Abstract

The growth by MOVPE of InSb epilayers on GaAs substrates has been studied. There are a number of well established problems in the growth of this material by MOVPE including the high lattice mismatch of 14.5%, the low melting point of InSb, and the poor thermal cracking of the commonly used antimony precursor, trimethylantimony. In addition, there is the problem with the stability of the “pick-up” from trimethylindium bubblers, leading to poor compositional control. We have investigated the use of a number of techniques including the application of low temperature buffer layers of InSb and GaSb, the use of a novel internally heated resistance heater with two independent zones to allow pre-cracking of the alkyls, and the use of alternative sources for both indium and antimony. Indium precursors used have been TMIn, EDMIn, a mixture of TMIn/TEIn and a saturated solution of TMIn in an involatile solvent. As well as TMSb, a new antimony source, tertiarybutyldimethylantimony (TBDMSb), has been assessed. The growth temperature was between 400 and 480°C, depending on the precursors used. A wide range of III/V ratios were investigated for the various combinations of precursors and combinations of pre-cracking and no pre-cracking. Epilayers were analysed for thickness, surface quality and electrical quality at room temperature and 77 K. In addition to the growth experiments, two analytical techniques have been used to monitor the InSb growth. The first involves the use of quasi-elastic light scattering of a He/Ne laser beam from the growing surface. The scattered light is very sensitive to the degree of surface roughness, so the transition from 2D to 3D growth can be easily seen by monitoring the scattered light from the surface with a silicon photodiode and an optical filter to allow only the laser light to be collected. Some modifications to the growth cell were necessary to keep the top area clean for access by the laser beam. The relative effectiveness of various growth regimes and buffer layers can be investigated by this method and the results correlated with optical and SEM micrographs. The second technique involves the application of ultra-violet spectroscopy to study the effective pick-up rates from a variety of alkyl sources. Metalorganic sources such as TMIn absorb conveniently in the 300-180 nm range, so a simple UV set up incorporating a single pass cell with a 10 cm path length has been used to record the spectra of the various alkyls used in this study, and to compare the pick-up rates from different sources and different designs of bubbler, concentrating on various indium sources. This technique also allows “in-situ” monitoring of the concentrations of alkyls present in the reactor, but this has not been pursued in this case.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call