On the other hand, the method with symmetryadapted functions always requires more parameters. Its use is mainly interesting when the molecular orientations are not really localized, because the librational anharmonicity is then taken into account. However, two problems may occur in this method which is available only for rigid molecules:(i) When the translational amplitude is very important, as for HMDS, the Bragg reflections are only observable for small values, even if the molecular orientations are localized. It is then very difficult to deduce from the refinements the orientational molecular probability.(ii) When the functions adapted to the molecular symmetry have negligible values for the atoms of the external shells of the molecule, the corresponding A,,m, terms are only related to the internal shells and are not then easily refinable. In this case, occurring in HME, the molecular orientational probability deduced from the refinements is less localized than in reality, mainly at low temperature.However, the comparative use of these two different methods gives a very accurate view of the long-range order in plastic crystals.We would like to thank Professors Fouret and Lefebvre for their interest in this work and Drs Sauvajol, Gharby and Gors for interesting discussions.
AbstractAb initio and CNDO/2 molecular-orbital calculations have been performed for the positional isomer pairs of phenylfuroxan, methylfuroxancarboxamide, isopropyl N-(methylfuroxanyl)carbamate, and chloro(phenyl)furoxan, for the molecules of dimethylfuroxan, diphenylfuroxan, and furazan and for an idealized unsubstituted furoxan molecule. The furoxan ring appears to be electron-overcrowded, in * Furoxan is furazan N-oxide. t To whom correspondence should be addressed. particular in furoxan itself, which has not been synthesized so far. This excess of charge is released toward the substituents in the derivatives. In all derivatives the furoxan group remains the most reactive part of the molecule.