Photocapacitance measurements have been used to determine the electron photoionization cross section of the centers responsible for persistent photoconductivity in Te-doped ${\mathrm{Al}}_x{\mathrm{Ga}}_{1-x}\mathrm{As}$. The cross-section data, which have been obtained at various temperatures and for crystals of various alloy compositions, are fitted by a theoretical line shape that is valid for large lattice relaxation. The line shape and thermal broadening can best be fit by a binding energy of 0.10±0.05 eV and a Franck-Condon energy of 0.75±0.1 eV. These values are in good qualitative agreement with the large-lattice-relaxation model of persistent photoconductivity which we recently proposed. We show that the 0.10-eV binding energy is also consistent with experiments that locate this energy relative to the Fermi level. The dependence of the properties of the persistent-photoconductivity center on the donor doping of the samples leaves little doubt that this center involves a donor atom, but because the center is not effective-mass-like, we believe that it is a complex also involving another constituent. Accordingly, we designate it as a "$\mathrm{DX}$" center. The anomalously-large Franck-Condon energy (Stokes shift) and apparent fact that the unoccupied state of the $\mathrm{DX}$ center is resonant with the conduction band, yet sufficiently localized to produce a large relaxation, are thus well established. These considerations lead us to the propose that the most likely model for $\mathrm{DX}$ centers in ${\mathrm{Al}}_x{\mathrm{Ga}}_{1-x}\mathrm{As}$, and perhaps in other compound semiconductors as well, is a complex involving a donor and an anion vacancy. We show that such a model is qualitatively consistent with the overall trends in persistent-photoconductivity behavior observed in a variety of III-V and II-VI semiconductors.

• Received 25 January 1978

DOI:https://doi.org/10.1103/PhysRevB.19.1015