Mechanisms of surface faceting and coarsening

Equilibrium faceting formation in vicinal Al2O3 (0001) surface caused by annealing

“…Different defects are outlined and the merging points never appear isolated (see Figure 6 (B)). This is a consequence of coarsening process of the facet domains: when different faceted areas grow and overlap, the observed junctions are produced [19].…”

High‐Temperature Morphological Evolution of Lithographically Introduced Cavities in Silicon Carbide

“…These observations suggest that the { } 1210 faces are unstable, i.e., not a component of the Wulff shape, and thereby unstable to breakdown into a "hill-and-valley" structure, as originally described by Herring [21]. These observations have a parallel in undoped sapphire, where experimental evidence shows that { } 1210 -type surfaces are stable and develop ledges that increase in height with anneal time, but the orthogonal { } 1010 surfaces break down into a series of inclined facets via a nucleation and growth process [19,[22][23][24].…”

“…Recent calculations by Mullins and Rohrer [38] suggest that the barrier to the nucleation of new facet layers is extremely large for facets above a limiting size of ≈1 nm, making particles/cavities without alternative sources of ledges (dislocations) unable to adjust their shape. In contrast, if cavities are etched into unstable surfaces, e.g., the m { } 1010 plane in sapphire, decomposition of these surfaces into hill-and-valley structures is observed [23]; cavities bounded primarily by such unstable planes undergo rapid morphological change at rates approximately consistent with a surface diffusion analysis [19].…”

Final technical report on Award No. DE-FG02-92ER45465

“…Herring's theorem states that any surface which is not a part of the Wulff shape (high-index, higher-energy planes) will reconstruct into a hill-and-valley structure comprising the low-energy facets; edge effects are neglected. The reconstruction of high-index surfaces of AIz03 is well known [23,24]. In particular, the faceting of (1010) (the first-order (m) prism plane) of Alz0, has been studied in great detail.…”