Faraday Discuss. Chem. Soc., 1984, 78,49-55 Homolysis of Conjugated Carbanions and Carbenium Ions BY JOHN L. COURTNEIDGE, ALWYN G. DAVIES," PETER S. GREGORY AND SAFIEH N. YAZDI Chemistry Department, University College London, 20 Gordon Street, London WC 1 H OAJ Received 6th June, 1984 Methods are reported by which fulvalene and pentalene radical anions and cyclobutadiene, cyclopentadiene and buta- 1,3-diene radical cations can be prepared in fluid solution and studied by e.s.r. spectroscopy. These preparations illustrate the general principles that radical anions and cations can be regarded as the conjugate bases and acids, respectively, of neutral radicals and that bonds to metals and to hydrogen often show common properties. Most of the small hydrocarbon radicals that can be generated and studied by e.s.r.spectroscopy in fluid solution have hitherto been electrically neutral (e.g. methyl, vinyl and allyl), and the simplest common hydrocarbon radical ions were those of the arenes. This picture is changing rapidly with the development and application of methods for generating radical ions, particularly radical cations, of simple unsaturated hydrocarbons. We review here briefly some of our relevant work on neutral radicals, then present recent results of e.s.r. studies of the preparation of the radical anions and cations of simple T systems, particularly the radical cations of 1,3-dienes. PENTADIENYL AND CYCLOPENTADIENYL RADICALS Pentadienyl radicals can be generated by standard methods such as the abstrac- tion of hydrogen from a penta- 1,3- or penta- 1,4-diene, of halogen from a pentadienyl halide, or by the ring-opening of a cyclobutenylmethyl radical.' Cyclopentadienyl radicals are less easy to generate, as t-butoxyl radicals react with cyclopentadiene by addition rather than by hydrogen abstraction.It became possible to study these cyclopentadienyl radicals in detail only when it was found2 that the cyclopentadienyl derivatives of many metals were photosensitive, and irradiation of these compounds in solution, in the cavity of an e.s.r. spectometer, gave strong spectra of the corresponding cyclopentadienyl radicals (1): R (1) M = e.g. Li, $Hg, Bu3Sq3 Ph3Pb: Me2CpTi,' C ~ Z r c l ~ . ~ R = e.g. 'H, 2H, a l k ~ l , ~ Me,Si, C13Si, Me,Ge, Me3Sn.7 The magnitude of the I3C hyperfine coupling shows that these are wdelocalised radical^,^ and Huckel molecular-orbital theory has been used to interpret the sub- stituent effects on the e.s.r.spectra in terms of the electronic effects of the substituent 4950 CONJUGATED CARBANIONS AND CARBENIUM IONS group^.^,^-^ As we are dealing with neutral radicals in non-polar solvents, the effects of solvation are probably not important. No radicals are observed when cyclopentadiene itself is irradiated with ultraviolet light, but penta-alkylation of the ring confers a remarkable photosensitivity. If pentamethylcyclopentadiene is irradiated with ultraviolet light, unfiltered or filtered through Pyrex glass, a strong spectrum of the pentamethylcyclopentadienyl radical (2) is observed, a( 15H) = 6.35 G, and dihydrogen is evolved: This reaction exemplifies the principle that, in the same molecular environment, hydrogen often shows the same behaviour as a metal." Just as reaction (1) gives rise to the formation of a metallic radical, reaction (2) appears to involve the liberation of a free hydrogen atom from the ring, since in ethene solvent it can be trapped to form the ethyl radical.' I Experiments involving deuterium labelling show that, in forming dihydrogen, this hydrogen atom derives its partner not from the ring but from a methyl group." This reaction warrants a thorough photochemical study.FULVALENE AND PENTALENE RADICAL IONS This photolysis of cyclopentadienylmetallic compounds can be exploited in preparing radical anions of 5-annulene systems where the parent annulene itself cannot be isolated.Thus fulvalene polymerises in solution at concentrations > 0.00 1 mol dm-3. Dihydrofulvalene (3), however, is stable, and gives a stable disodium derivative. In line with reaction (l), if this is irradiated with ultraviolet light it shows the e.s.r. spectrum of the fulvalene radical anion (4), a(2H) = 1.55, a(2H) = 3.70 G:12 2Na' Na' (3) (4) Similarly, pentalene cannot be isolated, but dihydropentalene ( 5 ) can be prepared and converted into its dilithio derivative, and this, again on irradiation, generates the e.s.r. spectrum of the radical ion (6), a(2H) = 0.83, a(4H) = 7.73 G:13J . L. COURTNEIDGE, A. G. DAVIES, P. S . GREGORY AND S . N. YAZDI 51 CYCLOPENTADIENE RADICAL CATIONS Reactions (3) and (4) emphasise that protic radicals are formally related to the corresponding radical anions as acid and conjugate base [reaction (91, although in fact, in these reactions, the proton transfer preceded the homolysis: HB' $ B'-+H+ ( 5 ) Conversely, protic radical cations can be regarded as the conjugate acids of neutral radicals [reaction (6)]: A'+H+ s AH'+ (6) We anticipated therefore that the photolysis of precursors of cyclopentadienyl radicals under acid conditions should generate the corresponding cyclopentadiene radical cations, and indeed photolysis by unfiltered or Pyrex-filtered U.V. light of pentamethylcyclopentadiene in trifluoroacetic acid shows a strong spectrum of the corresponding pentamethylcyclopentadiene radical cation (7), a( Me) = 0.8, a( 1 H) = 1.6, a(2Me) = 4.0, a(2Me) = 15.0 G.I4 The hyperfine coupling constants of the methyl groups on the diene unit agree with the values calculated from simple Huckel theory, and the hydrogen atom on the ring, lying as it does in the nodal plane of the SOMO of the diene, is only weakly coupled: .. R R ' 'R When R = H, the reaction might be thought to proceed by formation of the pentamethylcyclopentadienyl radical [reaction (2)] (as does indeed occur in acetic acid solution), followed by rapid protonation of the neutral radical to give the radical cation. However, hexamethylcyclopentadiene, which is photostable in neutral solvents, shows a strong spectrum of the hexamethylcyclopentadiene radical cation (7, R = Me) when it is photolysed in trifluoroacetic acid (fig. 1). The reaction therefore appears, at least in this latter case, to involve the photolysis of a carbenium ion, as represented in reaction (7). We do not know yet whether a hydrogen atom is formed, or whether dihydrogen is liberated, to complete the analogy with reaction (2).CYCLOBUTADIENE RADICAL CATIONS The tetra-t-butylcyclobutadiene radical cation was prepared by treating tetra-t- butyltetrahedrane or tetra-t-butylcyclobutadiene with aluminium ch10ride.l~ Hogeveen provided a more generally useful route to these radical cations when he showed that the cyclobutadiene-AlC1, u complex (8), which is formed when but-2- yne is treated with aluminium chloride, shows the spectrum of the tetramethyl- cyclobutadiene radical cation (9) under U.V. irradiation, a( 12H) = 8.70 G.16 This route to an annulene radical by the homolysis of a ring-metal bond may be regarded as the cationic equivalent of reaction (1).It is not necessary to isolate the intermediate (8), and the whole reaction can be carried out in the e.s.r. tube.52 CONJUGATED CARBANIONS AND CARBENIUM IONS Fig. 1. E.s.r. spectrum of the hexamethylcyclopentadiene radical cation (7; R = Me), obtained by photolysis with U.V. light of hexamethylcyclopentadiene in trifluoroacetic acid at - 15 "C. Me AlCI, 2MeEMe CH,C.,b AICI, #"'* MeHMe Me Me Me Me (8) (9) By this means, using mixed alkynes, R + R', or mixtures of simple alkynes, R = R and R' = R', or of simple and mixed alkynes, cyclobutadienes carrying mixed alkyl groups (10-12) can be prepared: l7 R' )qR' R R R' )qR R R' R' )qR R R The e.s.r.hyperfine coupling constants can then be interpreted in terms of the electronic effect of the substituents, as was done for cyclopentadienyl radicals. The order of electron-releasing power by different alkyl groups appears to be different in the two series of compounds; the same sequence would not necessarily be expected, but with the radical ions the situation may be complicated by strong solvation e ~ ~ e c t s . ' ~ In view of the success of reaction (7) for generating cyclopentadiene radical cations, it seemed possible that protic acids might be able to replace the Lewis acid in reaction (8) to provide a new route to cyclobutadiene radical cations, and we find that if a solution of di-t-butylethyne or of di- 1 -adamantylethyne in trifluoroacetic acid is photolysed, strong spectra of the corresponding tetra-t-alkylcyclobutadieneJ. L.COURTNEIDGE, A. G. DAVIES, P. S. GREGORY AND S. N. YAZDI 53 n Fig. 2. E.s.r. spectrum of the tetra- 1 -adamantylcyclobutadiene radical cation (14, R = 1-adamantyl), obtained by photolysis with U.V. light filtered through Pyrex of a solution of di-1-adamantylethyne in trifluoroacetic acid at 8 "C; ca. 20% v/v of dichloromethane was added to promote solubility. radical cations (14) are observed. The spectrum of the tetra- 1 -adamantyl- cyclobutadiene radical cation is shown in fig. 2. The sequence of the various steps of protonation, homolysis and dimerisation has not yet been established, but a number of alkynes have been shown to be converted into cyclobutenyl cations (13) under super-acid conditions," and it seems likely that we are observing again the photohomolysis of a carbenium ion [cf: reaction (S)]: As in reaction (7), we do not yet know the fate of the hydrogen fragment, but again hydrogen is reproducing the behaviour of a metal, in this case aluminium.The radicals prepared by using protic acids [reaction (9)] are much longer lived than those prepared using aluminium chloride [reaction ( 8)].16,17,19 A wide variety of cyclobutenyl cations can be prepared by different methods,20 and it appears likely that their photolysis may provide a route to cyclobutadiene radical cations which have hitherto been inaccessible.54 CONJUGATED CARBANIONS AND CARBENIUM IONS BUTA-1,3-DIENE RADICAL CATIONS When an alkyne is treated with aluminium chloride in dichloromethane, no e.s.r.spectrum is usually observed until the mixture is photolysed. Sometimes, however, a spectrum can be observed before photolysis; for example, di-t-butylethyne may display a strong spectrum showing hyperfine coupling by two methyl groups of one type, and two pairs of another, a(2Me) = 4.2, a(2Me) = 10.55, a(2Me) = 10.70 G. The hyperfine coupling constants correlate with those calculated for the hexamethyl- butadiene radical cation, and indeed this diene generates the same spectrum when it is treated with aluminium chloride2’ or with mercury( 11) trifluoracetate in trifluoroacetic acid (Kochi’s reagent22). An alternative route to the same 1,3-diene radical cation is the photolysis of a solution of the alkyne itself with Kochi’s reagent:23 The detailed mechanism of the reaction involving aluminium chloride is not known, but Olah has shown that di-t-butylethyne rearranges to hexamethylbutadiene under super-acid conditions.18 One possibility therefore is that adventitious hydro- lysis of A1C13 produces a trace of HA1C14, which induces a double nucleophilic rearrangement.CONCLUSION We have shown here that a variety of carbenium ions RH+ (or RM’) can be induced to undergo homolysis to give the corresponding radical cations Re+. This suggests the exciting possibility that the many further types of carbenium ions which have been identified in acid or super-acid solution24 may provide an entree to new families of radical cations. We are pleased to acknowledge the support of the S.E.R.C.I 2 3 4 5 6 7 8 9 10 II 12 K. U. Ingold, B. Maillard and J. C . Walton, J. Chem. SOC., Perkin Trans. 2, 1981, 970; A. G. Davies, D. Griller, K. U. Ingold, D. A. Lindsay and J. C. Walton, J. Chem. SOC., Perkin Trans. 2, 198 1 , 633. A. G. Davies and M-W. Tse, J. Chem. Soc., Chem. Commun., 1978, 353. P. J. Barker, A. G. Davies and M-W. Tse, J. Chem. SOC., Perkin Trans. 2, 1980, 941. A. G. Davies, J. A-A. Hawari, C. Gaffney and P. G. Harrison, J. Chem. SOC., Perkin Trans. 2, 1982, 631. P. B. Brindley, A. G. Davies and J. A-A. Hawari, J. Organomet. Chem., 1983, 250, 247. J. M. Atkinson, P. B. Brindley, A. G. Davies and J. A-A. Hawari, J. Organomet. Chem., 1984,264, 253. P. J. Barker, A. G. Davies, R. Henriquez and J-Y. Nedelec, J. Chem. SOC., Perkin Trans. 2, 1982, 745.A. G. Davies, E. Lusztyk and J. Lusztyk, J. Chem. Soc., Perkin Trans. 2, 1982, 729. A. G. Davies, J. P. Goddard, E. Lusztyk and J. Lusztyk, J. Chem. SOC., Perkin Trans. 2, 1982,737. C. Eaborn, Organosilicon Compounds (Butterworths, London, 1969), p. 125; A. G. Davies, J. Organomet. Chem., 1982, 239, 87. A. G. Davies, E. Lusztyk, J. Lusztyk, V. P. J. Marti, R. J. H. Clark and M. J. Stead, J. Chem. SOC., Perkin Trans. 2, 1983, 669. A. G. Davies, J. R. M. Giles and J. Lusztyk, J. Chem. SOC., Perkin Trans. 2, 1981, 747.J . L. COURTNEIDGE, A. G. DAVIES, P. S. GREGORY AND S. N. YAZDI 55 l 3 D. Wilhelm, J. L. Courtneidge, T. Clark and A. G. Davies, J. Chem. SOC., Chem. Commun., 1984, 810. J. L. Courtneidge, A. G. Davies and S. N. Yazdi, J. Chem. SOC., Chem. Commun., 1984, 570. Q. B. Broxterman, H. Hogeveen and D. M. Kok, Tetrahedron Lett., 1981, 22, 173. J. L. Courtneidge, A. G. Davies, E. Lusztyk and J. Lusztyk, J. Chem. SOC., Perkin Trans. 2, 1984, 155 and unpublished work. 14 l 5 H. Bock, B. Roth and G. Maier, Angew. Chem., Znt. Ed. Engl., 1980, 19, 209. 16 17 '' G. A. Olah and H. Mayr, J. Am. Chem. Soc., 1976, 98, 7333. l 9 Q. B. Broxterman, H. Hogeveen and R. F. Kingma, Tetrahedron Lett., 1984, 25, 2043. 2o G. Olah, J. S. Staral, R. J. Spear and G. Liang, J. Am. Chem. SOC., 1975, 97, 5489. J. L. Courtneidge and A. G. Davies, J. Chem. SOC., Chem. Commun., 1984, 136. 22 W. Lau, J. C. Huffman and J. K. Kochi, J. Am. Chem. SOC., 1982, 104, 5515. 23 W. Chan, unpublished work. 24 Carbonium Ions, ed. G. A. Olah and P. von R. Schleyer ( Wiley-Interscience, New York, 1968-1976), vol. I-v.