Zinc is a metal widely used as a coating for anticorrosive protection in the steel industry [1]. Despite it has been used for this scope for many years, a complete understanding of its deformation mechanisms is still lacking. Moreover, modelling its mechanical behaviour is an extremely challenging task. At the microscale, it presents strongly anisotropic elastic and plastic behaviours. At the atomistic scale, it presents a Hexagonal Closed Packed structure with an unusually large c/a ratio, equal to  1.856,  13.65% larger than the ideal ratio for a perfect HCP structure [2]. To date, to the best knowledge of the authors, there is no interatomic potential for Zn that can predict simultaneously the HCP crystal structure as the equilibrium structure, the lattice parameters, and the elastic constants. These properties, along with Stacking Fault Energy surfaces and Universal Binding Energy relation, are fundamental to understanding the plastic and the cracking behaviour of materials. Moreover, being able to correctly model dislocations and their glide at the atomistic scale, gives us the chance to translate this information to a continuum model at the microscale, such as Crystal Plasticity.
In this work, we make us of a recently developed Machine Learning scheme, which is Atomic Cluster Expansion [3], in combination with Hyperactive Learning Algorithm [4], to develop an interatomic potential for Zn that is DFT-accurate. By using the newly developed potential, we extract dislocation properties such as the Peierls Stress, which can be translated into the Critical Resolved Shear Stress to activate dislocation glide and hence plastic deformation in Crystal Plasticity.
Cristal Plasticity can then shed light on the deformation mechanisms of Zinc coatings with complex microstructures, improving our understanding of their mechanical behaviour and tracing guidelines for improved designs.

References
[1] A.R.,Marder; Progress in materials science, 2000, 45(3), 191-271.
[2] M. ,Yoo; Metallurgical transaction A, 1981, 12, 409-418.
[3] R, Drautz; Physical Review B, 2019, 99.1, 014104.
[4] C., van der Oord et al.; arXiv preprint, 2022, 2210.04225.