Accelerated discovery of the Al–Ag2Al–Al2Cu eutectic coupled zone through genetic optimization

Microstructure control in multi-component alloy systems is made complicated by the multiplicity of phases as well as the non-equilibrium nature of solidification. In particular, primary phases in a eutectic matrix may have a detrimental effect on the properties of the cast state. Instead of searching for wholly eutectic microstructures within a vast combinatorial space, we harness genetic algorithms (GA) to identify the stability field (coupled zone) of the three-phase Al–Ag2Al–Al2Cu eutectic, using previously reported observations as input. The output maps of the coupled zone have an accuracy of greater than 85% when compared against experimental data. In parallel, we conducted our own directional solidification (DS) experiments in order to assess the validity of our models. GA predictions of the eutectic coupled zone were further refined by incorporating our experimental observations into the constitutive growth laws. Our iterative approach of GA and DS not only improves the accuracy of the shape of the eutectic coupled zone in composition-velocity space, but also provides predictions on the microstructures of Al–Ag–Cu alloys where no such observations exist.