Studies in the System Alumina-silica-water

THE investigation discussed in this paper concerns phase equilibria in the alumina-silica-water system. Studies in this system are part of a re¬search project sponsored by the Geophysics Branch of the Office of Naval Research2 directed toward increasing knowledge of the stability relations of metamorphic mineral assemblages. Inasmuch as metamorphic rocks consist largely of aluminosilicate struc¬tures, many of them hydrated, the system alumina-silica-water is a logical starting point for the systematic laboratory study of metamorphic mineral systems. Although logical, this is not, however, an easy approach to the problem because of the very great difficulty of obtaining transformations among structures in response to changes in temperature, pressure, and con¬centration. Commonly equilibrium among the phases in this system is exceed¬ingly difficult to attain. New laboratory techniques are required to produce certain structural rearrangements, such as inversion of diaspore to boehmite or the transformation of alumina-silica mixtures to andalusite or kyanite. Phase equilibrium data for the alumina-silica-water system are consequently only partly complete, and only a summary progress report is presented here. Crystalline phases in the alumina-silica-water system are numerous and many are common minerals. The oxide components exist as: quartz, tridymite, cristobalite, corundum, water, and ice. Binary compounds occurring naturally are: sillimanite, andalusite, kyanite, and mullite in the alumina-silica system and gibbsite, boehmite, and diaspore in the alumina-water system. The most abundant among the ternary compounds is the composition A12O3.2SiO2.¬2H2O, existing in the four structural modifications, kaolinite, nacrite, dickite, and halloysite. The hydrated form of halloysite, A12O3.2SiO2.4H2O, is here called endellite. Pyrophyllite, A12O3.4SiO2.H2O, is the third ternary compound of natural occurrence. In addition to these minerals, two other ternary crystalline phases have been encountered in this study. The first is a structure resembling montmorillonite and corresponding perhaps to the pure aluminian end member of the beidellite series. The other has a composition approximating A12O3.2SiO2.H2O but its composition and structure are still in doubt. LABORATORY PROCEDURE In these laboratory studies, a starting material such as an alumina-silica gel or one or more of the crystalline phases is wrapped in platinum foil,