Crystallographic orientation dependence of impurity incorporation into III-V compound semiconductors grown by metalorganic vapor phase epitaxy

Abstract

This article presents a comprehensive study of the dependence of impurity incorporation on the crystallographic orientation during metalorganic vapor phase epitaxy of III-V compound semiconductors. We performed doping experiments for group-II impurities (Zn and Mg), group-VI impurities (Se and O), and a group-IV impurity (Si form SiH4 and Si2H6). The host materials were GaAs, Ga0.5In0.5P, and (Al0.7Ga0.3)0.5In0.5P grown on GaAs substrates. We examined the doping efficiency on the surfaces lying between {100} and {111}A/B. Even though we grew epitaxial layers in a mass-transport-limited regime, the doping efficiency significantly depended on the orientation, indicating that the surface kinetics plays an important role in impurity incorporation. Comparing our results with other reports, we found that acceptor impurities residing on the group-III sublattice and donor impurities residing on the group-V sublattice, respectively, have their own distinctive orientation dependence. Si donors exhibit orientation dependences which are either negligible or are similar to group-VI donors, depending on the growth conditions. We constructed a model for the orientation dependences, considering atomic bonding geometries between impurity adsorbates and adsorption sites.

Oxygen incorporation in AllnP, and its effect on P-type doping with magnesium

“…24 This has been shown for O 2 -doped (Al .7 Ga .3 ) .5 In .5 P by Kondo,et al 13,25 who demonstrated that oxygen incorporation efficiency decreases rapidly for miscut toward (111)A. Kondo et al 25 have pointed out the similarity of this behavior to that of column VI donors such as S, Se, and Te which incorporate substitutionally on the column V sublattice. For a (111)A surface, group V sites encountered by O are weak adsorption sites with only one bond to an underlying group III atom.…”

Zn-Doped InGaP Grown by the LP- MOCVD

“…The hole concentration is supposed to be proportional to the concentration of zinc atoms in the layers. This substitution of Zn atoms by the hole concentration is possible as the activation coefficient of Zn was measured in previous works s, 9,14,17 and its value was near unity (in the range of accuracy of measurements). The data taken from literature can also be seen in the Fig.…”

Surface orientation dependency for AlGaAs growth rate with/without HCl in MOCVD

“…In particular, the (0 1 1) surface shows a clear difference in growth rate. This difference in growth rate with the growth surface orientation is probably related to the change in the surface terrace from (1 0 0) to (0 1 1) when the angle is tilting from (1 0 0) toward /0 1 1S A/B is 25.2 [2,14]. The step edge for a (1 0 0) terrace is reported to have (1 1 1) A/B steps and these step edges provide reaction sites, unlike the (1 0 0) terrace area which should be less reactive.…”

“…The step edge for a (1 0 0) terrace is reported to have (1 1 1) A/B steps and these step edges provide reaction sites, unlike the (1 0 0) terrace area which should be less reactive. A (0 1 1) terrace also has (1 1 1) A/B steps, but the (0 1 1) terrace itself has been reported to be more chemically active than the (1 0 0) surface [14]. As a result, growth with in a (0 1 1) terrace is thought to encourage island growth.…”