G. WILSE ROIilNSON 27 



tween two electronic states causes the energy difference between the 

 states to increase (the surlaces "repel each other") , depending upon 

 the magnitude ot the perturbation. This means that an extremely 

 strong perturbation is likely to cause such a divergence of the po- 

 tential energy surfaces that the first criterion above can no longer 

 be met. On the other hand, states can exactly cross only if there is 

 no pertinbation which mixes the states, but the transition proba- 

 bility for non-radiative transfer between such states is exactly zero. 

 There therefore appears to be a third important criterion for rapid 

 non-radiative transfer of energy. The perturbation lohich mixes two 

 electronic states between which non-radiative energy transfer is to occur 

 must be of the correct magriitude, not too large and not too small, 

 so that the energy difference between the interacting states is of the 

 order of environmental energy states. If the environment does not 

 act as a "catalyst," the perturbation must be an intramolecular one. 

 The effect of zero, weak, and strong interactions upon the energy 

 transfer mechanism is depicted in Fig. 3. Perhaps the spin-orbit 

 perturbation operator which mixes states of different multiplicities 

 is of optimimi magnitude to promote energy transfer between states 

 which approximately cross. ^"^ It is known experimentally that energy 

 transfer between excited singlet and triplet states in certain rigid 

 environments occurs in an average time much shorter than 10~^ 

 seconds. On the other hand, non-radiative transfer from the first 

 excited triplet to the ground singlet can be extremely slow in the 

 same environments. In benzene, the geometries and energies of the 

 triplet and ground singlet are different, and there is likely to be no 

 potential surface intersection (in the absence of environmental per- 

 turbations) ; in addition, the zero-order spin orbit perturbation, be- 

 cause of symmetry, vanishes between these two states. Similar state- 

 ments can be made about other conjugated ring hydrocarbons. It is 

 no wonder such processes are rare imtil a fluctuating environment is 

 introduced. On the other hand, the excited singlet and triplet proba- 

 bly have nearly the same geometries, have similar energies, and the 

 probability of potential surface intersection is high. Furthermore, in 



" One possible example (ref. 7) which has been observed but not fully discussed 

 is in the spectrum of the HNO molecule near ~')0 ni^. Rotational lines (normally 

 single) are split by a perturbation in the excited state of H\0. Since the elec- 

 tronic structure of HNO is very simple at this low energy, it can be surmised that 

 the only possible perturbing state is a triplet state. The maximum energy splittings 

 are of the order of 1 cm-' and it is expected that the singlet-triplet radiationless 

 intercombination process would be very rapid from these perturbed rotational 

 levels. For larger molecules the region of interaction may be very extensive. 



