Thermal transformation of chrysotile from Uruaçu District, state of Goiás, Brazil, heated in dry conditions at temperatures from 600°C to 1300°C was studied by high resolution electron microscopy and selected area electron diffraction (SAD). Up to 600°C, no morphological or SAD pattern changes were observed. At 600°C, the fibrils were still crystalline with the characteristics of the clinochrysotile. In addition, a new fringe system of 10–15 Å spacings appeared sporadically parallel to the 7.3 Å fringes of chrysotile. Areas of these extra fringes seem to constitute favorable sites for the nucleation of forsterite. At 650°C, forsterite nuclei appeared inside the nearly amorphous fibrils in the shape of patches consisting of flaky crystallites. At 700°C the chrysotile structure had disappeared; the new spots present in the SAD pattern were indexed as those of forsterite. Between 800–900°C the crystallinity of the patches was clearly demonstrated. From the lattice images in the patches, topotactic relations between chrysotile and forsterite were analyzed. At 1000°C very tiny grains of enstatite were formed mixed with forsterite grains. The SAD pattern is complex due to the coexistence of forsterite, enstatite, and silica-rich amorphous areas. From 1100°C to 1300°C the tridimensional growth of enstatite was promoted. The present results support the topotactic relations between chrysotile and forsterite found by X-ray analysis although differences up to several degrees may exist when these phases are observed microscopically. Evidence suggesting a topotactic growth between forsterite and enstatite was also obtained.

The <0.1-μm size fraction of an Ordovician K-bentonite from northern Kentucky was characterized by X-ray powder diffraction (XRD). Using A.I.P.E. A. criteria for interstratification nomenclature and Reynolds’ computer algorithm the dominant clay mineral proved to be an R2 ordered illite/smectite. The best fit of observed and calculated XRD tracings was obtained using 12 > N > 5, where N is the number of layers within a diffracting domain.

Sections of the K-bentonite were prepared by ion-beam milling and examined in an analytical transmission electron microscope (ATEM). One-dimensional lattice images observed parallel to the a-b plane showed subparallel packets, about 50–100 Å thick, each of which consisted of about 10-Â thick unit layers. Somewhat thicker unit layers (as much as 14.5 Å) were also seen. The former are presumed to be illite, whereas the latter may be partially collapsed smectite. Selected-area electron diffraction patterns suggested simultaneous diffraction from several packets, each containing at least five layers. Both h0l and 0kl spacings were usually present, indicating that the stacking of the subparallel packets was random. Quantitative analysis by AEM and electron microprobe show the clay to be low in tetrahedral Al but high in octahedral Mg, the latter presumably contributing largely to the interlayer charge responsible for K fixation. The TEM data are broadly reconcilable with the accepted XRD interpretation of a two-component, mixed-layer clay. Alternatively, the TEM images may be interpreted as a single phase having numerous packet boundaries, the latter being responsible for swelling behavior. These two interpretations will not be fully reconciled until greater analytical precision and resolution permit individual packets to be studied. This work suggests that mineral definitions based purely on XRD interpretations may have to be reconsidered as more electron microscope data become available.

Na-saturated bimessite was synthesized by oxidizing an alkaline MnCl2 solution with gaseous O2. Transmission electron microscopy (TEM) showed no morphological change upon K-saturation of the bimessite, but selected-area diffraction (SAD) revealed structural disorder. Stepwise heating of the K-birnessite to 800°C yielded cryptomelane, as indicated by X-ray powder diffraction (XRD). Various degrees of transformation to the final cryptomelane product were observed by TEM and SAD, but not by XRD. Unaltered bimessite formed chiefly as thin, platy crystals of apparent hexagonal outline. The cryptomelane crystals that formed by heating the K-saturated bimessite were acicular. Bimessite crystals, partially transformed to cryptomelane and displaying non-integral diffraction spots, formed plates having linear striations and rods twinned at 60° to each other. These intermediate products commonly retained some of the original hexagonal appearance of the parent bimessite. Structural disorder was detected in the partially transformed crystals by SAD. TEM revealed 6.8- and 4.8-Å spacings of the (110) and (200) planes of cryptomelane, respectively. The basal planes of bimessite layers (001) appeared to correspond topotactically to the (110) plane of cryptomelane during these transformations.