No abstract

Gallium phosphide has been grown epitaxially by open tube vapor transport using the H2/PCl3/normalGa system. Although normalGaAs substrates have been used for the majority of the work, the more recent use of Czochralski grown normalGaP has enabled a reduction in both strain and stacking fault density of the epitaxial layer to be achieved. Sulfur, tellurium, and zinc have been used as dopants in the preparation of both n‐ and p‐type layers, and relationships have been established between the dopant level in the vapor stream and the carrier concentration in the grown crystal. A Schottky diode technique has been used to detect variations in doping level throughout the depth of a slice and, after modifications to the growth process, uniform doping levels have been established. Measurements of the dependence of mobility on carrier concentration over the range 1014–1019 cm−3 have been made on this substantially homogeneous material and consistent results obtained. The levels of compensation in n‐type material have been shown to be between 1 and 2% over the range of carrier concentration 8×1016−2×1018 cm−3 . The previously reported dependence of doping level on substrate orientation has been confirmed for (111) surfaces and extended to the (100) and (110) surfaces for homoepitaxy.

No abstract

Layers of epitaxial gallium phosphide doped with either tellurium, sulphur or zinc over the range 10 '~ to 10 '8 cm-2 have been grown from gallium solution using a vertical dipping system. These layers of thickness 60 to 80 /~m have been produced on the (1 00) and (111)B faces of gallium phosphide single crystal substrates at high growth rates. Doping gradients have been investigated by a Schottky barrier technique over an angle lapped region of the grown slice, measuring the capacitance voltage characteristics of the individual barriers. Significant changes in doping level have been observed throughout the thickness of the layers and these are related to the variation of distribution coefficient, the losses of impurity from the system and the presence of competing background impurity systems.

No abstract