Weathered perthite and mixed muscovite-kaolinite from a kaolinitic granite at Trial Hill in east Queensland and kaolinized sericitic alteration from a granite from the Ardlethan Tin Mine of New South Wales were examined by optical, scanning electron (SEM), and transmission electron microscopy (TEM) to determine the alteration process of muscovite to kaolinite and kaolinite to halloysite (7Å). Muscovite was found intimately interleaved with kaolinite in a variety of proportions on a sub-micrometer scale. The contact was generally parallel to the (001) layers of both minerals, and the thickness of the contact layer alternated between 10 and 7 Å over short distances. Where the kaolinite to muscovite contact was at an acute angle to the muscovite layers, a small angle existed between the layering of the two phases, consistent with a topotactic alteration of muscovite to kaolinite. One tetrahedral sheet in the muscovite appeared to have been removed over 50–100 Å, converting a 10-Å layer to a 7-Å layer. The mica near the contact with kaolinite was easily damaged in the electron beam and showed Al loss during analytical transmission electron microscopy; thus, H3O+ probably substituted for K+ in this transitional phase.

An SEM examination of completely weathered plagioclase showed kaolinite plates having attached, parallel, polygonal rods of halloysite (7Å), which had planar sides and a central void, partly fused with the surfaces of the kaolinite crystals. TEM study showed that the kaolinite altered to halloysite, and that, where the kaolinite was partly altered to halloysite, a series of sharp kinks were present in the kaolinite plate in which alteration had occurred. These kinks were interspersed with linear kaolinite relics, 0.1–0.2 μm long, which appear to have provided local rigidity to the clay packet. Apparently, the altered clay first curled into loosely wound spirals, which ranged in cross-section from triangles to irregular octagons, with pentagons and hexagons being most common. The tendency to pentagons and hexagons compares well with a statistical study of the angles, which were most commonly grouped around 120°. As alteration of the kaolinite relics progressed, the linear parts of the spiral lost their rigidity and became circular or oval shaped. The long axis of the halloysite spirals was parallel to the X axis of the kaolinite. Halloysite spirals formed most readily if they had space to curl; if space was not available, the halloysite formed sheaves. Rare, thin layers of muscovite were present projecting through kaolinite into halloysite. Where muscovite relics reached open spaces, the 10-Å structure expanded to 14 Å.