160 PENETRATION PHENOMENA IN LIQUID WATER 



the outer atomic electrons. They must also influence the ultraviolet 

 absorption spectrum, but about the latter, too, very little is generally 

 known. It has been observed that dissolving a substance in a non-polar 

 solvent generally affects its electronic absorption spectrum only slightly, 

 and this fact might seem to point to only small differences between the 

 stopping power of a molecular solid or liquid and the corresponding 

 vapor. It should be realized, however, that such observations have been 

 restricted, in effect, to the portion of the spectrum readily accessible, 

 that is, that in the visible and near ultraviolet regions. Light absorption 

 to higher levels, and especially to ionized states, would undoubtedly he 

 strongly modified. And it is just these higher excitation processes which 

 are decisive in determining the stopping power. (The pronounced ef- 

 fect on the stopping power of an alteration in the distribution of oscil- 

 lator strengths throughout the continuum has been noted previously — 

 for example, in connection with helium.) The absence of empirical 

 studies or theoretical analysis of this alteration * evidently calls for 

 some caution in assuming the accurate applicability of the Bragg rule 

 to molecular solids or liquids of low molecular weight. 



The above remarks view the influence of the medium as a perturba- 

 tion acting on a single atomic or molecular entity. In the case of a con- 

 densed substance one must also deal with collective effects arising from 

 the mutual coupling of several or many such entities. To such effects 

 are attributed the fact that the excited energy levels of a crystaUine soHd 

 are vastly different from those of the corresponding vapor. (The altera- 

 tion exceeds that arising from binding in a polyatomic molecule.) The 

 electronic levels of liquids are virtually unknown, but would without 

 doubt be rather similar to those of analogous solids. Admittedly, liquids 

 possess a lower degree of order (although not greatly lower), but this is 

 not very important. (Thus liquid metals exhibit almost as great elec- 

 tronic conductivity as the corresponding solids.) 



For polar media the influences referred to above are greater, and it is 

 surprising that the behavior of such media with respect to penetrating 

 radiation has practically never been studied. The alkali halide crystals 

 would seem an especially attractive case for study, for their dispersive 

 properties have been extensively investigated. Lithium fluoride, in par- 

 ticular, is likely to exhibit pronounced deviation from the Bragg rule,t 



* Instructive attempts to understand the modification of lower transitions have 

 been made. An outstanding instance is the red shift of the center of gravity of the 

 resonance absorption line of mercury atoms dissolved in water. Cf., for example, 

 the note of Phibbs (26) and literature there cited. 



t As an hypothetical and extreme model one might consider a medium of hydro- 

 gen in which the constituents are alternately H+ and H~. Here the electronic 

 stopping power for fast particles would exceed that of H2 by roughly 50 per cent! 



