X‐Ray, Photoluminescence, and SIMS Characterization of InGaAs / InP Grown by Vapor Phase Epitaxy

Abstract

X‐ray double crystal diffractometry (DCD), low temperature photoluminescence (PL), and secondary ion mass spectrometry (SIMS) have been used to characterize single epitaxial layers grown by hydride‐, chloride‐, and levitation‐vapor phase epitaxy (VPE). Coherent hydride layers exhibited both bound exciton and C related donor‐acceptor peaks, and the deduced bandgaps and alloy compositions were found to agree with published values. For coherent hydride‐VPE layers shifts of no more than 1 meV in the PL spectra for various positions on the wafer were observed. A layer found to be incoherent by DCD, on the other hand, exhibited luminescence shifted by up to 9 meV for various positions on the wafer. We conclude that large variations in PL spectra as a function of position on a wafer are indicative of an incoherent epitaxial layer. Incoherent layers were found by PL to contain C and Si acceptor impurities and by SIMS to have an increased S impurity content compared to coherent layers. A heavily Zn doped sample exhibited PL features similar to those of heavily Zn‐doped normalGaAs which were explained by invoking contributions from k‐nonconserving transitions. One of the chloride‐VPE layers also exhibited C and Zn related donor‐acceptor recombination. The PL spectrum of a levitation‐VPE sample exhibited both bound exciton and donor‐valence band recombination and, in addition, a weak Si acceptor related donor‐acceptor peak. This levitation‐VPE sample was found by SIMS to have a large Si impurity content. We infer that substitutional Si can reside on either a III or a V site in this material and is largely self‐compensating.

X‐Ray, Photoluminescence, Stoichiometry, and Thickness Mapping of In1 − x Ga x As y P 1 − y

“…For incoherent epitaxy, an additional term proportional to the in-plane mismatch must be added to the right side of Eq. [2]. However, for the presently studied wafers this correction is either zero or is quite small, as judged from the measured magnitude of (ha~a), the thickness of the mismatched layer, and the full width at half maximum (FWHM) of the Bragg peak.…”

“…(We defer a discussion on how to obtain the mismatch from rocking curves and how to obtain the bandgap from photoluminescence spectra to the next section). Equation [2] for the lattice mismatch and Eq. [4] for the bandgap with the corrections given in Eq.…”

Determination of epitaxic-layer composition and thickness by double-crystal X-ray diffraction

“…Detectable parallel mismatch has only been reported for crystals where the perpendicular mismatch is greater than +0.2%. Macrander & Swaminathan (1987) have noted that an exaggerated broadening of the 004 rocking-curve peaks of both the epitaxic layers and substrate signals the onset of incoherent epitaxic crystal growth. All the samples used for this study showed very narrow rocking curves.…”