because of capto-dative stabilization. The radical becomes locked in a twisted conformation upon complexation with SnCL4 with two non-equivalent β-protons (0.746 and 1.030 mT). Aminoalkyl radicals, (H2NCHR), containing an acceptor substituent (R) should also be capto-dative stabilized. These radicals have two non-equivalent hyperfine splittings from the amino hydrogens. A two jump model has been used to estimate the barrier to rotation about the C-N bond in these radicals.
Addition and abstraction reactions are both widely used in the preparation of radicals. Addition to a C:C bond invariably occurs at the least substituted carbon atom (to give a tertiary or a secondary radical). However, 'head' addition to methylenecyclo-alkanes becomes thermodynamically more favoured as the size of the ring decreases. Indeed addition of MeSiO' to methylenecyclo-propane appears to be predominantly 'head' (3) giving (4) as the radical observed [a(2α-H) 2.23, a(2β-H) 2.63, and a(δ-H) 0.034 mT]. Addition of Cl2-,SO4-, and ·OH also occurs readily to the C=C bond in alkenols and alkenoates. However, the addition of SO4: or Cl2: (or ·OH at pH < 1) to Me2C:CH(CH2)2CH(Me)OH leads to the observation of radical (5), formed via a cyclization reaction. Aliphatic alcohols, such as 3-buten-2-ol, also yield radicals by ·OH addition to the C=C bond. However, three different radicals are formed by hydrogen atom abstraction from 1 ,3-butanediol [CH(OH)CH2CH(OH)CH3, CH2(OH)CHCH(OH)CH3, and CH2(OH)CH2C(OH)CH3]. Vinyl and propenyl ethers can also give radicals by either addition or abstraction. For example, the t-butoxyl radical reacts with MeCH20CH:CH2 to give MeCH2OCH=CH2 [a(H) 1.532 and 0.014, A(3H) 2.248, and a(2H) 0.123 mT] and MeCH2OCHCH2OCMe3 [A(H) 1.628 and a(2H) 0.738 and 0.240 mT]. The photochemical hydrogen atom abstraction from alcohols by benzophenone has also been reported.
There is increasing evidence for hydrogen shifts in certain radicals. For example, the addition of MeCHOEt (generated from Et2O) to an alkyne, such as butynedioic acid, gives a spectrum with a(H) 2.55, 3.23, and 3.22 and a(2H) 0.04 mT. It is proposed that this spectrum arises from the cyclic radical (6) and that the first formed vinyl radical undergoes a rapid 1,5-shift followed by ~.d..Q cyclization to the observed radical. There is evidence for a 1,6-shift, again followed by cyclization, upon addition of CH20CMe3. A 1, 5-shift has also been observed in the rearrangement of RCH2(CH2)CCl2 to RCH(CH2)3CHCl2 (R = Me, Et, or MeO).
There is an interesting report by Symons et al, that γ-irradiation of MeCl in CD3CN at 77 K gives a methyl radical-chloride ion adduct. The spectrum of the adduct clearly exhibits coupling to 35Cl and 37Cl nuclei and it is proposed that the adduct is formed by electron capture. The methyl radical is irreversibly formed on annealing. Metal assisted radical reactions, such as polymerization with Lewis acids, are of some interest and a report has been made that the spin population in Me3CCH2C(SEt)CN varies upon coordination with SnCl4.
Davies et al. have studied the radicals formed upon photolysis of carboxylic acids either neat or in CCl4. When R = H, Me, Et, or Pr only RCHCO2H radicals are observed. However with pivalic acid only the Me3c radical is observed. Two models are proposed to explain the regiospecific formation of RCHCO2H.
A common radical reaction is fragmentation. A nice example of this type of reaction has been observed by Gilbert et al. when studying the photolytic decomposition of peroxydisul phate in the presence of carboxylic acids and related compounds. With ethanoic acid both CH3 and CH2CO2H are observed with their relative proportions varying with the ethanoic acid concentration. It appears that attack at the carboxy group is followed by decarboxylation. With propanoic acid, however, the primary reaction appears to be abstraction of a β-hydrogen atom, al though attack at the carboxy group also appears to occur. An example of a de hydration reaction is also reported. The initial α, β-dihydroxyalkyl radical formed from ethane-1,2-diols undergoes an acid catalysed dehydration reaction to give the corresponding carbonyl-conjugated CR2R3C(O)R1 radical. A kinetic study of these reactions indicates the importance of the electronic effects of substituents.
The thermal rearrangement of 2,2-bis(ethylthio)-3,3-dimethyl-pent-4-enal into 2,2-bis(ethylthio)-5-methylhex-4-enal is an example of a [1,3] carbon-carbon shift in an acyclic system. The observation of the (EtS)2CCHO radical during the course of this reaction supports a radical pathway for this isomerization.
Alkanes react with CC1 4 by a radical chain mechanism and three groups have studied various aspects of this process. Griller et al. have photolytically generated alkyl radicals (cyclopentyl and t-butyl) in CCl4· The spectra of the alkyl and the CCl3 radical are both observed. The rate constant for reaction of the alkyl radicals is greater in solution than in the gas phase possibly due to the solvation of a polar transition state. Similar reactions have been studied at 353 K using the spin-trapping technique and the rate constants for the reaction of Me2CCN and Me2CPh with CBr4 and CBrCl3 determined. The reaction between 1-hexene and CCl4 has also been studied using the spin-trapping technique and the presence of both CCl3 and Cl3CCH2CHCMe3 radicals confirmed.
The knowledge of absolute rate constants and activation parameters is necessary for a proper understanding of chemical reactivity. Several different research groups have reported results relevant to this topic. Roberts et al. have used laser flash photolysis in conjunction with e.s.r. spectroscopy to determine the activation energy for the decarboxylation of the t-butoxycarbonyl radical (45 - 50 kJ mol-1) and for the hydrogen atom abstraction by ButO. from cyclopentane (14 - 25 kJ mol-1). Ruegge and Fischer have also studied the former reaction and report an activation energy of 49 kJ mol-1. The self-and cross-termination reactions of the Me3COCO and Me3C radicals are both diffusion controlled (Ea ca. 8 kJ mol-1). Griller et al. have studied the reaction between simple alkyl radicals and 1,4-cyclohexadiene and find rate constants of ca. 104 - 105 mol-1 l s-1 at 300 K and activation energies of ca 21-30 kJ mol-1. Employing a spectrometer operating at 2 MHz magnetic field modulation Bolton et al. have determined the rate constant for the trapping of Me2COH by 5,5-dimethyl-1-pyrroline 1-oxide (1.1 ± 0.1 x 108 mol-1 l s-1).
The simplest member of the series of fully conjugated cyclic radicals of the general formulae C2n+1H2n+1 is cyclopropen-2-yl. The e. s. r. spectrum of the radical formed following y-irradiation has now been observed at 20 K. The spectrum consists of a large 1:1 doublet indicating that the "ethylenic" structure (7) is preferred over the "allylic" structure (8). The structure of the cyclopropyl radical has aroused interest. In this radical the 13C hyperfine couplings provide information about the degree of nonplanarity. In cyclopropyl and 1-methyl-cyclopropyl the a(α-13C) values of 9.59 and 9,8 mT respectively confirm these radicals have a 'bent' structure. In contrast, however, values of a(13C) 4.09 and a(4β-H) 2.70 mT in 1-(trimethylsilyl)cyclopropyl indicate that this radical is planar, or nearly so.
