Every 30 to 40 million, or so, operating hours a destructive reaction is observed in one of the approximately 4000 large compressors that move UF6 through the gaseous diffusion plants. Despite its infrequency, the reaction can be costly in terms of equipment and time. Laboratory experiments reveal that the presence of moderate pressures of UF6 actually cools heated aluminum, although thermodynamic calculations indicate the potential for a 3000-4000-degrees-C temperature rise. Within a narrow and rather low (< 100 torr; torr = 133.322 Pa) pressure range, however, the aluminum is seen to react with sufficient heat release to soften an alumina boat. Three things must occur in order for aluminum to react vigorously with either UF6 or UO2F2. 1. An initiating source of heat must be provided. In the compressors, this source can be friction, permitted by disruption of the balance of the large rotating part or by creep of the aluminum during a high-temperature treatment. In the absence of this heat source, compressors have operated for 40 years in UF6 without significant reaction. 2. The film protecting the aluminum must be breached. Melting (of UF5 at 620 K or aluminum at 930 K) is a way to do this in the laboratory. In contrast, holding Al samples in UF6 at 870 K for several hours produces only very moderate reaction. Rubbing in the cascade can undoubtedly breach the protective film. 3. Reaction products must not be allowed to build up and smother the reaction. While uranium products tend to dissolve or dissipate in molten aluminum, AlF3 shows a remarkable tendency to surround and hence protect even molten aluminum. Hence the initial temperature rise must be rapid and sufficient to move reactants into a temperature region in which products are removed from the reaction site.
