Nitrogen isoelectronic trap inGaAs1−xPx: II. Model calculation of the electronic statesNΓandNXat low temperature

Abstract

A semiphenomenological theory of the isolated N isoelectronic trap in $\mathrm{Ga}{\mathrm{As}}_{1\ensuremath{-}x}{\mathrm{P}}_{x}$ is presented, based on an extended (multisite), one-band Koster-Slater model of the electron-impurity interaction, including the effects of both the central-cell atomic pseudopotential difference and the spatially extended lattice distortion surrounding the substitutional nitrogen impurity. The alloy host is treated in a virtual-crystal approximation. The parameters of the model are determined by fitting low-temperature photoluminescence data from ion-implanted materials of two compositions selected near $x\ensuremath{\approx}0.35$. The model yields both a spatially localized ${N}_{X} (or A)$ state which evolves continuously with decreasing $x$ from the $A$ line of GaP, and a spatially-diffuse state ${N}_{\ensuremath{\Gamma}}$ which is present in near-direct and direct-band-gap alloys ($0.3\ensuremath{\lesssim}x\ensuremath{\lesssim}0.5$). The theory quantitatively describes the energies of these luminescence lines as a function of alloy composition $x$. In addition, good agreement with the data is found for the following calculated quantities: (i) the composition dependences of the ${N}_{\ensuremath{\Gamma}}$ and ${N}_{X}$ luminescence intensities, (ii) the pressure dependences of the ${N}_{\ensuremath{\Gamma}}$ and ${N}_{X}$ intensities, (iii) the composition dependence of the ${N}_{X}$ lifetime, and (iv) the binding energy of N${\mathrm{N}}_{1}$ pairs in GaP. The model leads to a tentative interpretation of ${N}_{\ensuremath{\Gamma}}$ luminescence as originating, at least in part, from excitonic molecules.