No abstract

Trends and perspectives in solid-state wetting / Mouillage solide-solide : tendances et perspectives A model for solid-state dewetting of a fully-faceted thin film Un modèle pour le démouillage solide-solide d'un film mince complètement facetté

Dewetting is a well-known degradation mechanism for thin films at elevated temperatures. It is driven by surface energy minimization and occurs while the film is solid. The dewetting process is characterized by the formation of holes, retracting edges, and the formation of thickened rims on retracting edges. In anisotropic single-crystal thin films, holes are initially faceted. It is often observed that the corners of the holes retract faster than the edges of the hole, leading to dendritic or star-shaped holes. This so-called “corner instability” is one of the defining morphological characteristics of the dewetting process, and an understanding of this instability may lead to new film patterning techniques. In this work, we present a study of the growth of natural and patterned initially square holes in single-crystal Ni thin films on MgO substrates. A characteristic structure near the corners of the holes was observed, and a model for the growth of faceted holes was developed based on these observations. Despite its simplicity, the model reproduces the observed phenomenology and is in quantitative agreement with experiments. The model reveals that the corner instability arises from a redistribution of mass to create a new hole perimeter, which can only be created at the corner. The consequence is that the corner reaches a steady-state constant retraction rate while mass accumulation at the rims causes their retraction rate to continuously decrease.

The capillary force drives the edges of solid thin films to retract. The distance a film edge has retracted over time is usually fitted to a power law. However, experiments and numerical simulations suggest that edge retraction does not follow a power-law. In this work, a simple geometric model for edge retraction is presented that reproduces the retraction distance versus time scalings of simulations for both isotropic and highly-anisotropic films, and is consistent with experiments. The earliest time at which a power-law fit becomes a reasonable approximation is calculated as a function of substrate-film contact angle.