[001] imaging of rutile: A problem in high-resolution microscopy
HRTEM images of the [001] zone of rutile (fig. 1) show 0.32 nm fringes near the edge of the crystal, but these rapidly change to 0.46 nm in the thicker parts of the crystal. This change in spacing is only possible if the intensities in the dynamically forbidden {100} reflections become comparable to the intensities of the {110} reflections. The {100} reflections are dynamically forbidden because the structure has 2-fold screw axes parallel to a and b and n-glides perpendicular to a and b. The presence of 0.46 nm rather than 0.32 nm fringe spacings in images of the thicker crystal regions presents a severe problem in matching the images to computer simulations. Fig. 2 shows [001] zone axis images for thin and thick crystals. As expected from symmetry, the computed images show only 0.32 nm spacings. In an attempt to explain the mismatch between computed and experimental images several effects not normally included in image calculations, and which could cause a change in the symmetry of electron diffraction patterns, were investigated, all without success.