The chemical vapor deposition of compound semiconductors requires volatile precursor molecules, usually organometallic (om) liquids or metal hydride (mh) gases, which can be therMally decomposed on a heated substrate to produce the requisite epitaxial Material growth. An inherent disadvantage with tiie om's is that their decomposition produces hydrocarbon species which can contribute carbon impurities to the growing crystal. An inherent disadvantage of the mh gases is the difficulty of storing and handling toxic and flamMable gases safely at high pressures. Furthermore, certain hydrides are too unstable to contemplate storage as gases at ambient temperature, leaving no apparent choice except the om option. However, where there is a choice between mh gases and om liquids, the mh's (e.g. ash[3], ph[3]) are still generally preferred due to their higher purity and inherent absence of carbon contaminants. Point-of-use mh gas generators are proposed to overcome limitations of both classes of cvd precursors. In addition to the obvious cases of ash(3) and ph(3) generators, the possibility of generating more unusual and less stable hd's like h(2)te and ga(2)h(6) appears feasible. Development of such mh source would expand the options for mocvd precursors for deposition of hg/cdte and gaas crystals. the phase I objective is to evaluate alternative chemistries for a point-of-use generator and to determine the constraints involved in getting a gas stream containing h(2)te into a cvd reactor without preMaturely decomposing the h(2)te component. The contemplated source chemistries would also eliminate the storage of extreme toxics at high pressure.
