The structures and reactions of radical cations of prototype alkanes such as ethane, propane, isobutane, neopentane, etc. produced in SF6 and in other halocarbon matrices irradiated at 4 K have been extensively studied by ESR. The geometrical structures as well as the unpaired electron orbitals of these cations are unequivocally characterized based on the simple symmetry considerations as well as the INDO MO calculations. The ESR hyperfine coupling constants determined for these cations as well as their deuterated homologs give conclusive evidence for the interpretations. The unpaired electron in linear alkane cations is delocalized over the in-plane C–H and C–C σ bonds forming delocalized σ radicals, whereas that in highly branched alkane cations is more confined to one of the C–C bonds forming localized σ radicals. The H1s spin density, giving a large hyperfine coupling in the linear alkane cations, is largely contributed from a direct participation of the C–H bond in the unpaired electron orbital, whereas that in the highly branched alkane cations is due to mainly hyperconjugation. The alkane cations undergo two types of reactions: one is deprotonation to form neutral alkyl radicals in SF6 and in CFCl2CF2Cl and the other is the H2 and CH4 elimination to form olefinic π cations in CFCl3. Based on the matrix effects on the structure of propane and isobutane cations, the site preference in forming alkyl radicals, and other observations, it is concluded that the alkane cations undergo reactions through the breakage of the bond in which the unpaired electron is highly populated.

E.s.r. evidence has been obtained that cyclopropane cations exhibit a static distortion from Dgh to C2" symmetry in a matrix at 4.2 K; the 3e' (D3h) highest occupied degenerate orbital splits into 6a, and 3bl (C2") orbitals and the unpaired electron occupies 6a1, giving dominant spin densities on the two basal carbon atoms. Recentlywe have found that C2H6+, a Jahn-Teller active molecule, trapped in SF, shows a reversible temperature t Preliminary results were presented at the 46th Autumn Annual Meeting of the Chemical Society of Japan, 3rd October, 1982, Niigata, Abstract p. 13; details were presented at the 21st E.s.r. Symposium, 26th October, 1982, Tsukuba, Abstract p. 48.change which may be ascribable to static and dynamic Jahn-Teller effects? Cyclopropane (cyclo-C3H6) with Dab symmetry possesses a degenerate highest occupied orbital (3e') so that cyclo-C,H,+ is also a Jahn-Teller active molecule. Recently Shida and CO-WOrkerS2 have observed the e.s.r. spectrum Of cyc10-C3H6f produced in CFC13 at 77 and have explained the extremely small hyperfine couplings giving an

In order to elucidate the structure of methacrylic acid radicals, the change with observation temperature of the ESR spectrum of free radicals trapped in solid methacrylic acid γ‐irradiated at −196°C was studied. Below −80°C, we found a 9‐line spectrum, which is similar to the ordinary 9‐line spectrum observed in irradiated poly(methacrylic acid) or poly(methyl methacrylate), but which differs in the stronger intensity of the so‐called 4‐line component. Our 9‐line spectrum changes reversibly into a 13‐line spectrum above −80°C. With broad‐line NMR measurements of methacrylic acid, it was found that there is such an unusual crystalline transition around −30°C that the line width is narrower in the lower‐temperature region (phase II) than that in the higher‐temperature region (phase I). The change of the ESR spectrum can be interpreted in terms of the exchange of the two β‐protons due to the hindered oscillation around the CαCβ bond of the single radical ···CβH2Cα(CH3)COOH if one assumes the gradual change of the hindering potential barrier caused by the crystalline transition and the lower barrier in phase II. The modified Bloch treatment gave the hindering potential barrier to be 7.2 kcal/mole in phase I and 1.5 kcal/mole in phase II. The difference between our 9‐line spectrum and the ordinary one with the very weak 4‐line component comes from the difference of the surrounding matrix.

An adaptation of the method of least squares to the refinement of molecular structures, based on gaseous electron-diffraction sector-microphotometer intensity data, is presented. The method has the advantages of permitting simultaneous refinement of many more parameters than can be handled by trial-and-error methods and of opening the possibility of obtaining realistic standard errors and error correlations in place of the more subjective 'error limits'. Although the problem of weighting of the observations is not solved, it is felt that the method provides a more straightforward way to take account of factors known to introduce uncertainties than does the alternative approach based on radial distribution functions.A brief account of general experience with the method is given.

No abstract