Characterization and elimination of surface defects in Gax In1−xP grown by organometallic vapor phase epitaxy

Synchrotron topography and X-ray diffraction study of GaInP layers grown on GaAs/Ge

“…The center dot areas of the hillock strain fields in Fig. 8(c) may be related to the polycrystalline hillock core identified by Matragrano et al in hillocks on GaInP layers using cross-sectional transmission electron microscopy [5].…”

“…This phenomenon can be explained by the increased strain inflicted into the GaInP lattice by the hillocks causing MD formation. The idea that the hillocks may act as sources for MD formation was first suggested by Matragrano et al based on cathodoluminescence observations in [5]. Their hypothesis is strongly supported by the relation between the MDs and the hillocks observed in Fig.…”

“…Some differences in the shapes of the X-ray maxima arising from the asymmetric strain is observed in RSMs of Figs. 3 and 4, but due to the relatively small strain in samples A and B, the differences in peak shapes between Figs Similar hillocks in Ga 1Àx In x P=GaAs with 0:388 r x r0:552 have been reported in [6], and they are a well-known problem in GaInP epitaxy [4,5,7]. However, it is possible to grow hillock-free GaInP layers using PH 3 as P precursor, if the PH 3 flow is modulated during the MOVPE growth [17].…”

“…However, it should be noted that there may well exist less ordered parts of the hillocks not visible in the XRD spectra and containing materials other than InP. In the GaInP/GaAs structure studied in [5] similar hillocks are observed to consist of an In-rich base region and of Ga-rich particulates. However, due to the relatively strong XRD intensity recorded from the InP crystallites in this work, it is safe to conclude that significant parts of the hillocks are made of rigorously ordered InP crystallites, which are connected to each other by twinned grain boundaries.…”

“…The GaAs/ Ge structures used as substrates in this work are of a high crystal quality and have been previously studied in [2,3]. The main idea for using the GaAs/Ge substrates is to apply prior knowledge about growing GaInP on GaAs [4][5][6][7] to Ge substrates with only minor modifications. Also, problems with anti-phase domains (APD) appearing in GaAs epilayers on Ge were already solved in a previous publication [2], and by using GaAs/Ge substrates instead of plain Ge substrate the possible problem of APD formation on GaInP/Ge interface is probably avoided.…”