Vapor pressures of gallium in the literature are too high because Ga2Ofalse(vfalse) formed in the reaction of gallium with the quartz effusion cells used for the measurements. Gallium was found to be less reactive with alumina than with quartz. Measurements in an alumina effusion cell gave vapor pressures for gallium that obey the expression logP )(normalatm=−14,900/T−0.515logT+7.34 Values for δHf298false[normalGa2Ofalse(vfalse)false]normalof−17.4±0.7 normaland−19.7±0.7 normalkcal/normalmole were found for the reaction of gallium with quartz and magnesium oxide, respectively. The latter value was employed to calculate the equilibria in forming Ga2Ofalse(vfalse) from the following pairs of reactants: gallium‐gallia, graphite‐gallia, tin‐gallia, gallium‐alumina, and gallium‐water vapor. The vapor pressures of silver and tin were determined and compared with literature values to check the constants of the system.

No abstract

The reaction of gallium with quartz to produce gallium suboxide, Ga2O , and silicon is shown by equilibrium calculations to be a mechanism by which gallium arsenide can become contaminated with silicon, and possibly oxygen, when synthesized in quartz systems. The contamination of gallium arsenide with silicon by this reaction is governed by the rate at which the gallium suboxide vapor is removed. The silicon content commonly observed in normalGaAs (up to a few ppm) can be accounted for by this mechanism. Suggestions for minimizing the attack are given. The apparent transport of normalGaAs in a system containing traces of water vapor can be explained by reaction of water with gallium in the hot zone to form Ga2O and hydrogen, which react with arsenic in the cold zone to produce normalGaAs and regenerate the water.

Chipman1 recently reviewed five lines of evidence that indicate that the presently accepted heat of formation of silica (-209.9 kcal./mole)2 calculated from the heat of combustion of silicon in an oxygen bomb may be as much as 5 kcal./mole too positive.Still more recently, Good3 measured the heat of formation of aqueous fluosilicic acid and calculated a value of -217.5 ± 0.5 kcal./mole for AHf°m (c, quartz) which is 7.6 keal./mole more negative than the literature value. Some recent effusion

Oxidation of chemically polished high-purity aluminum in dry oxygen, water vapor, and moist air at temperatures from 450 ~ to 640C ~ is characterized by a near-linear reaction rate to a weight gain of 3 ~g/cm 2, followed by a rate that decreases rapidly with further weight gain. Oxidation beyond the 5-7 ~g/cm ~ weight gain range is very slow. Oxidation is slightly faster in the moist atmospheres than in dry oxygen at temperatures above 550~ Oxidation of a commercial aluminum-magnesium alloy (5052) in dry oxygen and in moist air is much faster than for high-purity aluminum and proceeds to much higher weight gains. The rate does not conform to any recognized oxidation law, but it is more nearly parabolic than linear. The much higher weight gains obtained by other workers who used mechanically polished samples of both high-purity aluminum and aluminum-magnesium alloy are attributed to surface roughness. Electron diffraction examination of oxidized specimens show only eta alumina on the high-purity metal and magnesium oxide on the aluminum-magnesium alloy.Previous studies of the oxidation of solid aluminum at elevated temperatures showed differences in the amount of oxide formed and in the oxidation law followed. Of the more comprehensive studies, Gulbransen and Wysong (1), Smeltzer (2), and Aylmore, Gregg, and Jepson (3) found that weight gains of metallographically polished high-purity aluminum in oxygen exceeded the 30 ~g/cm ~ level before the rate diminished sharply. Gulbransen and Wysong found in 2-hr tests that a parabolic law was obeyed from 350 ~ to 475~ and a linear law from 500 ~ to 550~ Smeltzer found that oxidation followed a two-step parabolic law during the early stages, then decreased to a lower rate after the weight gain had reached 30-40 ~g/cm ~. The time required for completion of the initial rapid rate decreased from 20 hr at 450~ to 1 hr at 600~Aylmore, Gregg, and Jepson found that oxidation curves were parabolic after the first 2 hr at 400~ but had three distinct branches at higher temperatures, each with a succeedingly lower rate. They explained their results in terms of an amorphous oxide which forms initially and crystallizes to eta alumina.Using chemically polished samples of high-purity aluminum, Hunter and Fowle (4) learned that, upon oxidation, eta alumina formed to a thickness of 160-210A (2.7-3.4 ~g/cm -~ for an oxide density of 3.4 g/cm ~) at temperatures above 475~ Below 450~ the oxide was amorphous and did not grow beyond 50A. They noticed no difference between oxidation in air and in oxygen, but oxidation in moist air from 125 ~ to 275~ was slower than in dry air, although the same ultimate thickness was reached. Their oxide thickness determination was to measure the voltage required to cause normal current leakage through oxide films in a nondissolving electrolyte, and to convert this to oxide thickness by the relationship that 1 v = 14A for barrier oxides on aluminum.A possible explanation for the differences in the references just cited was contained in the work of Lewis and Plumb...