The rate constants were determined from the slopes of the lines and their temperature dependency is shown in Fig. 6.The energies of activation were calculated from the Arrhenius relationship to be 65.6 kcal. mole-' for the decomposition of potassium nitrate and 34.3 kcal. mole-' for the reaction between potassium nitrite and oxygen. The corresponding frequency factors are l O l o J set.-' and 103,58 set.-', respectively.These values are further confirmation of the suggested mechanism involving an initial reaction between nitrite and molecular oxygen, since an activation energy which is approximately equal to the dissociation energy of oxygen would be required if the primary step of the oxidation process involved the dissociation of molecular oxygen.The corresponding activation energies for the deconiposition of sodium nitrate and for the reaction between sodium nitrite and oxygen6 are 44.7 and 20.7 kcal. mole-'. The difference in the energies of activation for the decoiiipositions of sodium and potassium nitrate is apparently larger than could be accounted for by the difference in dissociation energies of the N-0 bond. This however may be attributed to the more extensive polarizing power of the sodium ion. The sodium ion, which induces a relatively greater positive and negative charge in the oxygen molecule a t the surface of the melt could cause oxygen to be more strongly attracted to the nitrite ion and accordingly the steepness of the potential energy profile would be effected. Therefore the height of the barrier is a function of both, the K-0 bond strength and the repulsion between the nitrites and oxygen.Acknowledgment.-The author expresses his gratitude to J. Campisi for carrying out the X-ray analyses and to S. Gordon for reviewing the manuscript.SEWARK 2, SEW JERSEV

No abstract

The unimolecular reactions that give rise to mass spectra are controlled by spatial relationships and energy considerations. Inmolecules that contain a heteroatom, eliminationreactions, involving bond-making as well as breaking, are often prominent, e.g. loss of water from alcohols. The ease of such reactions depends on spatial relationships in the molecule, and the resultant ion intensities in the spectra of geometric isomers can consequently be correlated with differences in geometry and hence furnish a basis for assigning structures. Processes that do not involve bond-making do not have such rigorous geometric requirements, but depend rather on attainment of a transition state defined in terms of a minimum energy content. Common product ions from stereoisomeric hydrocarbons seem often to arise via a common transition state. When this condition holds, the difference between the enthalpies of the isomers is reflected in the relative appearance potentials and-though the cause-and-effect relationship here is less direct and more readily obscured by other factors-relative intensities of the common product derived from the isomers. In both classes of processes, the spectra of stereoisomers can be simplified and made more distinctive by lowering source temperature and ionizing voltage.

Field-ionization mass-spectral analysis of the volatile fractions of cationically initiated polymers of propylene, 1-butene, cis-2-butene, and isobutylene reveal that substantial fractions of these polymers (20 to 80%) do not have the expected polymeric formulas. Specifically, they exhibit carbon numbers other than integral multiples of those of the monomers from which they were prepared. Although products with anomalous carbon numbers have been reported for triptene dimers1 and cis-2-butene oligomers,2"4 our data suggest that formation of such products during cationically initiated polymerization of olefins is more general than has been recognized. Aluminum chloride as a polymerization initiator appears to give more anomalous carbon-number product than boron trifluoride. Chain transfers involving carbon-carbon bond cleavages are responsible for their formation. These reactions appear to be more frequent at higher polymerization temperatures.

No abstract