E), in excellent agreement with the experimental value (0. 11 eV).In conclusion, it seems to us that the present computations of e~(E) agree with the experimental data above the absorption edge much better than computations based on Kane's model for what concerns the most relevant feature of the resonant structure, i. e. , the energy position of the peak with respect to the shoulder. Indeed, taking into account the approximations inherent to the use of Eqs. (4) and (5), as well as the fact that no phonon processes were considered, it is not surprising that both theoretical width and oscillator strength of our peaks turn out to be smaller than the experimental ones. We wish to point out that the Hermanson method allows for an accurate adjustment of the band structure, the method being extremely helpful in solving ambiguities about critical-point assignments.We wish to thank G. Samoggia for stimulating discussions and D. Barbieri for continuous assistance during the machine computations.The dispersion of the Brillouin-scattering cross section of fast-TA phonons in GaAs was measured in detail in the range 8850 to 11 500 A at 295 K. Acoustoelectrically amplified domains supplied very intense acoustic flux in the range 0. 2-1.0 GHz; this permitted a detailed study of the dispersion with a monochromator and a continuous light source'. Close to the intrinsic absorption edge, we see a resonant contribution to the scattering cross section. Further from the edge, a deep minimum indicates that the contributions to the elasto-optic coefficient P44 from resonant and nonresonant terms have opposite signs. At still longer wavelengths, the dominant trend is determined by the expected & dependence of the scattering cross section. The effect of increasing phonon intensity on the dispersion curve was studied and correlated with an associated broadening of the intrinsic absorption edge.

We report an extensive set of experimental measurements on electric-field echoes in insulating and semi-insulating crystals. Previously suggested mechanisms cannot explain the results. In particular, three-pulse echoes with decay times T t of several hours have been observed at low temperatures in semi-insulating 1I-VI compounds. We propose a mechanism based upon the redistribution of trapped electronic space charge.

Unexpected lines in the acoustic paramagnetic resonance spectrum of KMgF 3 :Ni 2+ are assigned to second nearest neighbour and more distant Ni 2+ ion pairs. Exchange parameters are deduced for the pairs, and the mechanism for their spin-phonon interaction is inferred to be the single-ion, ( S 1 . d . S 1 ) + ( S 2 - d . S 2 ) process.

Using an interference pulse-echo technique at 9 GHz, we have observed acoustic paramagnetic resonance absorption and dispersion from the spin systems KMgF3:Ni2+, KMgF3:Fe2+ and KMgF3:Co2+, and have confirmed that, in the absence of saturation, the functions may be related quantitatively through the Kramers-Kronig expressions. The moduli of the quadrupolar magnetoelastic coefficients | G11 | and | G44 |, for Ni2+ and Fe2+, were inferred to be 58 and 39, and 1340 and 1000 (cm−1 unit strain−1), respectively. For Co2+ we found the dipolar coefficients to be | F11 | = 64 and | F44 | = 35 (unit strain−1). The observed angular dependence of the absorption in Co2+ cannot be accounted for by the simple spin-lattice Hamiltonian. The experimental values of the magnetoelastic coefficients have been compared with point-charge crystal field calculations.