INTERDIFFUSION IN ALLOYS OF THE GAINASP SYSTEM

Interdiffusion experiments and results for InP-GaInAs(P) heterostructures are reviewed and discussed within a thermodynamic model. Important factors affecting interdiffusion in the GaInAsP system are shown to include the (1) miscibility gap, (2) different diffusivities on each sublattice in each material, (3) Fermi level or impurity induced changes in diffusivity and perhaps diffusion mechanism, and (4) experimental technique chosen. With a miscibility gap present, the activity coefficients and solubilities of all species vary near a heterojunction and cause the interdiffusion to become strongly composition dependent. At commonly used growth and annealing temperatures, many superlattices are expected to equilibrate as two quaternary superlattices rather than a homogeneous alloy. Different diffusivities on each sublattice in a superlattice can lead to either a widening or a narrowing of quantum wells. When this occurs, optical measurements of the band gap energy are likely to be misleading because of quantum size effects. Diffusivity on each sublattice can be changed by the presence of group II, IV, or VI dopants. Diffusion on the group III sublattice in p-type GaInAsP is found to be consistent with an interstitialcy mechanism, but the mechanism remains unknown with n-type doping and for the group V sublattice. Diffusion experiments performed in closed, open, and capped environments are discussed. Poorly designed and controlled experiments are found to be associated with large discrepancies in observed diffusivities, unreliable concentration profiles, and new condensed phases appearing in the solid. Experiments to date indicate that the Cu-Pt ordered structure often found in GaIn(As)P epilayers are unstable, and not strain stabilized, relative to the disordered structure at commonly used growth and annealing temperatures.