|M.Sc Student||Gorelik Leon|
|Subject||Mechanical Properties of Thin Coatings in Molecular|
|Department||Department of Mechanical Engineering||Supervisor||Professor Dan Mordehai|
|Full Thesis text|
Understanding the mechanical properties of contacts at the nanoscale is key to controlling the strength of coated surfaces. Despite its importance, quantifying the plastic yield of coated surfaces at the nanometer scale under load remains a big challenge. Mechanical properties of specimens at the nanoscale are different from their bulk counterparts, e.g., strength may be substantially higher, size and shape dependent. Therefore, to study the plastic yield at the nanoscale one should consider the atomic structure of the coated surfaces, which is beyond most of the current computational capabilities. In this work, we solve simplified problems in order to explore to which extent existing continuum models describing contacts with coated surfaces can be extended to the nanoscale. Molecular dynamics simulations of hollow cylinders or coated rigid cylinders under diametrical compression are performed and compared with models at the continuum level, as two representative extreme cases of coating which is substantially harder or softer than the substrate, respectively. We show here that the geometry of the atomic-scale contact is essential to capture the contact stiffness, especially for hollow cylinders with high relative thicknesses and for coated rigid cylinders. The contact pressure profiles in atomic-scale contacts are substantially different than the one proposed in the continuum models for rounded contacts. On the basis of these results, we formulate models whose solution can be computed analytically for the contact stiffness in the two extreme cases, and we show how to bridge between the atomic and continuum scales with atomically informed geometry of the contact. Finally, we study the predictability of continuum-based models to identify the onset of plasticity based on a resolved shear stress criterion and discuss their limitations in predicting the strength.