Burial diagenetic reactions of di- and trioctahedral clay minerals were compared in Brazilian offshore, basinal sediment sequences of Cretaceous age. Originally dioctahedral smectite-rich shales of three basins—Potiguar,Ceara, and Ilha de Santana—exhibited the classical smectite-to-illite burial pattern. Trioctahedral clay-rich shales and trioctahedral clay-mineral cements in sandstones, however, showed a burial sequence of saponite to mixed-layer chlorite/saponite with progressive increase in the percentage of chlorite layers with increasing burial depth.

The change from disordered to ordered interstratifications of chlorite/saponite occurred in the temperature range 60°-70°C at a vitrinite reflectance value of about 0.65. These values are lower than those found for the ordering of illite/smectite clays. Increasing substitution of Al for Si in tetrahedral sites, followed by fixation of interlayer hydroxide sheets was found to be the major chemical change promoting transformation of saponite to chlorite via corrensite.

Clay fractions separated from the A2, B, and C horizons of a soil formed on granite gneiss showed X-ray powder diffraction (XRD) spacings characteristic of trioctahedral illite. The trioctahedral illite was derived from biotite, and its development through various stages of weathering was followed by optical and electron microscopy combined with electron microanalysis. In the initial stages of weathering, Fe2+ within biotite was oxidized, without the loss of much K. During this process, biotite flakes became slightly buckled and fractured. Solutions moved into the damaged flakes leading to chemical weathering and exfoliation along cleavages and angular fractures. Major exfoliation broke up the flakes into segments, which themselves contained minor exfoliations and alterations along cleavage planes. The extent of exfoliation and alteration continued until thinner and shorter segments consisted almost wholly of thin (< 0.25 µm), parallel wafers separated by less compact layers of particles and microaggregates. The segments finally lost their shape and divided into clay-size particles. Parts of the thin wafers had the same chemical composition (and structure) as the original, intact flakes of oxidized biotite. The same parts of wafers retained some of the optical properties of the original biotite, and, when broken down to clay, they produced the XRD spacings of 10 and 1.54 Å, typical of fine-grained, trioctahedral mica (illite).

A system based on variation of the octahedrally coordinated cations is proposed for graphical presentation and subdivision of tri- and dioctahedral K micas, which makes use of elemental differences (in a.p.f.u.): (Mg – Li) [= mgli] and (Fetot + Mn + Ti – VIAl) [= feal]. All common true tri- and dioctahedral K micas are shown in a single polygon outlined by seven main compositional points forming its vertices. Sequentially clockwise, starting from Mg3 (phlogopite), these points are: Mg2.5Al0.5, Al2.167□0.833, Al1.75Li1.25, Li2Al (polylithionite), Fe22+Li, and Fe32+ (annite). Trilithionite (Li1.5Al1.5), Li1.5Fe2+Al0.5, Fe22+ Mg, and Mg2Fe2+ are also located on the perimeter of the polygon. IMA-siderophyllite (Fe2+2Al) and muscovite (Al2□) plot inside.

The classification conforms with the IMA-approved mica nomenclature and differentiates among the following mica species according to their position in a diagram consisting of mgli and feal axes plotted orthogonally; trioctahedral: phlogopite, biotite, siderophyllite, annite, zinnwaldite, lepidolite and tainiolite; dioctahedral: muscovite, phengite and celadonite. Potassium micas with [Si] <2.5 a.p.f.u. including IMA-siderophyllite, KFe22+ AlAl2Si2O10(OH)2, and IMA-eastonite, KMg2AlAl2Si2O10(OH)2 seem not to form in nature.

The proposed subdivision has several advantages. All common true, trioctahedral and dioctahedral K micas, whether Li-bearing or Li-free, are shown within one diagram, which is easy to use and gives every mica composition an unambiguously defined name. Mica analyses with Fe2+, Fe3+, Fe2+ + Fe3+, or Fetot can be considered, which is particularly valuable for microprobe analyses. It facilitates easy reconstruction of evolutionary pathways of mica compositions during crystallization, a feature having key importance in petrologically oriented research. Equally important, the subdivision has great potential for understanding many of the crystal-chemistry features of the K micas. In turn this may allow one to recognize and discriminate the extent to which crystal chemistry or bulk composition controls the occurrence of some seemingly possible or hypothetical K mica.