J . Chem. SOC., Furaday Trans. I , 1984, 80, 21 1-21 6 Electron Addition to Triphenylmethyl Arsonium Iodide An Electron Spin Resonance Study BY MARTYN C. R. SYMONS* AND GLEN D. G. MCCONNACHIE Department of Chemistry, The University, Leicester LE 1 7RH Received 27th May, 1983 Exposure of crystalline Ph,AsMe+I- crystals to s°Co y-rays at 77 K yields an electron-adduct having e.s.r. parameters characteristic of methyl radicals weakly interacting with a Ph,As mol- ecule. We now find that, on annealing, this species dissociates to give normal methyl radicals, whilst solutions in methanol (CD,OD) give, on electron addition at 77 K, the normal arsoranyl radical [Ph,AsMe] together with methyl radicals, there being no trace of the adduct formed in the crystalline material. Reasons for these differences are discussed and it is suggested that the arsoranyl radical rearranges to a o* species as a necessary step in the dissociation process. Geoffroy and Llinaresl have recently established that electron addition to Ph,AsMe+ ions in triphenylmethyl arsonium iodide crystals at 77 K yields a species having e.s.r.parameters (lH, 13C and g) characteristic of methyl radicals, except that there was also a well defined splitting from a single 7 5 A ~ nucleus. They estimated a spin density of ca. 0.75 on carbon for this adduct. No further species were detected on annealing. They drew attention to the surprising difference between this species and the phosphoranyl radicals 'PL, normally formed by electron addition to similar phosphorus compounds. Our interest in these results stemmed from a recent controversy regarding possible structures for the radical anions of alkyl halides and fluorinated alkyl Discussion centred around the fact that (F,C'-ha1)- anions have a well defined o* structure, but similar alkyl halide anions, (R' -hal)-, have never been prepared.We argued that a major contributory factor must be the tendency for alkyl radicals to prefer a planar structure, whereas ' CF, radicals are pyramidal. Clearly another factor is the electron-withdrawing power of the fluorine atoms, which lowers the energy of the carbon o level thus increasing the covalency of the anion. These factors are illustrated in fig. 1. The force of these arguments is nicely illustrated by results for the isostructural ammonia derivatives, [H,N'-hal].12 These species are o* anions, with considerable spin density on the halogen and with significant 2s contribution from nitrogen indicating pyramidality for the H3N group.Delocalisation onto halogen and 2s character on nitrogen increase on going from C1 to I, thus establishing the expected trend in orbital hybridisation. The aim of the present study was to discover conditions under which the As-Me bond does not break on electron addition and to learn more about the structure of the adduct. This work links with our recent study of the (Br-CN)- radical anion which was shown to exist in a linear, o* structure on initial electron addition, but rather than undergoing dissociation this rearranged to a bent, onb structure14 on annealing above 77 K.15 8 21 1 F A R 1212 E.S.R.STUDY OF Ph,AsMeI + adduct SOMO fi (CH,) C(~P,) 2. E Y g m v M .$ (CH3) c(sp3)- s 2 b) 4 u* anion SOMO c .- Y .- (CF,) C(sp3) .- - Ih Y .- z 41 ir \(I Fig. 1. Bonding scheme for H,C’----hal- ‘adduct’ and for (F,C’-ha1)- CT* radical anion. The vertical arrow shows how the rehybridisation that occurs as the CH, unit flattens aids in weakening the C-ha1 bond. 1 3264 G u u u u -Yz -Y* .‘/2 .3/2 T- Fig. 2. First-derivative X-band e.s.r. spectrum for Ph,AsMe+I- after exposure to s°CO y-rays at 77 K showing features assigned to Ph,As- - - -. CH, adducts. EXPERIMENTAL Triphenylmethylarsonium iodide was obtained from Lancaster Synthesis. The pure compound and its solutions in CD,OD were exposed to s°Co y-rays at 77 K in a Vickrad y-cell with doses up to 1 Mrad.E.s.r. spectra were measured at 77 K with a Varian El09 spectrometer calibrated with a Hewlett-Packard 5246L frequency counter and a Bruker B-H 12E field probe. Samples were annealed by decanting the liquid nitrogen from the insert Dewar and recooling whenever significant spectral changes were observed.M.C.R. SYMONS AND G. D . G. MCCONNACHIE i 3250 G v I I 1 3260 G 100 G - H I I - 3 4 L 1 - I ,H 10 G , 213 Fig. 3. First-derivative X-band e.s.r. spectra for a dilute solution of Ph,AsMe+I- in CD,OD after exposure to s°Co y-rays at 77 K, (a) showing features assigned to ph,AsMe] arsoranyl radicals and (b) showing features assigned to methyl radicals. The centre features are largely due to solvent radicals. 8-2214 E.S.R. STUDY OF Ph,AsMeI RESULTS AND DISCUSSION The e.s.r.spectrum obtained from the pure compound is shown in fig. 2. The data derived therefrom are very close to those obtained by Geoffroy and Llinares from their single-crystal experiments. In marked contrast, solutions in CD,OD gave the spectrum shown in fig. 3. The central region of this spectrum is shown in fig. 3(b). The four features shown in fig. 3(a) are assigned to the arsoranyl radical Ph,AsMe. 7 5 A ~ has I = $!, but because of the large magnitude of the hyperfine coupling the four features corresponding to M , = f $, -k$ are unequally spaced. The features exhibit parallel and perpendicular components with a small x, y splitting for the M , = -f line. The MI = +f line is fortuitously almost isotropic. The results have been analysed using the axially symmetric spin Hamiltonian Table 1.E.s.r. hyperfine parameters hyperfine coupling to 7 5 A ~ , g- tensor Br or lZ7I/Gu componentsa lH hyperfine coupling radical All A , Aiso gll g, constant/G ref Ph,AsMe 592 505 54 1 Ph,As 592.6 505 54 1 Ph,As'-Me 102f2 98+2 99k4 525 524.8 Ph,As' -Me 97 88 90 CH,Br- 58 -28 0.7 86 CH,I- 108 -60 -4 1.97 2.0 14 - this work 1.97 2.014 16 2.0028 2.0017 (11)22.7 this work 2.0010 (1)22.0 2.0021 2.0014 - 2.0009 2.027 2.027 - 1 - 20.6 22 - 20.6 22 - - a Calculated using the Breit-Rabi equation. The resulting tensor components are given in table 1, together with those for the Ph,As ' radical16 for comparison. The four narrow features shown in fig. 3(b) are characteristic of 'free' methyl radicals, with no trace of any residual interaction with the Ph,As molecule.In their single-crystal study Geoffroy and Llinares were unable to detect any break-down products from the [Ph,As- - - - -Me'] adduct. However, using our technique of rapid quenching outlined above we have been able to obtain well defined features for normal methyl radicals on annealing, together with features assigned to H$AsPh, radicals, presumably formed by Ph,AsCH, + * CH, -+ Ph,AsCH, + CH,. (2) ASPECTS OF MECHANISM Irradiation of the pure salt, Ph,AsMe+I-, is expected to give electron-gain and electron-loss centres. The former is clearly the methyl radical addult, there being no trace of either the arsoranyl radical or of the substituted benzene anion derivativesM.C.R. SYMONS AND G. D. G. MCCONNACHIE 215 such as are usually formed from the phosphorus ana10gues.l~ The electron-loss centre is almost certainly I;, with very broad, ill defined features such as are often found in ' loose ' matrices in which extensive libration can occur.In marked contrast, electron addition to Ph,AsMe+ ions in a CD,OD matrix gave no sign of the methyl radical adduct, but low yields of methyl radicals together with relatively high yields of arsoranyl radicals. Ph Ph When the electron comes into the sphere of attraction of the 'tetrahedral' cation it can be stabilised either by bond stretching to give a o* radical, structure (I),l8?l9 or by bond-bending to give an arsoranyl radical, structure (11) or (111). We suggest that structure (I) is a necessary precursor to the formation of methyl radicals, whilst structure (11) or (111) must convert to structure (I) before giving methyl radicals.This suggestion is similar to our postulate for the formation of phenyl radicals and halide ions from ha10genobenzenes.l~ Initial electron addition is partitioned between the n* aromatic orbital and the C-ha1 o* orbital. The latter anion can move smoothly to give products, but the former cannot so a n* + o* excitation is required. For the arsine, a similar excitation is required for structure (11), which is probably the favoured structure. However, for structure (111) there is already extensive delocalisation onto the two axial groups so direct loss of methyl could occur without any electronic excitation. Hence the rate-determining step could well be thermal conversion of structure (11) to structure (111).ASPECTS OF STRUCTURE THE ARSORANYL RADICAL We cannot readily distinguish between structures (I)-(111) from the e.s.r. parameters alone. However, note that the parameters are identical, within experimental error, to those for the Ph,As' radical.16 Since the parameters are known to depend primarily on the nature of the axial ligands in such structures, this suggests that structure (111) should be favoured. This is also expected in terms of the observation that the more electronegative substituents favour the axial sites. The o* structure (I) is unlikely because it is not expected to resemble closely the Ph,As' radical. This is unlikely to have a o* structure, and even if it had it would not be identical with structure (I).However, the most important argument against the o* structure (I) is the one we have used to explain the absence of o* alkyl halide anions.10~13~14~20 If this structure represents a minimum along the reaction coordinate leading to Ph,As: + 'CH,, then given that these units are in close contact in the methyl radical adduct, we would expect 'recombination' to give the o* structure. In other words, we think it is unacceptable to have such a high barrier between the o* state and the methyl radical adduct state that the two species can be quite stable at 77 K with no tendency to interconvert. We therefore conclude that the arsoranyl radical centre detected in CD,OD has structure (111).216 E.S.R. STUDY OF Ph,AsMeI THE METHYL RADICAL ADDUCT It was previously concluded that this centre has ca.75% spin density on 'CH, and, presumably, 25% on the Ph,As group. In our view, however, the species is essentially a pure methyl radical undergoing a very weak charge-transfer interaction with the Ph,As molecule. Data for this adduct and for a range of other adducts are given in table 1. The lH hyperfine coupling is seen to be a sensitive probe for delocalisation, in which case there is almost zero loss of spin density for the Ph,As adduct. Certainly the anisotropic 13C coupling is less than that expected for stationary 'CH, radicals, but that can be understood in terms of libratory motions of these small radicals rather than delocalisation. Indeed, this is normal for 'CH, radicals at 77 K. If the isotropic coupling is used to estimate the spin density on carbon it must again be close to unity.Had there been some 25 % delocalisation, indicative of considerable a-bonding, the CH, unit should have become pyramidal, as for the NH, group in the hale-NH, cenfres.l2 This would lead to a reduction in IA(lH)I and a major increase in A(13C), neither being observed. We conclude that the centre is essentially a methyl radical. In that case the small coupling to 7 5 A ~ represents a slight electron transfer. If the isotropic and anisotropic coupling constants are analysed in the usual way21 we find ca. 0.018 4s character and ca. 0.033 4p character, in reasonable agreement with the model. Similar values were actually estimated by Geoffroy and Llinares, but they apparently favoured the 13C results.Since even this degree of delocalisation is not evidenced by the isotropic lH and 13C coupling constants, it is possible that the 75A~ results are due to a spin-polarisation mechanism rather than representing real delocalisation. M. Geoffroy and A. Llinares, Mol. Phys., 1980, 41, 55. E. D. Sprague and F. Williams, J. Chem. Phys., 1971, 54, 5425. S. P. Mishra and M. C. R. Symons, J . Chem. Soc., Perkin Trans. 2, 1973, 391. Y . Fujita, T. Katsu and K. Takahashi, J. Chem. Phys., 1974, 61, 4307. A. Hasegawa, M. Shiotani and F. Williams, Faraday Discuss. Chem. Soc., 1977,63, 157. D. J. Nelson and M. C. R. Symons, Chem. Phys. Lett., 1977,47,436. M. C . R. Symons, J. Chem. Soc., Chem. Commun., 1977,408. M. C. R. Symons, J. Chem. Res. (S), 1978, 360. J. T. Wang and F. Williams, Chem. Phys. Lett., 1980, 72, 556. lo M. C. R. Symons, Chem. Phys. Lett., 1980, 72, 559. l1 J. T. Wang and F. Williams, J . Am. Chem. Soc., 1980, 102, 2861. l2 M. C. R. Symons, J . Chem. Res. (S), 1981, 160. l3 M. C. R. Symons, Proceedings of the 6th International Conference on Radiation Research, 1979,6,238. l4 M. C. R. Symons, Pure Appl. Chem., 1981,53, 223. l6 M. C. R. Symons and S. P. Mishra, J. Chem. Soc., Furaday Trans. I , 1982, 78, 3019. l6 S. A. Fieldhouse, H. C. Starkie and M. C. R. Symons, Chem. Phys. Lett., 1973, 23, 508. S. P. Mishra and M. C. R. Symons, J. Chem. SOC., Perkin Trans. 2, 1976, 21. l* T. Berclaz, M. Geoffroy and E. A. C. Lucken, Chem. Phys. Lett., 1975, 36, 677. lS M. C. R. Symons, Chem. Phys. Lett., 1976,40, 226. 20 M. C. R. Symons, Radial. Phys. Chem., 1980, 15, 453. 21 M. C. R. Symons, Chemical and Biochemical Aspects of Electron Spin Resonance Spectroscopy (Van 22 M. C. R. Symons and I. G. Smith, J . Chem. Soc., Perkin Trans. 2, 1981, 1180. Nostrand Reinhold, Wokingham, 1978). (PAPER 3/862)