Search

Croissance de ZnSnP 2 sur differentes orientations de GaAspa epitaxie en phase liquide. Les meilleurs resultats sont obtenus sur les faces {100}. Determination des conditions de croissance permettant d'obtenir la phase chalcopyrite

A model is developed for the incorporation of donors and acceptors in epitaxial GaAs which indicates that the thin p‐type region often observed at the layer‐substrate interface in n−p+ structures is caused by the electric field associated with surface states and substrate doping. The model also predicts impurity gradients at the outer surface, about which little is known experimentally. Since results from the model are in agreement with many experimental observations, surface states and substrate doping are believed to be the major cause of impurity gradients in epitaxial GaAs.

Column‐IV (Si, Ge, and Sn) and column‐VI (S, Se, and Te) impurities are used to dope high‐purity layers of vapor epitaxial GaAs. Resistivity and Hall data on a large number of samples show that the common donors in GaAs (assumed to be column‐IV impurities on gallium sites and column‐VI impurities on arsenic sites) are compensated by acceptor centers which are donor impurity‐vacancy complexes. These acceptor complexes are tentatively identified as column‐IV impurities on gallium sites associated with single gallium vacancies and column‐VI impurities on arsenic sites associated with single arsenic vacancies.

Using back reflection Laue measurements we have examined the substrate orientation dependence of antiphase domain boundaries for ZnSnP2 (c= 2a) grown on GaAs by liquid phase epitaxy. These results together with those of other workers indicate that {211} substrates are best for the growth of chalcopyrite layers withc= 2a, while {100} substrates are best for those chalcopyrites withc

The chalcopyrite alloy ZnxCd1−xSnP2 is a potentially use-ful electronic material. In addition to having effective masses lower than and energy gaps similar to its III-V compound analogs, this alloy can also be lattice matched to InP. We have used an open-tube, sliding-boat, liquid-phase system to grow ZnxCd1−xSnP2 epitaxially on InP sub-strates. Unintentionally-doped layers have electron con-centrations as high as 3 × 1019cm−3 with mobility values of about 2,000 cm2/V-sec. These mobility values are sub-stantially larger than have been obtained in the equivalent III-V materials at similar concentrations.

The distribution of impurities incorporated in epitaxial layers during their growth is determined by diffusion due to concentration gradients and by drift in the built-in electric field. This problem has been previously treated in the approximation that the impurities reach their equilibrium distribution during growth. We have extended this calculation to treat impurity drift and diffusion under non-equilibrium conditions, a situation much more characteristic of most realistic growth condi-tions. In the new calculation, the solution of the Shockley-Poisson problem is exact and the boundary con-dition at the growth interface is an approximation based on the electrostatics of surface states. The new calcu-lation also permits consideration of outdiffusion from the substrate, a phenomenon of technological significance. It is also capable of modeling variations of source im-purity concentration and growth rate during epitaxial growth. Calculations modeling n-type GaAs epitaxy of practical interest are presented.

Zero‐bias Schottky barrier capacitance measurements indicate the presence of heavily‐doped surface impurity gradients in n‐type epitaxial layers of GaAs grown on n+ substrates. These impurity gradients are produced during epitaxial growth by the electric field due to 5×1011 cm−2 surface states in GaAs in 1000 K. For n‐type layers grown on Cr‐doped substrates, high‐resistance regions are observed at the surface. These regions are probably caused by the buildup of positively charged chromium atoms in the electric field at the surface. The redistribution of chromium atoms in ion‐implanted samples during annealing can also be explained by electric fields due to surface states and the implanted‐ion profiles.

Photothermal ionization measurements on Si‐doped GaAs grown by molecular beam epitaxy (MBE) have indicated that the previous identification of the Si donor peak was incorrect. Data leading to the new identification are presented and the results leading to the earlier identification are re‐examined. Implications of the new identification for the importance of Si as a residual donor in epitaxial GaAs grown by various techniques are discussed.