Image interpretations computer simulations

A special mention is in order of high-resolution electron microscopy (HREM), a variant that permits columns of atoms normal to the specimen surface to be imaged the resolution is better than an atomic diameter, but the nature of the image is not safely interpretable without the use of computer simulation of images to check whether the assumed interpretation matches what is actually seen. Solid-state chemists studying complex, non-stoichiometric oxides found this image simulation approach essential for their work. The technique has proved immensely powerful, especially with respect to the many types of defect that are found in microstructures. 

Interpretation of two-dimensional images of sheet silicates has previously been addressed. Computer-simulated images of chlorite (22) and biotite and muscovite (2A) show that two-dimensional images of these sheet silicates allow determination of the basal spacing (spacing perpendicular to the sheets) and stacking order (relative rotations and shifts between sheets). Furthermore, these calculations show that at the Scherzer (optimum) focus, many, but not all, aspects of the structures may be interpreted directly. 

Since no computer-simulated images of these structures have been published, interpretation of these images has been purely intuitive, based on the 0.7-nm periodicity. Furthermore, the correspondence between the dark and light bands and the underlying structure is not known, although it has been assumed that, at optimum defocus, the dark fringes correspond to the T-0 layers. 

The computer-simulated images presented above and in Guthrie and Veblen (12) have shown that one-dimensional images of clay minerals and sheet silicates can be interpreted and can provide important information. The simulations have also shown, however, that even these simple images must be interpreted carefully. 

A review of First Principles simulation of oxide surhices is presented, focussing on the interplay between atomic-scale structure and reactivity. Practical aspects of the First Principles method are outlined choice of functional, role of pseudopotential, size of basis, estimation of bulk and surface energies and inclusion of the chemical potential of an ambient. The suitability of various surface models is discussed in terms of planarity, polarity, lateral reconstruction and vertical thickness. These density functional calculations can aid in the interpretation of STM images, as the simulated images for the rutile (110) surface illustrate. Non-stoichiometric reconstructions of this titanium oxide surface are discussed, as well as those of ruthenium oxide, vanadium oxide, silver oxide and alumina (corundum). This demonstrates the link between structure and reactivity in vacuum versus an oxygen-rich atmosphere. This link is also evident for interaction with water, where a survey of relevant ab initio computational work on the reactivity of oxide surfaces is presented. 

---