78 CHEMICAL REACTIONS IN THE GAS PHASE 



nation of relative intensities by the simple measurement of peak heights, 

 also indicates that the production of ions having appreciable kinetic 

 energy is far less frequent here than in diatomic molecules. Very careful 

 measurements (20) have shown that all the members of any one isomeric 

 series of molecules, such as the octanes, show the same total ionization 

 when measured under the same conditions. This can be interpreted in 

 two ways. Either there are essentially no products formed with kinetic 

 energy, or every isomer in the series produces about the same amount of 

 products with kinetic energy. In view of the striking differences in mass 

 spectra of a series like the octanes this second alternative seems far less 

 likely. (See Table 1.) 



When an attempt is made to correlate appearance potentials with 

 thermal data on bond dissociation energies, the agreement is generally 

 fair (21, 22). The discrepancies, as might be expected, are all in a 

 direction which indicates that the products of dissociation have a certain, 

 amount of excess energy of the order of a few tenths of a volt. This 

 energy could conceivably be either vibration-rotation energy or kinetic 

 energy of translation. In view of the previous discussion it seems likely 

 that it is mostly vibrational, and indeed the apparent magnitudes of 

 the excess energy for heavy molecules seem never to be too large to be 

 interpreted in this way. It is frequently necessary in small molecules 

 such as methane to assume such dissociations as (21) 



CH4+ -^ C+ + 4H 



in which the uncharged fragments come off as atoms. However, in a 

 paper by Delfosse and Bleakney (22) on the mass spectra of propane, 

 propylene, and allene it is shown that here no such assumption is 

 necessary for any of the ions measured, which include all the important 

 ions. Every appearance potential measured is best explained by as- 

 suming that the minimum energy process possible occurs. That is, 

 wherever possible, the uncharged fragments come off as molecules. 



Also, in the mass spectra of all large hydrocarbon molecules one finds 

 a number of metastable peaks (23, 24). These metastable peaks con- 

 stitute a large fraction of the diffuse peaks occurring at non-integral 

 masses. These peaks have been shown to be formed by delayed dis- 

 sociations which occur after the original ion has been accelerated but 

 before or immediately after its entry into the analyzer (25, 26). Exami- 

 nation of these peaks in a very large number of hydrocarbons (23, 24) 

 shows that the uncharged fragments of the dissociation almost always 

 are in the correct ratio to form a stable molecule. For most of these 

 molecules, we have no way of knowing the state of association or dis- 

 sociation of the neutral part, but the fact that the fragments are of such 



