Photoelastic trends for amorphous and crystalline solids of differing network dimensionality

A single model is developed for the different photoelastic response of Ge-family materials and chalcogen-based molecular solids. If is the "Grneisen" parameter for the electronic susceptibility, experiment shows that <0 for the former group, while >1 for the latter. In addition, several group IV-VI compounds have 0 1. In our model the dielectric constant is calculated, within the Drude formalism, using one Penn-Phillips oscillator for Ge-family solids and two for molecular chalcogenides. The model predicts that should depend linearly on 2 Eg, with Eg the Penn-Phillips gap and dimension-less determined from experiment. Reliable values of Eg, and other relevant parameters are tabulated for a large number of materials. New experimental results are also presented for ZnTe. The experimental evidence provides support for the model. A plot of versus 2Eg exhibits the predicted linear correlations for materials with <0 and >1; the slopes are in excellent agreement with measured band-gap volume derivatives. These correlations pertain to amorphous and crystalline solids alike. For the molecular chalcogenides, it is concluded that band-broadening influences through a uniform "red shift" of the lower-energy oscillator with respect to the stationary upper oscillator. The observed photoelastic trends are related to bonding topology by analogy with arguments previously applied to phonons. >1 follows from the bonding strength dichotomy in (<3D)-network structures, whereas <0 obtains for covalent 3D-network solids. It is suggested that can serve as an indicator of network dimensionality for these two cases.