POLYATOMIC MOLECULES 83 



give rise to rapid shifting of the electrons from one bond to another and 

 also rapid interconversion of electronic and vibrational energy. Some 

 of these states will be highly repulsive and cause the immediate dis- 

 sociation of the ion with no opportunity for "wandering" of the elec- 

 tronic energy and with the production of fragment ions having appreci- 

 able kinetic energy. As has been mentioned, the evidence supports the 

 conclusion that relatively few ions are produced with high kinetic 

 energy. We conclude that the majority of the tremendous number of 

 possible electronic states to which transitions are probable are at most 

 "weakly repulsive." That is, although the potential energy surface may 

 have a "pass" through which dissociation can occur with little or no 

 energy of activation, the pass will be narrow and the height of the 

 asymptotic region outside the pass will be at most only slightly lower 

 than the potential surface height corresponding to the normal con- 

 figuration of the ion. Under such circumstances, the parent ion (and 

 then the fragment ions) will, in general, not dissociate immediately on 

 formation, but will "wander" around through various electronic states 

 until it happens into one where its nuclear configuration and momenta 

 are appropriate for dissociation. The dissociation will then occur with- 

 out removing appreciably more energy from the system than is essential ; 

 then the ion fragment will repeat the process and dissociate further, 

 until its energy is too low to cause further bond breakage. 



We w^sh to make still another argument based upon the existence of 

 this large number of states. Transition probabilities from the ground 

 state of the molecule to the states of the ion will no doubt vary over an 

 enormous range. There will be some molecules where the transition 

 probability to one or to a very small number of states will be comparable 

 to the sum of all other transition probabilities. This might arise in a 

 highly unsaturated, conjugated hydrocarbon or in a molecule containing 

 a heavy, easily ionized atom. More generally, however, there will be a 

 large number of states to which the transition probabilities will be large 

 and of the same order of magnitude. For a molecule as light as propane, 

 if this group were to constitute only one-tenth of 1 per cent of the total 

 it would still contain about a thousand states. These states will tend 

 to be concentrated among the lower states of the ion, since transitions 

 to the upper states will, in general, require change of a large number of 

 electron spins. Except for this requirement there is no reason to expect 

 these states with high transition probabilities to be distributed in any 

 other than a uniform manner among all the states of the ion. The 

 probability of transition to states of the ion will tend to cluster about a 

 mean value, the density of states being greatest for energies near this 

 mean and least near the energies corresponding to the most stable and 



