SOME COMPARISONS OF CHEMICAL BEAM EPITAXY WITH GAS SOURCE MOLECULAR-BEAM EPITAXY

This paper seeks to illustrate the major differences between chemical beam epitaxy (CBE) and gas source molecular beam epitaxy (GSMBE). The reasons for these differences nearly all eminate from the use of group III metal alkyls in CBE whereas GSMBE uses conventional elemental group III sources. From this difference can be traced the dissimilarity in growth mechanisms which can be further exemplified in some of the more important features of CBE. Using as an illustration the growth of GaAs from triethyl gallium, the model is extended to include the reduction in growth rate brought about by the addition of small quantities of In to the growing layer at temperatures in excess of 500-degrees-C. This is technologically important in the subsequent growth of InGaAs lattice matched to InP. The positive advantages of CBE, such as area uniformity and true selective area growth are also discussed in detail. InGaAs layers have been grown with a compositional uniformity of +/- 0.03% In and a thickness variation of < 1% across a 2 inch wafer. Selective area growth of InP has been achieved in windows opened in a silicon nitride mask. This phenomena has been studied as a function of substrate temperature and phosphorus overpressure. The technique of GSMBE is in essence the same as conventional MBE in terms of growth mechanism. However, its ability to use two group V elements is its particular stength and this is illustrated by a description of the growth of large area arrays of InGaAs/InP MQWs for use as quantum confined Stark effect modulators. The extension to complex structures, some containing 800 interfaces is reported as is the growth of such structures on both sides of an InP wafer.