The existence of grain boundaries (GBs) in many advanced ceramics can significantly alter their performance, rendering the study of GB structure an essential part of understanding ceramic properties on a fundamental level. Furthermore, the debate regarding the stability of [001] twist GBs in rocksalt-structured oxides provides the impetus for comparing experimental observations of the Σ=5 [001] twist GB structure in NiO with the stable GB structure calculated by Tasker and Duffy (Fig. 1). The Tasker and Duffy structure was obtained by introducing a Schottky defect in each of the GB unit cells, resulting in a 20% decrease in atomic density at the GB. High resolution TEM (HRTEM) employing Fresnel fringe and lattice imaging techniques (Fig. 2) supplemented with image simulations of six possible configurations for the NiO Σ=5 [001] twist GB structure using the TEMPAS program (Fig. 3) were the primary experimental methods used in an attempt to detect this reduced atomic density in NiO Σ=5 [001] twist GBs manufactured by a modified Schober-Balluffi method. Since the Fresnel fringe technique consists of a change in Fresnel fringe contrast with defocus due to a change in mean inner potential at the GB, it becomes necessary to consider the contribution of the various causes of a change in mean inner potential to the resulting contrast. These causes include: (i) a change in atomic density or a change in atomic structure at the GB, (ii) the existence of a space charge at the GB, (iii) the existence of a thin amorphous film at the GB, (iv) the presence of impurities at the GB or a change in composition across the GB, and (v) GB grooving. The contribution of a GB space charge to the change in mean inner potential at the GB is assumed to be small, given that the space charge associated with with a GB in an ionic crystal is on the order of a few tenths of a volt, whereas the mean inner potential at the GB is on the order of 15 - 20 V. An attempt to address the impurity and grooving issues was made by performing EDS and EELS across the GB in an ultra-high vacuum, high resolution STEM.
Read full abstract