Famday Discuss. Chem. SOC., 1984, 78, 57-69 Spin Delocalization in Radical Cations of Oxygen-containing Organic Compounds as Revealed by Long-range Hyperfine Interactions and Solvent Effects BY LARRY D. SNOW AND FFRANCON WILLIAMS* Department of Chemistry, University of Tennessee, Knoxville, Tennessee 37996- 1600, U.S.A. Received 10th May, 1984 E.s.r. evidence is presented for long-range ' H hyperfine interactions and thermally revers- ible solvent effects involving mainly the radical cations of aldehydes and ketones in Freon matrices. Long-range couplings to S protons are found to be surprisingly large in the cations of acyclic, cyclic and polycyclic carbonyl compounds. The conformational requirements for these large couplings are discussed in terms of facile spin transmission via a trans arrangement of C-C and C--H u bonds, the assumption of this trans effect providing a rigorous and consistent analysis of the experimental results for both non-rigid and rigid molecular geometries.The matrix perturbation observed at low temperatures is believed to arise from a weak complex formed between the radical cation and the solvent, the hyperfine interaction occurring mainly with a single chlorine nucleus. The reversible loss of the e.s.r. substructure associated with this complex at elevated temperatures is attributed to motional averaging of the chlorine hyperfine anisotropy rather than to dissociation of the complex. ~ During the past few years solid-state e.s.r. studies have provided a wealth of information regarding the structure of radical cations derived from simple organic molecules.This broad advance was triggered by the discovery of a simple method for the generation of these cations by y-irrradiation of dilute solutions of the parent molecules in Freon matrices,14 these solvents functioning as efficient positive hole carriers to the solute by virtue of their high ionization potentials. The other desirable properties of Freons for this application are their high electron attachment rates and their ability to provide a chemically inert environment for the highly reactive solute radical cations. Also of great importance is the fact that the e.s.r. spectra of the cations are frequently well resolved in these matrices. Using this technique, the radical cations of several oxygen-containing organic compounds have been charac- terized, including those of simple ethers,576 a c e t a l ~ , ' ~ ~ aldehydes,"*' ketones,I2 esters of carboxylic a ~ i d s , ' ~ , ' ~ epoxide~'~-'' and vinyl monomers.'* Except where molecular rearrangements are i n ~ o l v e d ~ .' ~ , ' ~ the aforesaid cations are either 7 ~ ~ - ~ or a10-14 oxygen-centred radicals showing appreciable hyperfine couplings to p hydrogens as a result of efficient hyperconjugation. Another manifes- tation of the tendency for some of these radical cations to undergo spin delocalization comes from the e.s.r. observation of significant matrix hyperfine interactions. In the case of the methyl formate ~ a t i o n ' ~ " ~ this solvent interaction is very strong and has been interpreted in terms of the formation of a (T* radical, the hole being shared with a filled chlorine lone-pair orbital in a Freon molecule.Here we are concerned with 'weak' hyperfine interactions as evidenced by thermally reversible solvent effects and coupling to remote ( y and 6) hydrogens, with examples of such effects drawn primarily from radical cations of carbonyl 5758 HYPERFINE INTERACTIONS AND SOLVENT EFFECTS compounds. Our preliminary reports' ',I2 have already commented on the surpris- ingly large magnitude of the hyperfine couplings which are involved, given the secondary character of these effects. Here we present additional results that fully confirm the generality of these interactions, particularly in regard to long-range proton hyperfine couplings which are interpreted according to previous empirical and theoretical generalization^.'^-*' Such experimental studies should eventually contribute to a greater understanding of the mechanisms of orbital interactions through bonds and through space.'2923 EXPERIMENTAL Dilute solutions (0.1-1 mol%) of the solutes in various Freons were prepared in Spectrosil or Suprasil e.s.r.tubes (3 mm i d . ) on a vacuum line. Acetaldehyde, propionaldehyde, cyclo- hexanone, adamantan-2-one and bicyclo[3.3. I lnonan-9-one were obtained from the Aldrich Chemical Co. The various deuterated compounds were supplied by Merck, Sharp and Dohme Isotopes and purified by trap-to-trap distillation on the vacuum line immediately before use. The Freon solvents included trichlorofluoromethane and 1 , 1 , 1 -trichlorotrifluoroethane, both from Aldrich.The sample tubes were irradiated at 77 K by exposure to y-rays from a 6oCo source (Gammacell 200, Atomic Energy of Canada Ltd) for doses up to 0.5 Mrad (5 x lo3 J kg-'). Subsequently, the sample tubes were transferred to a variable-temperature Dewar insert mounted in the cavity of an e.s.r. spectrometer (Bruker ER 200 D SRC), the ESR-900 cryostat (Oxford Instruments) being used to obtain temperatures below 77 K. Spectra were recorded over a wide range of temperature as the sample was progressively annealed up to the softening point of the particular matrix. Additional e.s.r. spectra were recorded during the course of the work with a Varian V-4502- I5 instrument. The microwave frequency was measured with a Systron-Donner counter (model 6054 B) and magnetic-field markers were recorded with an n.m.r.gaussmeter (Walker model G-502 or Bruker microprocessor-controlled model ER 035 M). RESULTS ACETALDEHYDE A N D PROPIONALDEHYDE RADICAL CATIONS We have previously reported e.s.r. studies on the acetaldehyde cations CH,CHO', CH3CDO', CD3CHO' and CD,CDO'.'' These isotopic experiments proved that there is no resolved hyperfine coupling to the hydrogens of the methyl group in the acetaldehyde cation, the interactions being confined to the aldehydic hydrogen and the Freon matrix. As shown in fig. 1, the spectrum of CD,CHO' in CFC13 at 50 K possesses a large doublet splitting of 136 G together with a highly anisotropic substructure arising from the matrix interaction. Although this substructure is lost reversibly at ca.120 K, the large coupling to the aldehydic hydrogen remains virtually unaltered.' ' It was therefore of interest to carry out a similar e.s.r. study of the propional- dehyde radical cation, and the salient results are presented in fig. 2 and 3. In contrast to the single reversible change observed for the acetaldehyde cation, the e.s.r. spectrum of the propionaldehyde cation in CFC1, undergoes a series of reversible changes between 100 and 160 K. As shown in fig. 2( c), the spectrum of CH3CH2CHO' at 40 K is distorted by anisotropic features which are similar to those observed in the low-temperature spectra of the acetaldehyde cations and previously assigned to a chlorine interaction.' ' Also paralleling the behaviour of the acetaldehyde cations is the reversible disappearance of these extraneous peaks on warming to 120 K, as indicated by comparison of the CH3CH2CHO+ spectra inL.D. SNOW AND F. WILLIAMS 59 9433.4 MHz Fig. 1. First-derivative e.s.r. spectrum of the CD3CH0 radical cation recorded at 50 K. The cation was generated by y-irradiation of a 0.6mol0/0 solution of the parent compound in CFC13 at 77 K for a dose of 0.5 Mrad. fig. 2 ( a ) and ( c ) . Instead of changing to a simple doublet, however, the spectrum of CH3CH2CHO+ at 120 K becomes a doublet of doublets. That the smaller doublet splitting arises from a methyl rather than a methylene hydrogen is established by the identical spectrum of CH3CD2CHO+ in fig. 2(b). Further changes are observed in the e.s.r. spectra of CH3CH2CHOf and CH3CD2CHO+ at higher temperatures, as shown in fig.3. At 140K the smaller doublet splitting of 12.5 G is no longer resolved [fig. 3(c)], but additional structure becomes evident just before the softening point of the CFC13 matrix (ca. 160 K) is reached. Although not fully resolved, the spectra of CH,CD2CHO+ [fig. 3(a)] and CH3CH2CHO+ [fig. 3( b ) ] at this limiting temperature reveal that a quartet substruc- ture [A(3H) =: 4.6 GI has replaced the sharp 12.5 G doublet in the corresponding spectra at 120 K (fig. 2). Since this change is reversible it can be attributed to the onset of methyl-group rotation about the C(2)-C(3) bond. Thus deuterium labelling and studies of motional effects show that coupling to the single non-aldehydic hydrogen in the 120 K spectrum of the propionaldehyde cation can definitely be assigned to a methyl hydrogen.This situation in which a methyl group adopts a preferred conformation at low temperature such that only one of its hydrogens is strongly coupled to the electron spin has also been observed for the isobutyraldehyde and 2,4-dimethylpentan-3-one radical cations.12 Here, non-aldehydic hyperfine couplings to two and four equivalent 6 hydrogens, respec- tively, were detected, consistent with the presence of two locked methyl groups in each isopropyl group. At this point we digress from the main theme to note that the high-temperature e.s.r. spectra from CH3CD2CH0 and CH3CH2CH0 solutions in fig. 3 contain fairly strong signals in the centre which are readily identified as coming from the radicals CH3CDCH0 and CH3CHCH0 or their conjugate acids formed by protonation of the carbonyl group.The hyperfine pattern for CH,CDCHO (or its conjugate acid) indicated by the stick diagram in fig. 3(a) is particularly clear, the couplings A(3H) = 22.1 G and A( D) = 3.2 G being characteristic of this type of carbon-centred radical. If these radicals originate by the loss (or transfer) of a proton from the60 HYPERFINE INTERACTIONS AND SOLVENT EFFECTS T=120K T = 4 0 K 9 2 7 3 . 5 MHz I I 9 4 2 8 . 8 MHz Fig. 2. First-derivative e.s.r. spectra of the’ radical cations of CH,CH2CH0 [ ( a ) and (c)] and CH3CD2CH0 ( b ) at the temperatures shown. The spectra were recorded after exposure of CFC13 solutions of the parent compounds (0.5 mol %) to y-irradiation at 77 K for a dose of 0.5 Mrad in each case.L.D. SNOW AND F. WILLIAMS 61 L I I 9273.3 MHz r;l T=140K i v 9272.9 MHz 0 v) c) d c) 0 r- Fig. 3. First-derivative e.s.r. spectra of the radical cations of CH3CD2CH0 ( a ) and CH,CH,CHO [( b) and (c)] recorded at higher temperatures than in fig. 2. The central regions of the spectra ( a ) and ( b ) also include signals from the radicals CH,CDCHO and CH,CHCHO, respectively (see text).62 HYPERFINE INTERACTIONS AND SOLVENT EFFECTS Hz I Fig. 4. First-derivative e.s.r. spectra of the ( a ) cyclohexanone and ( b ) [2,2,6,6-2H,]cyclo- hexanone radical cations at 130K. In each case the cation was generated by y-irradiation at 77 K of a 1 mol% solution of the parent compound in CFC13 for a dose of 0.5 Mrad. corresponding radical cation, the results imply that deprotonation occurs mainly from the methylene group.However, kinetic proof of this step is lacking. CYCLOHEXANONE RADICAL CATION Because of their greater conformational rigidity, ring structures are intrinsically more suitable than acyclic compounds for the detailed exploration of angular- dependent long-range couplings.'' As a first step cyclohexanone was used to test the above finding that the strongly coupled non-aldehydic hydrogens in carbonyl radical cations reside in 6 rather than y positions. Accordingly, experiments were carried out on 2,2,6,6-tetradeuterocyclohexanone as well as the fully protiated compound, and the e.s.r. results are shown in fig. 4. The spectra of the two radical cations are seen to have identical 1 : 2 : 1 triplet patterns except for slightly narrower linewidths from the partially deuterated species, verifying that the two strongly coupled hydrogens [A(2H) = 27.5 GI are not in the y positions.Although these results do not exclude the possibility that the two E hydrogens are responsible for the hyperfine interaction, the experiments described in the following section make this assignment very unlikely, as do the empirical rules mentioned later in the Discussion. Therefore, by elimination, the strongly coupled hydrogens must be those in either the two equatorial or the two axial 6 positions.L. D. SNOW AND F. WILLIAMS 63 jli Fig. 5. First-derivative e.s.r. spectrum of the adamantan-2-one radical cation at 145 K. The cation was generated by y-irradiation of a 0.1 mol% solution of the parent compound in CFCl, at 77 K for a dose of 0.4 Mrad.9267.5 M H z (3 (3 I \ I r"' P Q, (D €9 €9 Fig. 6. First-derivative e.s.r. spectrum of the bicyclo[3.3. I]-nonan-9-one radical cation at 144 K. The cation was generated by y-irradiation of a 0.1 mol% solution of the parent compound in CFC1, at 77 K for a dose of 0.4 Mrad.64 HYPERFINE INTERACTIONS AND SOLVENT EFFECTS Table 1. E.s.r. parameters for radical cations of various aldehydes and ketones in the CFCI, matrix radical cation T / K g hyperfine splitting/G CH,CHO+ CH,CDO+ CD3CHO+ CD3CHO+ CH3CH2CHO+ CH3CH2CHO+ CH3CD2CHO+ ( CH3)2CHCHO+ 2,4-dimethylpentan-3-one+ cyclohexanone+ [2,2,6,6-2H,]cyclohexanonef adamantan-2-onet bicyclo[3.3. 1]nonan-9-onef I58 140 50 88 40 120 158 120 159 120 88 88 88 145 144 2.0048 2.0054 CQ.2.0050 ca. 2.0050 ca. 2.0050 2.0048 2.0048 2.0048 2.0048 2.0044 2.0033 2.0032 2.0032 2.0025 2.0025 1363 1 HP) 20.7( 1 DP) 136(1H,), 19(135C1) 135(1H,), 12.5(1H8), 15( 135C1) 1354 IH,), 1 2 3 1H8) 135.1(1Hp), 4.4(3H8) 135.3(1HP), 12.5(1Hg) 135.3( lHP), 4.8(3H8) 120.3( lH,), 20.4(2H8) 15.2(4H8) 27.5(2H8) 27.5(2H8) 22.9(4H8),7.2(2H,) 1 3 6 ( i ~ , ) , 2 1 ( 1 ~ ~ c 1 ) 22.3(4H8),6.9(2HY) RADICAL CATIONS OF ADAMANTAN-2-ONE AND BICYCLO[3.3.1]NONAN-9-ONE In addition to having rigid polycyclic character, these subject compounds possess C,, symmetry. Consequently, they offer an excellent opportunity for probing the couplings to remote hydrogens in a cyclohexanone-like system. In fact, the e.s.r. spectra of their radical cations in fig.5 and 6 possess very similar hyperfine patterns and each can be analysed into a 1 : 4: 6: 4: 1 quintet of 1 : 2: I triplets, the only difference being that the spectrum of the more rigid adamantan-2-one radical cation (fig. 5 ) has better resolution. Also, the e.s.r. parameters of the two cations, summar- ized together with data for other radical cations in table 1, are almost identical allowing for the appreciable error associated with the analysis of broad-line patterns. Given the nearly identical e.s.r. results for these two cations, the assignment problem is greatly simplified. Using the individual cyclohexanone-like rings to define the axial (a) and equatorial (e) positions, each cation has sets of 2HYe, 4Hse, 4Hs, and 2H,,. However, the B2 symmetry of the SOMO allows us to eliminate 2H,, from consideration because these hydrogens lie in the nodal plane of this orbital, i.e.the mirror plane perpendicular to the R=C=O plane (see Discussion). The triplet splitting of ca. 7 G can therefore be assigned to 2H,, while the ca. 22 G quintet splitting must be due to either 4H6, or 4Hs,. This is consistent with the analysis of the results for the cyclohexanone cation, the 27.5 G triplet splitting being assigned to the 2Hs, or 2Hs,. R DISCUSSION REVERSIBLE MATRIX EFFECTS The present study provides further evidence for thermally reversible solvent effects on the e.s.r. spectra of certain aldehyde radical cations in Freon matrices, these effects being characterized by the development of asymmetric substructures in the low-temperature spectra.Although these anisotropic features are difficult toL. D. SNOW AND F. WILLIAMS 65 analyse in detail,' ' there is general agreement that the quartet-like substructure in the doublet e.s.r. spectrum of the acetaldehyde cation (fig. 1) arises mainly from an anisotropic hyperfine coupling to a single chlorine nucleus.' This coupling shows a small but significant negative temperature dependence, All ( 35Cl) decreasing from 21.4 G at 50 K to 18.7 G at 88 K." For the propionaldehyde cation, a tentative analysis of the spectrum in fig. 2( c) gives an estimate of ca. 15 G for the corresponding C1 coupling at 40 K, indicating that the interaction is weaker than for CH3CH0. It has been suggested that the magnitude of the matrix interaction should be a function of the difference between the ionization potentials of the solute and solvent molecules, a smaller difference giving rise to a stronger intera~tion.'~"~ Our results are in accord with this hypothesis, the ionization potentials of CFCl,, CH3CH0 and CH,CH2CH0 being 11.78," 10.2624" and 9.85 eV,24" respectively.Also, the fact that no matrix interaction can be discerned in the e.s.r. spectrum of the isobutyraldehyde cation between 82 and 120 K (table 1) is consistent with the even lower solute ionization potential (9.74 eVZ4') in this case. Another significant feature of the results is that for both CH3CHO' and CH3CH2CHO+ the reversible loss of the chlorine hyperfine structure is accompanied by virtually no change in the large ( 135-1 36 G) isotropic H coupling to the aldehydic hydrogen.This would be a most remarkable finding if the collapse of the 35Cl hyperfine interaction were to result from the thermal dissociation of a weak (T bond between the cation and a CFC1, m01ecule.'~~~'~ The formation of (T* adduct radicals of this type is well known in y-irradiated solids,25 but in our experience the dissociation of such adducts is always accompanied by an alteration in the spin- density distribution within the radical part. It might also be expected that bond scission would lead to sudden changes in the spectrum, but this behaviour is not observed. Indeed, a careful examination of the propionaldehyde cation spectra in fig. 2 reveals that weak anisotropic features are still present at 120 K, albeit with loss of fine structure, suggesting a gradual rather than a sudden loss of the chlorine hyperfine structure with increasing tem- perature.The same conclusion can be drawn from a study of the published e.s.r. spectra for various deuterium-labelled acetaldehyde cations.' ' Alternatively, the reversible temperature effect can be attributed to a modulation of highly anisotropic couplings brought about by the intermolecular motion. Assum- ing the isotropic component of the 35Cl hyperfine interaction is small, this would explain the absence of resolved structure at high temperature. The marked insensitiv- ity of the isotropic aldehydic hydrogen coupling to the loss of 35Cl hyperfine structure is also understandable in terms of this description, since the actual solvent interaction would probably be unaffected by motional considerations.Thus it seems more reasonable to us that the reversible loss of the substructure with temperature results not from a dissociation of the cation-solvent complex but from a motional averaging of the chlorine hyperfine anisotropy associated with this complex.' ' Irrespective of the mechanism responsible for the loss of the solvent hyperfine structure at high temperature, it is generally agreed that the aldehyde-cation-CFC1, interaction is a relatively weak e f f e ~ t , ' ~ ~ ~ ~ ' ~ ' ~ especially in comparison with that observed for the methyl formate ~ a t i o n . ' ~ " ~ Accordingly, we previously described this weak solvent perturbation as a 'superhyperfine' interaction. " We feel that this description is an appropriate one because the e.s.r.spectra of some other cations are also observed to possess hyperfine splittings which do not originate from interaction with the magnetic nuclei of the cation. For example, the quintet e.s.r. spectrum of the ring-opened cation CH20CHl from ethylene oxide shows a curious fine structure in CFC13 which persists up to 150 K, whereas a different substructure 3566 HYPERFINE INTERACTIONS AND SOLVENT EFFECTS is observed in CF3CC13 below 140 K.16 However, the e.s.r. parameters of the main quintet pattern remain sensibly constant in the two matrices.I6 Evidently the effect is highly specific for the cation-matrix-temperature combination, but the structure of the cation does not seem to be significantly affected by these weak solvent perturbations.Despite the weak nature of the solvent interactions described here, it is possible that such complexes may play a significant role in radical cation chemistry.26 As mentioned above, e.s.r. studies show that the radical cation formed from ethylene oxide in the solid state is the ring-opened (2A2)2-oxatrimethylene cation with a symmetrical ( C2v) planar structure similar to that of the isoelectronic ally1 radical. Since a large symmetry-imposed barrier of 25-30 kcal mol-' was originally predic- ted27 for ring opening from the 2B1 ground state of the isolated ethylene oxide it seemed reasonable to consider the alternative of a symmetry-allowed reaction pathway from a complex between the Freon solvent and the 2A1 excited state of the cation.However, recent higher-level calculations by Clark2* indicate that the barrier to the 'forbidden' ring opening is only 3.7 kcal mol-I and may disappear completely at even higher levels. Therefore there is no compelling need to invoke a solvent effect to explain the formation of the ring-opened cation. LONG-RANGE PROTON HYPERFINE INTERACTIONS &HYDROGEN COUPLINGS. A common feature of our results is the detection of a surprisingly large hyperfine coupling to one or more 6 hydrogens in carbonyl radical cations at low temperatures. Considering the acyclic cations first, this long-range coupling is observed to one, two and four equivalent hydrogens in the cations of propionaldehyde, isobutyraldehyde and 2,4-dimethylpentan-3-one, respectively. The results clearly indicate that each of the strongly coupled hydrogens is associated with a stereospecific position in a locked methyl group, and this has been confirmed by studies of the temperature dependence of the e.s.r.spectra. Thus we showed previously that the well resolved quintet spectrum of the 2,4-dimethylpentan-3-one cation is observed only at low temperatures (ca. 100 K),l2 as expected when the equivalence of the S hydrogens requires a preferred conformation involving all four methyl groups. Similarly, the present study of the propionaldehyde cation establishes that the loss of the 6-hydrogen interaction is clearly accompanied by the onset of methyl group rotation. In accordance with many previous empirical and theoretical studies of long-range 'H coupling^,'^-^^ we adopt the simple rule that the 6-proton coupling in these cations is maximised when the interaction can be relayed from the major site of spin density at the carbonyl group uia a trans (W-plan) arrangement of C-C and C-H (T Specifically, the preferred conformations of the methyl groups in the acyclic cations are chosen such that the dihedral angle C,C,C,H, between the C,C,C, and CyCSHG planes is always 180".Although the adoption of this rule leads to a simple and self-consistent explanation of the results, it does not allow us to predict the overall conformation of the acyclic cations. For example, two possible conformations (1) and (2) of the propionaldehyde cation are shown below, these being the cis and trans rotamers corresponding to values of 0 and 180" for the dihedral angle OCBC,C, between the OC,C, and C,C,C, planes.In fact, the requirement that one of the S hydrogens in the methyl group occupies the trans position relative to the C,-C, bond places no restriction on the value of the dihedral angle obtained by rotation of this bond, at least when only one methyl group is present in the 6 position as for the propionaldehyde cation.L. D. SNOW AND F. WILLIAMS 67 In contrast to the propionaldehyde cation, the C,, symmetry of the 2,4-dimethyl- pentan-3-one cation implied by the equivalence of the four 6 hydrogens allows only two possible conformations to be achieved by rotation about the C,-C. bonds. These correspond to OC,CyCs dihedral angles of either *60 or f 120" for the two 6 methyls in each isopropyl group, the y hydrogen residing in the trans or cis position, respectively, with respect to the oxygen of the carbonyl group. The skeletal form of the latter conformation is very similar to that of cyclohexanone (3), especially as regards the OC,C,C, dihedral angle.leads to the assignment of the two strongly coupled S hydrogens in the cyclohexanone cation to those in the equatorial positions, as indicated in structure (3). The coupling of 27.5 G to these hydrogens is the largest we have observed for &hydrogen splittings (table 1) and almost certainly reflects the perfect trans arrangement associated with the chair conformation of the cyclo- hexanone ring, the dihedral angles C,C,C,H,, being close to 180". The use of the trans Similarly, the four strongly coupled S hydrogens in the cations of adamantan-2- one (4) and bicyclo[3.3.l]nonan-9-one ( 5 ) are assigned to those in the equatorial positions as defined by the individual cyclohexanone-like rings of these polycyclic compounds.The smaller Hs, coupling of 22.5G in these polycyclic cations is probably attributable to the fact that four protons are now involved instead of two as for the cyclohexanone cation, leading to an overall increase in the delocalisation of the spin from the carbonyl group. The same type of effect is well known in alkyl radicals where the &proton hyperfine splitting decreases upon progressive methyl substitution at the cy carbon for the series of ethyl, isopropyl and t-butyl radicals.2968 HYPERFINE INTERACTIONS AND SOLVENT EFFECTS The B2 SOMO of the adamantan-2-one cation3’ (6) illustrates how the spin is delocalized into the four equatorial 6 hydrogens, the mechanism involving ‘homohyperconjugation’ from u spin density induced into the C,-C, bond at the two C, positions.Clearly, the trans or alignment effect should be important for this kind of spin transmission, just as it is with the cos20 rule for the more common 7~ hyperconjugation. Y-HYDROGEN COUPLINGS. It remains to consider why resolved y-hydrogen split- tings of ca. 7 G are apparently observed for the polycyclic cations but not for the acyclic or cyclohexanone cations. Small spin populations at the y hydrogens are consistent with the symmetry of the SOMO in (6), so we suggest that the absence of hydrogen splittings in the cyclohexanone and acyclic cations is related to the lesser rigidity of the y hydrogens in these latter systems.Even for the fairly rigid cyclohexanone ring, the carbonyl-containing region of the ring is apparently more flexible than cyclohexane3’ so that the two equatorial y hydrogens can flex out of the OC,C, plane. This is impossible in the polycyclic molecules where a perfect symmetry of the two y hydrogens is maintained.L. D. SNOW AND F. WILLIAMS 69 We acknowledge helpful discussions and useful correspondence with Professors T. Clark, E. Haselbach, S. F. Nelsen, L. Radom and M. D. Sevilla. We also thank Prof. Nelsen for his continuing interest which has led to ongoing collaborative work on various aspects of the problems discussed here. This research was supported by the Division of Chemical Sciences, Office of Basic Energy Sciences, U.S.Department of Energy. ( a ) T. Shida and T. Kato, Chem. Phys. Lett., 1979, 68, 106; ( b ) T. Shida, Y. Egawa, H. Kubodera and T. Kato, J. Chem. Phys., 1980, 73, 5963. M. C. R. Symons and I. G. Smith, J. Chem. Res. ( S ) , 1979, 382. J. T. Wang and F. Williams, J. Phys. Chem., 1980, 84, 3156; see also A. Grimison and G.A. Simpson, J. Phys. Chem., 1968, 72, 1176. K. Toriyama, K. Nunome and M. Iwasaki, J. Phys. Chem., 1981,85, 2149; J. Chem. Phys., 1982, 77, 5891. H. Kubodera, T. Shida and K. Shimokoshi, J. Phys. Chem., 1981, 85, 2583. J. T. Wang and F. Williams, J. Am. Chem. SOC., 1981, 103, 6994. L. D. Snow, J. T. Wang and F. Williams, J. Am. Chem. SOC., 1982, 104, 2062. M. C. R. Symons and B.W. Wren, J. Chem. SOC., Chem. Commun., 1982, 817. K. Ushida and T. Shida, J. Am. Chem. SOC., 1982, 104, 7332. l o ( a ) M. C. R. Symons and P. J. Boon, Chem. Phys. Lett., 1982, 89, 516; ( b ) Chem. Phys. Lett., 1983, 100, 203. I ' L. D. Snow and F. Williams, Chem. Phys. Lett., 1983, 100, 198. L. D. Snow and F. Williams, J. Chem. Soc., Chem. Commun., 1983, 1090. I 3 ( a ) D. Becker, K. Plante and M. D. Sevilla, J. Phys. Chem., 1983, 87, 1648; ( 6 ) M. D. Sevilla, D. Becker, C . L. Sevilla and S. Swartz, J. Phys. Chem., 1984, 88, 1701. G. W. Eastland, D. N. R. Rao, J. Rideout, M. C. R. Symons and A. Hasegawa, J. Chem. Res. ( S ) , 1983, 258. 12 14 l 5 M. C . R. Symons and B. W. Wren, Tetrahedron Lett., 1983, 24, 2315. l 6 L. D. Snow, J.T. Wang and F. Williams, Chem. Phys. Lett., 1983, 100, 193. I n M. Tabata and A. Lund, 2. Naturforsch., Teil A , 1983, 38, 687. l 9 ( a ) G. A. Russell, in Radical Ions, ed. E. T. Kaiser and L. Kevan (Wiley-Interscience, New York, 1968), p. 87; ( b ) F. W. King, Chem. Rev., 1976, 76, 157. 2o ( a ) Y. Ellinger, A. Rassat, R. Subra and G. Berthier, J. Am. Chem. SOC., 1973,95,2373; J. Chem. Phys., 1975, 62, 1; (b) Y. Ellinger, R. Subra, B. Levy, P. Millie and G. Berthier, J. Chem. Phys., 1975, 62, 10. K. U. Ingold and J. C. Walton, J. Am. Chem. SOC., 1982, 104, 616. Hoffmann, Acc. Chem. Res., 1971, 4, 1. (a) K. Kimura, S. Katsumata, Y. Achiba, T. Yamazaki and S. Iwata, Handbook of He ZPhotoelectron Spectra of Fundamental Organic Molecules (Japan Scientific Societies Press, Tokyo, 198 l), pp. 144-146; ( b ) K. Watanabe, T. Nakayama and J. Mottl, J. Quant. Spectrosc. Radiat. Transfer, 1962, 2, 369. 25 ( a ) F. Williams and E. D. Sprague, Acc. Chem. Res., 1982, 15, 408; ( b ) E. D. Sprague and F. Williams, J. Chem. Phys., 1971, 54, 5425. 26 c f , P. S. Skell and J. C. Day, Acc. Chem. Res., 1978, 11, 381. 27 W. J. Bouma, J. K. MacLeod and L. Radom, J. Am. Chem. SOC., 1979, 101, 5540; see also W. J. Bouma, D. Poppinger, S. Saebo, J. K. MacLeod and L. Radom, Chem. Phys. Left., 1984,104, 198. T. Clark, J. Chem. SOC., Chem. Commun., 1984, 666. R. W. Fessenden and R. H. Schuler, J. Chem. Phys., 1963,39, 2147. T. Bally, S. Nitsche and E. Haselbach, Helu. Chim. Acta, 1984, 67, 86. 17 22 ( a ) R. Hoffmann, A. Imamura, and W.J. Hehre, J. Am. Chem. SOC., 1968, 90, 1499; ( b ) R. 23 M. N. Paddon-Row, Acc. Chem. Res., 1982, 15, 245. 24 28 29 30 B. Kovac and L. Klasinc, Croat. Chem. Acta, 1978, 51, 5 5 . 3' R. J. Abraham, D. J. Chadwick and F. Sancassan, Tetrahedron, 1981, 37, 1081.