Decay patterns of trapped electrons in aqueous glasses have been analyzed over the time span of more than eight decades in terms of a time-dependent "rate constant" of the form t ', 0~n~1, which can be derived from the long-tail hopping-time distribution of Scher and Montroll. The curve fitting of experiments in our model is quite satisfactory with n -0.1 and with another adjustable parameter which depends on the trapping species and varies within one order of magnitude from one system to another. The time-dependent spectral shift for the absorption spectrum of trapped electrons can also be fitted by a -0.1, providing fairly direct evidence that non-Gaussian diffusion of electrons from shallow to deep traps underlies the f dependence.

A numerical solution of the Smoluchowski equation for diffusion, reaction, and neutralization of ion pairs in a dielectric medium in the presence of a scavenger has been applied for analysis of recent electron scavenging experiments in alkanes. It is shown that these experiments can be quantitatively understood in terms of a simple exponential distribution of initial electron-positive ion separation distances. The credibility of this distribution is considerably enhanced when it is employed, in conjunction with Onsager's solution for diffusion of an ion pair in the presence of an external field, to calculate ion yields as a function of field strength. The agreement with experiment is once again excellent. The attenuation parameter which characterizes our initial exponential distribution is related to the liquid structure factor and gas phase molecular scattering cross sections through the multiple scattering theory of Lax. The wave nature of electron motion is explicitly considered in a discussion of the physical meaning of ionization in liquids.

The line shapes of absorption spectra of trapped electrons in metal-ammonia solutions, water, and alcohols are compared in detail with photodetachment spectra of spherical potentials, which have been applied extensively in liquid helium. A qualitative discussion is presented to point out some uncertainty in the role of the bound-bound transition ls -2p, which is predominant in the dielectric continuum model. Apart from the uncertainty related to the effective mass of the polaron, the coherence of the 2P state in the more realistic fluctuation potential is likely to be small. A molecular model for the trapped electrons is briefly described.

An attempt is made to formulate a theory of the reaction rate constant for electron scavenging which is valid for high mobility nonpolar liquids. It is proposed that the scavenging mechanism is an electron capture process which is completely analogous to the gas phase process. Application to the liquid state of theory developed for the use in the gas phase requires careful consideration of the relative energies of the quasifree (i.e., V0) and captured states of the electron and it is shown that a reasonable conjecture (that the polarization energy of a scavenger negative ion is the same in all hydrocarbons) allows a semiquantitative correlation of the known experimental results. When the energy of the negative-ion state of the scavenger in the Franck–Condon region is above the V0 level (e.g., C2H5Br in TMS), the reaction rate is shown to increase with V0, while if the former is below the V0 level (e.g., CCl4 in neopentane), the reaction rate decreases with V0. The temperature dependence of the observed rates is qualitatively consistent with the proposed model.

The use of a central-field type of molecular orbital for the electronic wave functions of simple molecules has been investigated and application has been made to H 2, H20 and CH4• Ionization potentials calculated with the use of these wave functions are generally in good agreement with experimental values. No empirical data have been used in these calculations except the known equilibrium dimensions of the molecules. Interactions of all electrons are included. Configuration interaction, however, is not used.Excitation potentials for H 20 and CH4 are not in very good agreement with experimental values, which possibly means that configuration interaction cannot be neglected in such calculations.