The pressure dependence of Schottky barrier height. at the Pt/GaAs interface has been studied using a diamond anvil cell. The pressure coefficient of the Schottky barrier height suggests that whatever states responsible for the Fermi level pinning follow the valence band edge under pressure. Within models. that simple intrinsic defects are responsible for the formation of Schottky barriers in GaAs, our results suggest that these. intrinsic defects may involve vacancies.

Deep centers induced by hydrostatic pressure in GaAs:Si have been studied by deep level transient spectroscopy and constant temperature capacitance transient techniques. The capture behavior of these centers has been studied in detail and found to be consistent with the multi phonon emission theory. The pressure coefficients of the ionization energy and the barrier height are consistent with the large lattice relaxation model proposed by D. V. Lang and R. A. Logan [Phys. Rev. Lett. 39, 635 (977)].Recently Mizuta et al. l found that when GaAs containing shallow donors is subjected to hydrostatic pressure in excess of 20 kbar a deep center similar in properties to the DX center in AIGaAs alloys appeared. These results stimulated much discussion concerning the nature of this pressure-induced deep center (to be abbreviated as PIDC here)

The pressure dependence of the DX center in Te-doped ternary alloy Ga& "Al As with x=0.15, 0.25, and 0.35 has been studied. The pressure coeKcients of the activation energies for emission and capture as well as the pressure coefficient of the thermal ionization energy were found to change sign when the band gap of Ga& "Al"As changed from direct to indirect. The compositional dependence of these energies at atmospheric pressure has also been extrapolated from the pressure dependence. Similarities between the effects of changing pressure and Al mole fraction suggested that the properties of the DX center depend mainly on the host semiconductor conduction-band structure. The difference in the pressure dependence of the DX center when the band gap is direct as opposed to indirect suggests that the defect wave function has considerable contribution from the conduction-band minima near the X point of the Brillouin zone and not just from the L-point minima as has been proposed by many authors.

concluded that these pressure induced deep donors (to be abbreviated as PIDD in this article) in GaAs are identical to the DX centers in GaAlAs.To substantiate this conclusion it is necessary to compare all the known properties of the DX centers with that of the PIDD in GaAs.The hydraulic press at room temperature. The pressure was determined by the standard ruhy fluorescence technique. 5Two optical measurements have been performed to study the properties of the PIDD in GaAs. In the first experiment the dependence of the electron photoionization cross section (a~) of deep centers on incident photon energy was measured. In the second experiment the thermal capture rates of optically excited carriers were determined at low temperatures. From these capture rates the decay times of free carriers in PPC can be calculated.In the first experiment light from a tungsten halogen lamp is focussed into a monochromator with a spectral width of 7 nm. The radiation from the monochromator is directed into the diamond anvil cell and scattered by the powder surrounding the sample. Since the diode is covered by metal electrodes on both the too and the bottom, only the 3 scattered light can enter the sample from the sides. As a result it is not possible to determine exactly the amount of light absorbed by the sample, however, it is still possible to measure the relative an() provided we assume that the light scattering efficiency is constant over the range of photon energies in this study. Except for this assumption we have corrected for the dispersion in the diamond absorption and in the lamp emission. As a test of the reliability of our measurement we have used the same setup to measure the dispersion of the u; of GaAlAs:Te at atmospheric pressure and compared our results with those reported by Lang et a1. 3We have used the method of Chantre et a1. 6 to measure CT~. The sample, whether GaAlAs at atmospheric pressure or GaAs under pressure, was first maintained at zero bias at room temperature in order to fill the traps and then cooled to 77 K in the dark. A reverse bias of 3 V was then applied to the diode. As the thermal emission rates of the DX center and the PIDD in GaAs at 77 K were both negligible, the traps in the depletion layer remained occupied. Next the sample was illuminated with radiation from the monochromator and the rate of change in the diode capacitance was measured. Since the change in capacitance was proportional to the change in the electron concentration in the deep centers (n 1 ) within the depletion layer we obtained in this way dnr/dt.If ~is the incident photon flux density, cT; and G"; are respectively the electron and hole photoionization cross sections for the deep center, then it has been shown that dnT/dt is given by:6 dn 1 /dt =£(hv) [a-;(hV)where NT is the concentration of deep centers and hV is the photon We note that the corresponding value for the DX center in GaAlAs:Si obtained by Lang and Logan was 1.25 eV 11 • On the other hand we could not obtain any reasonable fit to our results by assuming...