170 PENETRATION PHENOMENA IN LIQUID WATER 



Some studies of the ionization by high-energy radiations of non-polar 

 hquids such as carbon disulfide [cf., for example, the work of L. S. Taylor 

 (32)] have sometimes been cited in support of the view that the total 

 ionization in liquids is not much different from that in gases. Passing 

 over the difficulties in experimental measurements and in the interpreta- 

 tions thereof, which are formidable, we must state most emphatically 

 that the assertion that water behaves like a non-polar liquid such as CS2 

 is unjustified, to say the least. 



Of decisive importance in interpretations of chemical and biological 

 effects of radiations is the spatial distribution of the H and OH radicals 

 which are the secondary products of the energy transferred from ihe 

 radiation. In essence, it is believed that a pair of radicals formed by the 

 dissociation of an excited H2O molecule will most probably recombine 

 inside their so-called "Hquid cage," whereas such a pair formed sub- 

 sequent to an ionization act (namely, by dissociation of 1120"*" and in- 

 dependent dissociative capture of the electron by H2O) is separated by 

 a distance equal to that traversed by the ionized electron; this distance 

 is in general quite great enough so that the radicals will survive for sub- 

 sequent reactivity. Thus, grossly speaking, ionization acts are expected 

 to have chemical effectiveness, but excitation acts not. We must recall, 

 however, that the excited levels are all broadened in the liquid, so that 

 higher states (perhaps all but the lowest triplet) are actually continuous 

 and what is an excited state for a free molecule may correspond to an 

 "ionized" state in the liquid. (This phenomenon is familiar in the 

 spectra of gases as a reduction of the effective ionization potential by a 

 strong external electric field.) Indeed, because of the pronounced quasi- 

 crystalline structure of liquid water, the interaction of the excited mole- 

 cule may be with a number of closely coupled water molecules, so that 

 the energy level may approach in properties a conductivity band of short- 

 range order. Thus the "excited" electron may move through many 

 molecular diameters before being trapped — a phenomenon which has 

 been called internal ionization. The consequent chemical (and biological) 

 effects of such a process should closely approach those of a normal ion- 

 ization act. We thus are led to expect an effective W (in the sense here 

 adopted) much smaller than that for the vapor. For the vapor, W has 

 been found by Appleyard (33) to be 30 ± 1 ev. For the liquid we should 

 predict a value of the order of magnitude of half of this — roughly 15 ± 

 5 ev. [It will not be as small as the lowest excitation energy (-^7 ev) 

 because of the generous energy loss to vibrational modes and perhaps 

 also to the chemically non-effective lowest triplet state.] 



Considerations such as the above, with refinement of the meaning of 

 excitation and ionization acts in the liquid state and analysis of the 



