ENERGY EXCHANGE IN PHOTOREACTIONS 51 



peaked curve simulating a resonance curve. Compare for instance Fig. 

 1-16, the data of Zemansky plotted in this fashion, with Fig. 1-14. 



From the model and the expression for the rate constant, one readily 

 sees that the relation can become complicated when the different vibra- 

 tional frequencies lie in different molecules. For example, the trans- 

 mission coefficient or the potential-energy curve associated with section 

 III may change markedly as the quenching oscillator is changed. The 

 left side of Fig. 1-16- is characterized by primary collisions between mer- 

 cury and hydrogen atoms of the quenching molecules, so that the various 

 potential-energy diagrams for the different substances may be expected 

 to be quite similar, as the data indicate. It is not probable that the plot 

 of vibrational frequency against quenching efficiency will show the high- 

 est efficiency at the vibrational freciuency that just matches in energy 

 the electronic excitation energy. Indeed, in view of the numerous factors 

 that establish the collision requirements, it would be surprising to find 

 as good agreement as was found by Zemansky. 



Pressure broadening (Lorentz and Holtzmark broadening) of absorp- 

 tion spectra is similar to the quenching of fluorescence in many ways (see. 

 for instance, the discussion in Mitchell and Zemansky, 1934), though the 

 effectiveness of various substances in the two processes is not parallel. 

 The two phenomena, as well as a variety of luminescent processes, pro- 

 vide data for the study of diabatic collisions and are included in the 

 domain of the present theory. 



The "cracking" of highly excited molecules which occurs under elec- 

 tron bombardment in mass spectrometers provides interesting examples 

 of intramolecular diabatic processes. Hydrocarbons excited to high 

 energy states during ionization pass through many crossing points in a 

 single oscillation of the atoms. We have previously seen that the density 

 of crossing points will be very high. It appears probable that the ion 

 remains in any single state such a short time that the vibrational degrees 

 of freedom for that state never become fully developed. Consequently 

 the ion, moving as an isolated entity through the spectrometer, will dis- 

 sociate only when it finds itself simultaneously in a state in which dis- 

 sociation would normally occur and with the atoms of the dissociating 



Fig. 1-17. Potential-energy diagrams illustrating a mechanism by which a resonance 

 relation may be preserved for electronic-vibrational transfers of energy, (a) Con- 

 tour map of the upper electronic surface, (b) Vertical cross section through I illus- 

 trating nearly identical noncrossing Franck curves for upper and lower electronic 

 surfaces, (c) Vertical cross section through II illustrating difference in form of upper 

 and lower Franck curves at smaller Hg-(A'F) distances, (d) Variation in collision 

 energy Eo, with distances along the line of crossing points, line III. Eq is the mini- 

 mum collision energy along this line, (e) Vertical section through line V showing 

 crossing of Franck curves for upper and lower surfaces at the best gateway. (/) Col- 

 lision trajectory projected on a vertical plane. The dotted line is that part of the 

 trajectory lying on the lower surface. 



