Defect identification in theAsGafamily in GaAs

Ab initio total-energy calculations are presented for intrinsic defects in GaAs with a particular emphasis on hyperfine interactions in order to clarify the atomic structure of the various ${\mathrm{As}}_{\mathrm{Ga}}$-related defects. For the ${\mathrm{As}}_{\mathrm{Ga}}\text{\ensuremath{-}}{X}_{2}$ defect complex the interpretation as an ${\mathrm{As}}_{\mathrm{Ga}}\text{\ensuremath{-}}{\mathrm{V}}_{\mathrm{As}}$ antisite-vacancy pair as was considered so far is challenged. An ${\mathrm{As}}_{\mathrm{Ga}}\text{\ensuremath{-}}{\mathrm{Ga}}_{\mathrm{As}}$ antistructure pair is the most likely identification. It is also unlikely that the ${\mathrm{As}}_{\mathrm{Ga}}\text{\ensuremath{-}}{X}_{1}$ defect can be identified as a distant antistructure pair as was considered from magnetic resonance experiments. The theoretical results obtained for the isolated ${\mathrm{As}}_{\mathrm{Ga}}$ point defect agree with the experimental data reported for the defect identified as isolated ${\mathrm{As}}_{\mathrm{Ga}}$ and, with the exception of a small broadening of the nearest-neighbor lines and of a moderate splitting in the fifth shell, for the EL2 as well. We speculate that at room temperature the EL2 will be an isolated ${\mathrm{As}}_{\mathrm{Ga}}$ defect which lowers its symmetry attracting some other mobile defect at the low temperatures required to perform magnetic resonance experiments. We have calculated the binding energies of antisites bound to a distant shallow acceptor and the influence of the pairing on the hyperfine interactions. We show that this mechanism could explain the broadening of the nearest-neighbor lines but not the splitting in the fifth shell.