The triply incommensurate phase of 2H-TaSe2 obtained by cooling from the normal phase was investigated by TEM between 87K and 113K with the resolution of 3 A. Moire-like patterns observed in this phase were confirmed to be interference fringes due to the first and the second order diffraction beams from the incommensurate structure and were not due to the dark field diffraction contrast of domains of the commensurate structure as interpreted earlier. Lattice fringes (∼9Å) of this modulated phase do not show any discontinuity across the boundaries of regions of different contrasts of the moire-like fringes which is expected from domain boundaries. Instead, a periodic change in the spacing of the lattice fringes (phase slip region) expected from the superposition of split superlattice spots in forming the lattice image is observed. This is the first direct observation of the existence of the phase slip region which is also expected from the discommensuration theory.

In this short review we contrast and compare the relative merits of three imaging techniques in the CTEM and STEM, for detecting subnanometer size catalyst clusters. It is shown that incoherent dark field signals, such as those obtained by hollow-cone illumination in the CTEM, and the annular detector in the STEM, are significantly more sensitive to subnanometer size clusters than the bright field signal. Sensitivity to clusters of high atomic number increases as the illumination (or detector) semi-angle increases, although at the expense of decreased signal-to-noise. The STEM annular detector signal offers advantages over the CTEM hollow-cone signal in this respect since it has a higher electron collection efficiency. A further advantage of the STEM is that image contrasts may be enhanced by electronically combining the annular detector signal with the axial energy loss signal. This ratio method, or Z contrast technique, has been used with success to detect single Pt atoms on thin low atomic number supports.

The morphology of nickel-containing phases at each stage during the treatment of a 12%Ni/SiO2 catalyst was studied. Liquid impregnated catalyst initially shows nickel nitrate uniformly dispersed on the surface of the silica in the form of a film. After calcining, large rafts of NiO with faceted surface pits are seen. The reduced catalyst contains bunched nickel particles or large rafts on the surface of the silica.

Surfaces of commercially grown edge-defined film-fed growth sapphire (EFG α-Al2O3) were studied in the electron microscope using both reflection electron microscopy (REM) and conventional transmission electron microscopy (TEM). The as-grown sapphire surface, ostensibly {1120}, was characterized by “rooftop” structures which were often locally periodic. These rooftop structures consisted of alternating {1120} facets and additional facets inclined a few degrees. The crystallography of the surface facets was analyzed using REM imaging of bulk specimens, and trace analysis of back-thinned plan section TEM specimens. Surface roughness was measured by stylus profilometry. and these measurements were compared to the electron microscopy observations. Fine structural features parallel to <0110> directions were also observed in both REM and TEM experiments, and these were attributed to surface steps of atomic scales.

The reactive ion beam deposition of ceramic films onto unheated substrates can produce amorphous films with essentially molecular mixing. The annealing and hot isotatic pressing (hipping) of these films to produce crystalline phases have reproducable effects which are sensitive to the temperature and the density of the film. Experiments with titanium oxides indicate that it is principally the equilibrium phases that are formed and that hipping can be used to encourage the same transformations at lower temperatures.

Thin films of titanium oxide close to the stoichiometry of TiO2 were deposited onto unheated substrates of sodium chloride. Some of the films were removed from the substrate by floating them off in water and the remainder were either annealed or hipped to induce crystallization. The anneals were performed either in air or argon and the hipping was done under an argon pressure of about twenty thousand pounds per square inch. Several of the free standing films were annealed in the same atmospheres on nickel grids. All the specimens were prepared for transmission electron microscopy by the same floating technique and were examined in a Philips 400 T.E.M. at 125 keV. The as deposited films were amorphous and showed no visible texture other than that derived from a small amount of porosity. The films were sufficiently conductive that they could be examined directly in the T.E.M. without carbon coating provided they were supported on a grid of fairly fine mesh. One specimen was also examined in the Kratos 1.5 MeV high voltage electron microscope at the National Center for Electron Microscopy. The specimen was annealed in vacuum using an in-situ hot stage to directly observe the behavior of the film.

The post deposition annealing and hipping of these films reproducibly induced the crystallization of anatase below 800°C. This is the equilibrium phase [1] and the extent to which the films transformed and the morphology of the growing crystallites were determined principally by the film thickness. There was little difference between the responses of free standing films and films left on the salt substrate. They tended to transform at about the same temperature, which was reproduced in the in-situ hot stage experiment and the microsructures which formed were very similar. The dependence upon thickness was also reflected in all the microstructures of the different post deposition treatments and it was possible to complete the transformations that were very sluggish in some of the films by hipping them at the same temperatures.

Qualitative and quantitative analysis of small catalyst particles is possible in the analytical electron microscope down to analysis areas on the order of 10 nm in diameter. The location of elements in the image field can be determined either by placing the electron probe on a particu-lar image feature or by forming a digital x-ray image showing the distribu-tion of various elements. In either case analysis of specimens of well defined thickness such as microtomed thin sections preserves spatial relationships in catalyst particles and simplifies interpretation of single element x-ray images. Electron energy loss spectroscopy can be combined with x-ray spectroscopy to reduce the ambiguity in x-ray spectra caused by spurious x-rays generated by electrons scattered from the analy-sis area to regions of high concentrations of elements removed from the analysis area.

We have used electron energy loss spectroscopy (EELS) and electron diffraction to study the local atomic arrangement in amorphous SixC1−x:H in the composition range 0.37 < × < 1. In the thin films, which were pre-pared by radio-frequency glow discharge from a mixture of methane and sil-ane, the π at the K-edge of C does not show up even for the highest C-content, i.e., Si0.37C0.63, consistent with fourfold coordinated C in the whole composition range studied. Also the electron diffraction results suggest a tetrahedral network. Models where the minority element is sur-rounded by four atoms of the majority element, fit the experimental data better than models based on a random distribution of Si and C on the tetra-hedral network.

TICx films prepared by reactive sputtering using a Ti target and different methane partial pressures were characterized by analytical transmission electron microscopy. The films are polycrystalline, and the plasmon energy increases considerably with increasing carbon content. Combination of the information obtained from electron energy loss plasmon and core loss spectra, and electron diffraction indicates that x in TiCx increases linearly with methane partial pressure. We find that the face centered cubic TIC phase spans the composition from TiC0.2 to TiC1.0 and when x<l we have a mixture of TiC1.0 and amorphous C.

Specimen artifacts such as grain boundary grooving, surface damage of the specimen, and Si contamination are shown experimentally to arise from the ion milling used in the preparation of transmission electron microscopy specimens. These artifacts in polycrystalline, ceramic specimens can cause clean grain boundaries to appear to contain a glassy phase when the dark-field diffuse scattering technique, the Fresnel fringe technique, and analytical electron microscopy (energy dispersive spectroscopy) are used to identify glassy phases at a grain boundary. The ambiguity in interpereting each of these techniques due to the ion milling artifacts will be discussed from a theoretical view point and compared to experimental results obtained for alumina.

The microstructural changes that are associated with creep in silicon nitride and silicon carbide are described and related to the creep mechanisms. Creep in hot-pressed silicon nitride and sintered silicon carbide involves mechanisms of cavitation and crack propagation within the bonding phase when the temperatures are below those necessary to activate plastic deformation within the matrix grains. In hot-pressed, magnesia doped silicon nitride the bonding phase is amorphous magnesium silicate and the creep damage involves decohesion and sliding of the matrix grains within the soft bonding phase. In yttria-doped silicon nitride the bonding phase is crystalline and diffusional interactions between this phase, the matrix and the atmosphere dominate the creep process. A similar condition is found in siliconized silicon carbide whereby the creep is controlled by the behavior of the silicon phase.