A dense Ga203 film could have desirable insulating properties. Experimentally, we have been unable to produce good insulating films with our photochemical technique. Perhaps due to the oxide's alledged porosity, capacitance measurements do not reflect classical metal-insulatorsemiconductor behavior. Anodic oxides of GaAs are typically satisfactory insulators. We have formed anodic oxides by the AGW process (8) and attempted to unpin the Fermi level at the anodic oxide-GaAs interface using our photochemical process. But in order to unpin the interface Fermi level, we find that we must remove all the anodic oxide (for instance by using a higher water flow rate and little or no illumination) whereupon the photochemical process will work but a poor insulator is obtained. If we start with our photochemical oxide on GaAs, anodizing it results in rapid repinning.
ConclusionsThe growth of Ga203 films on photochemically unpinned GaAs is documented. The Offsey et al. technique is modified to produce large-area treated wafers. Attempts to create good insulating films with unpinned GaAs interfaces have been unsuccessful. The recently reported hydrated sodium sulfide treatment (7), in certain aspects of its unpinning and repinning behavior, resembles the photochemical oxides.
ABSTRACTSi oxidized sequentially using 1602 and 1802 isotopic tracers was studied in detail to investigate the oxidation mechanisms for dry 02 oxidation. Analyses of these samples with SIMS provided detail profiles of the 160 and 180 concentration in the oxide. At 1 atm and 1000~ we found 180 reacted both at the surface of the oxide as well as at the interface. A subsequent anneal of the 180 oxides revealed that 180 in both regions of the 8i02 was incorporated in the lattice. Both 180 peaks at the surface and at the interface grew with oxidation time. The same 180 peaks also increased in growth rate with lower initial 1602 oxidation time. The growth rates for both peaks increased at higher temperatures. Profiles of the 180 in the oxide revealed a solubility limit for the diffusing oxidant in the oxide at 2 • 1020 cm -3. The same profiles also showed very little oxygen diffusion in the network of the oxide (D < l0 -16 cm2/s).) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 138.251.14.35 Downloaded on 2015-03-23 to IP