Photoreflectance Study of Doped Gallium Arsenide Beyond the Bandgap

We investigated the photoreflectance (PR) response of doped GaAs grown by metalorganic chemical vapor deposition (MOCVD) and molecular beam epitaxy (MBE) by exploitation of the E_1 structure. The objective was to study the effects of the doping concentrations (5 times 10^{15 } - 2 times 10^{19}), the doping element (Si, Zn or Be) and the growth technique on this optical structure. We proved that, in situations where the study of the E_0 structure yields ambiguous results, an investigation of the E_1 structure would be helpful. There is an overlap of E_1 and E_1 + Delta_1 structures as the carrier concentration increases beyond a certain limit. Our innovative ten parameter curve fitting technique unravels the overlaping structures and allows for the accurate determination of the PR lineshape parameters. Our results indicate that: (a) There is no significant energy shift in the E_1 structure with increased doping concentration. (b) There is an increase in the broadening parameter of E_1 structure, Gamma_1, which have a linear relation with the logarithm of measured carrier concentration above 1 times 10 ^{17}cm^{-3}. (c) This linear relation is the same for all Si:GaAs samples and it is independent of the growth technique. (d) The linear relation is the same for Zn: and Be:GaAs samples, which have a significantly smaller slope than the one for Si:GaAs and it is independent of the growth technique. This linear relation has a potential application in contactless doping level determination of GaAs. The strength of this approach in evaluation of carrier concentration includes, measurements on layers as thin as 20 nm, applied to layers within the depletion width and application to heavily doped materials as well. We also studied the results of secondary ion mass spectroscopy (SIMS) measurements on a selected number of MOCVD grown Si:GaAs samples. SIMS result indicated that large compensations are associated with our samples for all doping levels. Based on both SIMS and PR results we concluded that the above mentioned linear relation is independent of the compensation ratio (number of Si acceptors divided by the number of Si donors).