A new solution system consisting of H3PO4, H202, and H20 was found useful for etching GaAs wafers. This solution system can be divided into four regions a-d, according to etching characteristics. The boundaries between the regions are given by a rfiole ratio (m 9 r) of H202 to HsPO4 of about 2.3 and a mole fraction (m 9 5) of H20 of about 0.9 at room temperature. Ratelimiting processes are: a, adsorption of H202 (m. r <_ 2.3, m. ] ~ > 0.9)', b, diffusion of H202 (m 9 r <~ 2.3, m 9 ] < 0.9) ; c, dissolution of oxidized products (m-r > 2.3, m-] <_ 0.9); and d, adsorption of H~PO~ (~-r _ > 2.3, m. ] > 0.9). Solutions in region a have a reproducible etching rate of 0.01-0.1 ~m/min, which is useful for MESFET processing. Crystallographic etching is also available with solutions in region c.

A new method to anneal implanted silicon wafers in a few seconds using a halogen lamp as a radiation source is proposed. Boron-implanted (200 keV, 1×1013 cm-2) silicon wafers were completely activated with little diffusion by radiation for 6 seconds. Conventional annealing for 15 minutes caused diffusion of boron. It was found that the activation of implanted boron was determined by the maximum temperature during annealing.

No abstract

InAs/GaAs single quantum well structures have been grown by metalorganic chemical vapor deposition. The grown wells are 1.03, 1.3, and 1.7 InAs monolayers thick. The 4 K photoluminescence spectra exhibit strong and narrow peaks, their energy decreasing smoothly with increasing well thickness. The noninteger value is interpreted on the model that the interface is macroscopically flat but has valleys and hills with their lateral extent smaller than excitons. The effective interface position is determined by their relative lateral extent.

Current–voltage characteristics of GaAs/AlxGa1−xAs /GaAs heterobarriers grown by metalorganic chemical vapor deposition were investigated for x from 0.32 to 0.46. Calculation of the current, which included just two components—a tunneling component and a thermionic component—agreed well with experimental results. The tunneling effective mass used is the same as the Γ point effective mass when x is 0.32; this value allows the calculated results to fit the experimental data. The tunneling effective mass becomes much larger than the Γ point mass as x increases. This fact suggests that the other band edge X participates in the tunneling process, making the effective mass larger as it approaches the Γ point edge. The barrier heights deduced from I–V and I–T relations are in good agreement. The results support a ratio of 60%–65% of AlGaAs/GaAs conduction-band discontinuity to the total band-gap difference.