Observations of the Solid Phase Epitaxial Regrowth for GaAs/Si
- PDF / 3,490,653 Bytes
- 6 Pages / 420.48 x 639 pts Page_size
- 8 Downloads / 223 Views
OBSERVATIONS OF THE SOLID PHASE EPITAXIAL REGROWTH FOR GaAs/Si. Philip Kightley, Peter J Goodhew, Peter D Augustus* and Robert R Bradley*. Department of Materials Science and Engineering, University of Liverpool, PO Box 147, Liverpool, England L69 3BX. *GEC-Mareconi Materials Technology Ltd., Towcester, Northants. ABSTRACT. The changes in defect structure caused by an anneal are studied by TEM for thin GaAs layers on Si. A low growth temperature of 4501C is always used for the first 300k. When it is first deposited this layer contains considerable disorder which consists of planar defects, 11121 epitaxy and an irregular distribution of dislocations. Annealing of the layer gives rise to a solid phase regrowth that consumes the misoriented and much of the twinned crystal and produces regular misfit dislocation arrays. Continued growth at high temperature can reduce the twin densities to below the detection limit of TEM. It is this solid phase transformation of this layer that allows good epitaxy to continue. INTRODUCTION. The use of a low growth temperature GaAs layer, followed by an anneal, permits the improved growth of GaAs/Si [1], although there have been relatively few observations on the mechanisms that allow this improvement in crystallinity. In this paper our observations of the generation of a regular misfit dislocation array, and of de-twinning mechanisms that are activated during the anneal, are presented. The anneal is best described as a solid phase epitaxial regrowth of the layer (SPER). The paper is set into three sections which concern the presence of {211 ) epitaxy, twinning and the generation of misfit dislocations. EXPERIMENTAL. The samples used were grown by MOVPE and are; Structure
Sample
50k GaAlAs/Si - island morphology. 300k GaAlAs/Si. (the pre-layer) 300A GaAIAs/Si Annealed (750'C/10 mins). 2 microns GaAs/300A GaAlAs (as for sample C)/Si
A B C D
The samples had their first 300A (or 50A for sample A) grown at 450'C. Samples C and D were then annealed at 750'C for 10 minutes. Sample D also had 2 microns of GaAs deposited onto it. This was performed at the normal growth temperature of 650"C. Sample D was etched to reveal the interface structure using 3:1:1 NH3 :H20 2 :H20. The first 300A of all samples was a 5 period 30A/30A AlAs/GaAs superlattice starting with AlAs. Transmission electron microscopy has been performed in both [001] plan view and [110] cross section at 120 or 150keV in various bright field, dark field and weak beam reflections. RESULTS. 1
Accidental f 1121 epitaxy.
The (001) zone axis transmission electron diffraction patterns from samples B and C are shown in figures la and b. For both samples the area selected for diffraction was -0.3g1m thick so much of the diffraction is from the silicon substrate. The double diffraction effect that gives multiple matrix spots is a clear indicator that the epilayer is at least partially relaxed for both samples. This arises from the diffracted (primary) beams from the first crystal acting as sources for diffraction in the second crystal
Data Loading...
