Indium-doped GaAs: A very dilute alloy system

The influence of indium incorporation in GaAs organometallic–vapor-phase-epitaxy (OMVPE) layers has been investigated in great detail. The results obtained concern the change in band-gap energy, the concentration of residual impurities, and the low-temperature (2-K) photoluminescence (PL) efficiency. For In concentrations ranging between 0 and 6.5×${10}^{19}$ ${\mathrm{cm}}^{\mathrm{-}3}$, both $A^0$X and $D^0$X bound-exciton lines could be resolved. Together with the near-band-gap transitions involving shallow impurities (DA- and eA-related recombination lines), they shift toward lower energies versus indium content. This indicates the formation of a ternary compound ${\mathrm{Ga}}_{1\mathrm{-}\mathrm{x}}$ ${\mathrm{In}}_{\mathrm{x}}$As, even at these extremely dilute indium concentrations. After a quantitative calibration of the indium content, linear relations have been found which connect the PL emission line energies and the indium concentration. They make low-temperature PL measurements the most quantitative, and nondestructive, tool for precise composition studies. In this case, care should be taken that the slope parameters are line dependent. For instance, we find a slight, but finite, discrepancy between the slope parameters corresponding to substitutional acceptors on Ga and As sites, respectively. This is discussed in terms of the two different sublattices by using a simple cluster model of 17 atoms. Lastly, we find the absolute PL intensities to increase versus indium concentration: This indicates an improvement in the optical quality of our samples. Since, on a relative scale, the PL signals involving ${\mathrm{Zn}}_{\mathrm{Ga}}$ and/or ${\mathrm{Mg}}_{\mathrm{Ga}}$ residual acceptors are not significantly affected by the amount of indium incorporated, but depend mainly on the growth sequence, we feel that indium in GaAs acts primarily by closing nonradiative-recombination paths which are not necessarily associated with gallium vacancies.