Extracting Anisotropy Strength and Interfacial Free Energy of Al-Zr Alloy Under Rapid Cooling Conditions Using Molecular Dynamics Simulations

Additive manufacturing is one of many growing industrial techniques with the benefit of creating new and unique materials, designs, and especially alloys. Dendrites are branching microstructural defects that will always form during the manufacturing of these alloys. Understanding and controlling the solidification to limit the formation of these dendrites is in high demand in the additive manufacturing field. The solid-liquid interfacial free energy and its associated anisotropy are key quantities to determine the microstructure formation of these dendrites. Other research has shown the addition of 0.7wt% Zr to an Aluminum Alloy AA5083 significantly increased the mechanical properties of the alloy. It is possible this is due to the dendrite formation on the atomic level and thus an Al-1at%Zr system is analyzed for this experiment. Molecular Dynamics (MD) simulations implemented in the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) were used for this atomistic modeling. Using an Embedded Atom Method (EAM) potential, a Nose-Hoover thermostat and barostat, and a time step of 1 fs, a large, thin structure of Al-1%Zr was created with a solid center and two liquid ends. The system was allowed to equilibrate under an NPT ensemble at the melting/undercooling temperature and two rough, quasi-1-dimensional solid-liquid surfaces were created. The Capillary Fluctuation Method (CFM) was then utilized to analyze the solid-liquid interface. CFM relates the Fourier Amplitude of the rough surface to the wavenumber, k, and with it, the interfacial stiffness, S, is calculated for three different crystallographic orientations. Through a cubic harmonic expansion, the interfacial free energy, γ, as well as the anisotropic parameters ε₁ and ε₂ are then determined. With this, the interfacial free energy and the anisotropic parameters are analyzed at various compositions to find any correlation.

Acknowledgements: The work is supported by ARL Grant No. W911NF-2020032 and used the Extreme Science and Engineering Discovery Environment (XSEDE) TACC at the stampede2 through allocation [TGDMR140131]. This work utilized resources from the University of Colorado Boulder Research Computing Group, which is supported by the National Science Foundation (awards ACI-1532235 and ACI-1532236), the University of Colorado Boulder, and Colorado.